CONCRETE GODE HANDBOOK

An explanatory handbook to
the Code of Practice for Structural Use of Concrete 2004

Structural Division
The Hong Kong Institution of Engineers
 FOREWORI)

It is my pleasure to represent the Structu¡al Division of the Hong Kong Institution of

Engineers to express ou¡ sin<.:ere gratitude to the author ofthis handbook, who has spent
more than ayear io prepare and write up this practical guide to the Code ofPractice fo¡
Structu¡al Use of Concrete 2004.

The Hong Kong SAR Government has recently issued several new codes ofpractice for
structural design. These codes ofpractice have incorporated the state-of-the-art know-
how and many good practices developed locally aad worldwide, and collectively they
¡epresent a major advancement in stuctr¡ral engineering. In many ways, these new
codes a¡e quite diflerent from the previous ones. I¡ order 1o help users of the new codes
to better understand the rationale behind the va¡ious changes and adapt to the new codes,
the Stuctural Division of the Hong Kong Institution of Engineers has called upon
experts in the ¡eleva4t fields to prepare a series of explanalory handbooks on the new
codes ofpractice. This particular handbook is the second ofthe series on the publication
list. The fìrst one on wind loading was published in June 2005. After this one, which rs
published herein, the third and the remaining handbooks will be published within several
months Anyone inte¡ested in purchasing these handbooks may visit the website of the
Structual Division for details

The author ofthis handbook has many years ofpractical experience He specializes in
both concrete materials and concrete structües. In this handbook, he has explained very
clearly the stuctwal concepts and rationale behind the various clauses of the code and,
where appropriate, provided typical worked examples to illustrate the new design
procedures. In addition, he has put in good practices for the design of high-performa¡oe
st¡ucturos. On the whole, the handbook has been written in a straight forward and easy
to undeßtand marurer. Apart from serving as a practical guide, it should also be a useful
reference for undergraduate students and young engineers to learn the design of concrete
slructurcs.
published by the structural Division of the Hong Kong Institution of Engineers
Lastly, would like to congratulate the author for his fine wo¡ks in writing up the
I
The content of this Handbook is an expression of opinion only. The guidance and handbook, which I believe would greatly benefit the industry in various perspectives,
recommendations given including quicker adaptation to the new code, avoidance ofmisirterpretation ofthe code
Handbook in the ligbt o requirements, easier grasping of the new design procedures, promotion of good design
Hong Kong Institution practices and upgrading ofthe design standard in Hong Kong.
agents assume no fespo
and its content.

No part of this Ha¡dbook may be reproduced, stored in a ¡etrieval system oI transmitted

in any form or by any means without prior permission of the structural Division of the
Hong Kong Institution of Engineers. k. P¡of. K.K. Choy
Chairma¡r,
Structural Division,
ß'irst publishetl 2006 HKrE (2005/2006)

ISBN-I3: 918-962-7 619 -19-2

ISBN-I 0 : 9 62-7 619 -19 -l
 Lovtc'{elø 6dg ¡tr,rvlUæf
PREFACE CONTENTS

After two years of intensive sfudy and painstaking new Code of Foreword I
Practice for Structural Use of Concrete was p Department,
Govemment of the Hong Kong Special Administrat 2004. Preface ii

ch is based on the limit state design philosophy and covers also high_ Contents llr
is a great advancement compared to the previous one published in
resembles BS8l10 a¡d Eu¡ocode 2 in certain ways. the new code has General
1.1 Scope I
(a) 1,2 References 1

1.3 Assumptions I
(b) 1.4 Definitions 2
1.5 Symbols 2
(c) Iocal requirements as stipulated in the practice Notes for Ap and RSE issued bv
Buildings Depafment throughout the years; and Basis ofdesign
(d) resea¡ch findings ofthe local unive¡sities on characteristics oflocal materials a¡d 2.1 Requirements 3
on structural use ofhigh-sfuengûr concrete. 2,1,1 Perfounance standards 3
2.1.2 Design working life
2.2 Principles of limit state design
2.2.1 General 5
2.2.2 Ultimate limit state (ULS) 6
2.2.3 Serviceability limit state (SLS) 12
23 Loads 13
2.3.1 Design loads 13
2.3.2 Loads for ultimate limit state t4
2.3.3 Loads for serviceability limit state
2.4 Materials 15
2.5 Analysis and verification 15
2,6 New and altemative methods 15
analysis aad when necessary additional scientific investigation,
Material¡
3.1 Concrete ló
3.1.1 General l6
3.7.2 Elastic defomation l6
3.7.3 Creep and shrinkage 17
3.1.4 Stress-str¿in relationships for design 18
3.2 Reinforcing steel 20
íke to tha¡k his colleague, I¡. Dr. Fraucis T.K. Au, fo¡ his carefr¡l 3.3 preshessingtendons 20
his resea¡ch students, Ms. S,L. Chau, N¿f¡. H.C. lVong, Mr. p.L. 3.4 Prestressingdevices 20
Mr. Y.K. Lam, for thei¡ assistance in the preparation of this 3.5 New materials 20

Durability anil fire resistance

4.1 Objectives 2t
4.1.1 Durability 2l
4.1.2 Fire resistance 2l
Albert K.H, Kwan 4.2 Requirements for durability ¿z
June 2006 4.2.1 General 22
7

4.2.2 Design for durability 25 7 .2.2 Confol of cracking without direct calculation (deemed-to-satisfr) 82
4.2.3 Exposure conditions )1 7.2.3 Assessment of crack widths 82
4.2.4 Cover 27 7.2.4 Early thermal cracking 84
4.2.5 Concrete materials and mixes 29 7.3 Deformations 89
4.2.6 Mix proportions 30 7.3.1 Generalconsiderations 89
4.2.7 Mix constiluents JI 7.3.2 Excessive response to wind loads 90
4.3 Requirements for fire resistance 31 7.3.3 Excessive vibration 91
7 .3.4 Limiting defl ection without direct calculation (deemed-to-satisf, ) q1
Structur¡l analysis 7 .3.5 Calculation of deflection 92
5.1 General provisions 5¿ '/.3.6 Calculation of cuwature 93
5.1.1 Generalprovisions 32
5.1.2 Methods of analysis 33 Reinforcement: general requirements
5.1.3 Load cases and combinations 5+ 8.1 General 94
5.1.4 Imperfections and second order effects 35 8.2 Spacing of reinforcement 94
5.2 Analysis of structu¡e J) 8.3 Permissible intemal ¡adii for bent bars 94
5.2.1 ldealisation of the stucture 35 8.4 Anchorage oflongitudinal reinforcement 95
5.2.2 Analysis of sections for ultimate limit stâtes 8.5 Anchorage oflínks a¡d shear reinforcement 97
5.2.3 Analysis of sections for serviceability limit states 38 8.6 Anchorage by welded bars 97
5.2.4 Simplifications 38 8.7 Laps and mechanical couplers 97
5.2.5 Monolitl¡ic frames not providing lateral restraint 38 II Additional rules for la¡se diameter bars 99
5.2.6 Frames providing lateral stability 40 8.9 Bundled bars 100
5.2.7 Slabs 8.10 Prest¡essing lendons 101
5.2.8 Corbels and nibs Àa

5.2.9 Redistribution of momenrs À^

Detailing of members and particular rules
5.3 Second order effects with axial loads +J 9.1 General 104
5.4 Shea¡ walls 43 9.2 Beams 104
5.5 Transfer structu¡es 44 9 2.1 Longitudinal reinforcement 104
5.6 P¡ecast elements 44 9 2.2 Shea¡ reinforcement 106
9.2 3 Torsion reinforcement 106
Ultimafe limit states 9.f Solid slabs t07
6.1 Members in flexu¡e 45 9.3.1 Flexuralreinforcement 107
6.1.1 General 4) 9.3.2 Shea¡reinforcement 108
6.1.2 Beams À<
9.4 Cantileveredprojecting structues 108
6.1.3 Solid slabs supported by beams or walls 50 9.5 Columns 110
6.1.4 Ribbed slabs 5l 9.5.1 Longitudinal reinforcement il0
6.1.5 Flatslabs 52 9.5.2 Transverse reinforcement 110
6.2 Membe¡s axially loaded with or without flexure 54 9 6 Walis 111
6.2.1 Columns 54 9.6.1 General 111
6.2.2 lValls 58 9.6.2 Vertical reinforcement ll1
OJ Torsion and combined effects 6l 9.6.3 Horizontal reinforcement 111
64 Design for robusñess against disproportionate collapse 63 9.6.4 Transverse ¡einforcement 111
65 Corbels and nibs 65 9.6.5 Plain walls 111
66 Stai¡cases 65 9.7 Fou¡dations I2
67 Foundations 65 9.'l.l Pile caps L2
68 Beam-column joints 68 9.7 .2 Column and wall footines t2
9 7,3 Tie beams 12
S erviceability limit súates 9.8 Co¡bels r2
7.7 General 78 9.9 Detailing for ductility 13
7.2 Cracking 80 9.9 | Beams l5
7.2.1 General 80 9.9.2 Columns 15
10 General specifi cation, construction and workmanship 12.3.6 Deflection of beams t42
10.1 Objectives t17 12.3.7 Ultimate lirnit state for beams in flexure t42
10.2 Construction tolerances r17 12.3.8 Desigl shear resistance of beams l+5
10.3 Concrete tr7 12.3.9 Torsion 144
10.3.1 Constituents TI7 12.4 Slabs t44
10.3.2 Mix specification 1t7 12.5 Columns 144
I 0.3.3 Methods of specification, production cont¡ol and transpoÍ 118 12.6 Tension members r44
10.3.4 Sampling, testing and assessing conformity 118 12.7 Prestessing t45
10.3.5 Placing and compacting r20 12.8 Loss ofprestrsss, other than frictiou losses 145
10.3.6 Curing 122 12.8.1 General 145
10.3.7 Conc¡eting in hot weather t¿> 12.8.2 Relaxation of steel 145
10.3.8 Formwork and falsework t¿t 12.8.3 Elastic deformation of ooncrete 145
10.3.9 Surface finish lZO 12.8.4 Sbrinkage of concrete 146
10.3. 1 0 Construction joints 127 12.8.5 Creep of concrete r46
10 3.1 I Movement joints 127 12.8,6 Draw-in during anchorage 146
l0 4 Reinforcement 128 12.9 Loss ofpresstress due to friction r46
10.4.1 General 128 12.9.1 General 146
10.4.2 Cutting and bending 128 12.9.2 Friction in jack and anchorage 146
10.4.3 Fixing 128 I 2.9.3 Friction in the duct due to unintentional va¡iation
10,4.4 Surface condition r28 ftom the specified profile 146
10.4.5 Laps andjoinls 129 12.9.4 Friction due to curvature offendons 147
10.4.6 Welding 129 12.9.5 Lub¡icants r48
10.5 Preshessing steel t?o 12.10 Transmission lenglhs in pre-tensioned members 148
10.5.1 General t29 12.1 1 End blocks in post-lensíoned members 148
10.5,2 Transport and storage 129 12.11.1 General 148
10.5.3 Fabrication 130 12. 1 1.2 Serviceability limit state 148
10.5.4 Placing 130 12.1 Ultimate limit state
1 .3 149
10.5.5 Tensioning 131 12. 1 2 Considerations affecting design details 149
1 0. 5.6 Protection and bond of prestressins tendons t32 12.12.1 Genenl 149
l0 5.7 Grouting 132 L2.l2.2Limilzlío:ns on area of prestessing tendons 149
72.12.3 Cove¡ to prestessing tendons 149
lt Qu¡lity assurance and quality control 12.12.4 Spacng of prestessing tendons and ducts 150
11.1 Scope 135 12.12.5 Longitudinal rei¡forcement in prestressed concrete beams 150
I 1.2 Quality assurance 135 12,12.6 Li¡ks in prestessed concrste beams 150
1 1.3 Classification of the control measures 135 12.72.7 lmpactloadtng 150
11.4 Verification syslems 1)6
1 1.5 Control of each stage of design and construction process 136 t3 Lo¡d tests ofstructures or prrts ofstrucúures
ll.6 Cont¡ol ofdesiga 136 13.1 General 151
I17 Contol ofproduction and constuction 136 13.2 Testloads 151
13,3 Assessment of results 151
L2 Prestressed concrete 13.4 Test criteria t52
't5,
12.1 Basis ofdesign 138 13.5 Specialtests
12.2 Strucnues a¡ld sfuctural ûames 139
12.3 Beams 140
12.3.1 General 140
12.3.2 Slender beams 140
12.3 3 Continuous beams 140
12.3.4 Serviceability limit state for beams 140
12.3.5 Stress limitations atûansfer for beams 141
 GENERAL

1.1 Scope

This code covers mainly the aaalysis, design and construction of reinforced and
prestressed concrete buildings, where the concrete is made of normal-weight
aggregate. Precast components are not covered; for the design of precasl
components, a separate code is refened to.

Any parts of the buildings subjected to highway loadings, which may be more
appropriately regarded as highway slructules or bridges, a¡e also not covered in
this code; for any such parts, a highway design code is refened to.

Although not explicitly stated, this code does not deal with earthquake loads or
earlhquake resistant design. Earthquake ¡esístance is not norrrally a desiga
requirement for Hong Kong. If earthquake resistance is demanded, it is
¡ecommended to consult other codes ofpractice for the design. Nevertheless, it is
considered prudent to pay particulal attention to the detailing of the concrete
structures so as to provide a ¡easonable amount of ductility, which is needed not
only for eafihquake resistance, but also for robustness and overall safety of the
struclures.

1.2 References

This code rnakes refe¡ence to the other codes ofpractice currently in use in Hong
Kong and the Hong Kong Constr¡ction Standards CS i and CS2.

As for codes/stand¿rds elsewhere, iÎ is not possible for these codes/standa¡ds to

cover every aspect of concrete building struchles. tühere provisions cannot be
found in these codes/standards, the engineer may resort to other acceptable codes/
standards elsewhe¡e. However, it should be borne in mind that the cha¡acte¡istics
of concrete made in different places are not quite the sarne because of the
differences in raw mate¡ials, curing conditions and concreting practices, and that
the design standards in different places may vary quite substantially. Therefore,
when applying codes/standards elsewhere to Hong Kong, the engineer will have
to provide justifìcation or evidence that the mate¡ial parameters adopted are
appropriate under the Hong Kong conditions and that the ove¡all standard of the
design is not inferior to that stipulated in this code.

1.3 Assumptions

Concrete is a variable materiai, of which the quality is sensitive not only to

variations in the mix proportions, but also to the wo¡kmanship and the curing
conditions. Hence, the reliability of the material is dependent on the quality
@lank page) conl¡ol of the concrete production and tbe site supewision during concreting.
Adequate quality control and sile supervision have been assumed in this code.
 The engineer responsible for site supervision should satisff himseiflherself that
the concrete in the fi¡ished structuÌ€ is of the required quality and keep all
records of the concrete tests. If there is any special construction sequence or
curing regime designed 1o mitigate cracking or to achieve certain special
properties, it should be advisable to b¡ief all the site staff involved beforehand
and cany out monitoring to ensure stict adherence to the specified procedu¡es.
ofa concrete struclure can be fairly complicated. Even
The structural behaviou¡
when the applied load is relatively small, the concrete stucture might have
already cracked. Hence, when estimating the stiffness of the concrete stn¡ctu¡e,
the possibility of premature cracking should be considered. Moreover, a conotete
struchue has in general only finite and sometimes rather limited ductility, which
is dependent on the concrete grade and the details ofthe reinforcement provided.
Hence, when applytng inelastic or plastic analysis, due considerations should be
given to whether the concrete stntcture has suffrcient ductility to meet the
ductility demand In any case, the engineer should ascertain himself./herselfthat
the assumptions made in the structural analysis are valid. When unconventional
mate¡ials or slructural forms are used, it may be necessa¡y to verifr the validíty of
the design assumptions by rigorous analysis or even prototype testing.
It has been assumed in the code that the design is ca¡¡ied out by an appropriately
qualified design engineer and that the stuctu¡al modelling and design
calculations contain no mistakes. It is the design engineer,s responsibility to
au-ange design scrutiny of the strucn¡ral modelling and independent checking of
the correctîess and accuracy of all the design calculations. The various safety
factors provided in the code are not supposed to cover any mistakes o¡ er¡ors in
the stuctural modelling and design calcuiations.
1.4 Definitions
The va¡ious terms in the code may have specific meanings as defined in Seclion
1.4 of the code. Common sense interpretation should not be relied on and this
section should always be referred to whenever there is doubt to the exact meaning
of any term.
1.5 Synbols
The majority of symbols used are taken f¡om 858110: Parts I aud 2. Other
symbols are taken ûom Eurocode 2. Where the same symbol has been used in
these two codes fo¡ more t¡an one definitions, a new symbol has been defined so
that each symbol used in the code has only one specific meaning.
2 BASIS OF DESIGN
2-I Requiremenús
2.1.1 Performance standards
In simple technical lerms, the basic perfonnance requirements of a building
srucnEe ¿üe:
. slractural adequacy (have sufficient strength and stability to resist the
largcst possible loads and sufficient endurance to resist repeated loads
during the expected life);
o semiceability (free ofexcessive deformation, vibration and cracking etc);
t durability (able to last the expected life with only normal maintenance
required);
c fire resistance (able to withstand fire up to a cefain period oftime); and
o robusîness (will not cause disproportionate collapse when subjected to local
rlamage due to accidental overload or impact).
the main function of the code is to sel the desip standa¡ds for the above
performance requirements. However, it should be bome in mind that setting the
design sta.ndards is actually a social-economical issue. None of the above
perfonnance requirements can be demanded or achieved in absoiute sense. For
i¡stance, there is no absolute stuctural safety (encompassing both stuctural
adequacy and robustaess) in this world. No matte¡ how high the safety stauda¡d
is, there is always a probability, albeit small, that the structure will fail. The
safety standard oould be set higher so that the risk of füilu¡e would become
smaller, but at the sa.rre time the cost of constuction would soa¡ and the structure
might become excessively bulky to defeat its intended fr¡nction. Likewise, there
is no absolute durability because it is physically impossible for a structure to last
forever. Therefore, in reality, i1 is necessa¡y to d¡aw a line somewhere and set
performance slanda¡ds that are high enough to ensure an appropriate degree of
reliability and yet practicable and econorrical.
The design standards set in the code have been developed over a long period of
time with the best of our engineering knowledge a¡rd skill embodied. They are
generally comparable to most national standards, especially the British Standards,
and may therefore be considered in line with intemational practices. On the
whole, they represent a reasonable balance between the perforrnance standa¡ds of
buildings and the economic/sustainable development ofûre society that is so fa¡
acceptable to the general public.
The desip standards stipulated in the code a¡e recommended minimum standa¡ds
for general buildings under no¡m¿l conditions. If, for some re¿rsons, such as high
susceptibilíry to unusual loading conditions or simply high value/importance of
the building, it is considered prudent to design the building stnrcture to a higher
standard, it is up to the desiga engineer to do so, wilh perhaps consent tom the
client.
2.1 .2 Desiga working life
structures a¡rd other common structufes, the indicative design working life is 50
years while for design working life category 5 structures, such as monumental
Regarding th the code specifies the expected life in building structures, bridges and other civil engineering structu¡es, the indicative
te¡ms of the assumed ìo be 50 years for general design working life is 100 years.
buildings an :turcs. The usage of this terminoiágy is
explained below.
From the above, it can be seen that for normal buildings, the design life specified
in BS7543 is 60 years whereas the design working life specified in ENl990 is 50
lerminologies and definitions have been used for years. However, the difference between 60 and 50 years does not really carry
building, narnel¡ design life, service life and desisn much significance because up to now the¡e is no rigorous scientific basis for
these values. In actual practice, the environmental and usage conditions of the
designrireandservice,,J-"::î#tåä;ïjiliJ¿l.'tr*:l',år,,ï:,,i'iå buildings could vary from one extreme to another and similar buildings designed
in EN documents, as depicted below
to the sane durability stand.ard might end up with very different degrees of
deterioration after 30 or 40 years. Hence, tle specified design iife or design
lnBS7543:1992,
. working life should be t¡eated only as a nominal value. A value of 50 years has
d3sien lfe is the period ofuse intended by the designer (e.g.
'
" as stared by the
designer to the cl ient to support specification decisiãns),
been adopted in the new code, This does not mean that the buildings designed as
per ÎÌre new code would not last longer than 50 years. Like many existing
o seryice âf is the actual period of time during which no excessive buildings that have survived longer than 50 years, it is quite possible for the
expenditure is required on operation, maintenance oirepair ofa component, buildings, after major repairs or retrofitting, to suwive up to 60 years or so.
or re-construction, and
¡ required service life.is the service life specified to meet users'
requiremetrts
(e'9, as stated in the crient's brief foì a project or 2.2
in the p"Ã."t-uo". Principles of limit state design
specification).

ln EN documents, design yorking @ is employed. This is the period 22.1

of time General
during which a structure that has undergon"ìonnal mainter".ce-i"
*lik"ly ì;
require major repairs.
The new code is based on the limit state design philosophy. According to the
limit state design philosophy, a structure has to be designed to satisfu not just one
I_n new code, the terms design life, service life and required service perfonnance requirement but a multitude of performance requirements, or in
_the life
defined in 8s7543 are not used because in practice the other words, to be designed to stay within limit slates, which define the bounds of
designìife should not be
different f¡om the required service rife and tîre purposeful differentiation acceptable performance limits, under all possible usage condilions
between
design ce life would *îy .u"r" confl¡sion. Relatively, it
should to just use the tenn design working life, ;hicíis Limit states to be considered in the design are mainly the ultimate limit state and
more cl ENstandards. the serviceability limit state. The ultimate limit state is related to slructu¡al safety
limit state is related to proper functioning of the
whereas the serviceability
Apart from the diference in terminorogies, the British standa¡ds stucture. The ultimate limit state is generally considered more impofant.
and the EN
documents also differ in the specifiecr aesign life or design
working rtfr p*"dr, Hence, most engineers choose to first design fo¡ the ultimate limit state, thinking
as depicted below.
that the ultimate limit state is likely to be the most critical límit state, and then
check that the rernaining limit states will not be reached.
any design
for category However, the va¡ious structural modellìlg assumptions made during ultimale
ldings, new limit state design, such as the neglect of the torsional stifftress of frame members
ign life is a and the neglect of the out-of-plane stifftress of shear walls, which a¡e on the
conservative side ûom stuctural safety point ofview, are not necessa¡ily on the
civic and o_tler high quatiry-buildings, å""::1iä,,t,\:'i,iffHT:n
years' ln 855400: Part 1: 1988, the design life ofsteel, concrete
J::Ï; conservative side when applied to sewiceability limit state design. When the
serviceabiliry limit state is considered, the torsional stiffrress of frame members
bridges is ¡ecommended to be 120 years. "ia.".p"rrt. and the out-of-plane stiffrress of shear walls should not be neglected because the
torsional moment and out-of-plane bending moment induced may cause cracking
On the other band, ENl990: 2002 @) recommends tliat (see Table 2.1 of of the concrete. The common practice of using the structural model originally
the
document) for design working life category 4 sfuctures, developed for ultimate limit state design also for serviceability limit state design
such as buildins
may not be acceptable.
2.2.2 Ultimate limit state (JLS)
ULS design is concemed mainly with strength, stability and robustness of fhe
shucture, although collapse, overtuming and buckling have also been mentioned
as if they are sometling differenl in Clause 2.2.2,1 of the code (somehow
robusbess has been missed out in this clause). Collapse may be a coDsequence
of strenglh, stability or robustness failure while oveduming arid buckling are both
consequences ofstability failure. Hence, collapse, overtuming and buckling are
only duplicated concepts. It should be more systematic to just concent¡ate on
strength, stabilþ and robusb.ess, which already cover all aspects of ULS design,
shear walls (due to small lateral stiffness in that direction). Likewise, if all the
shear walls are ananged to be i¡tersecting at one vertical axis, tle building will
tend to rotaþ about tlat axis (due to small torsional stifhess). Another
possibility is the connection of a cantilever beam to a shear wall at a direction
perpendicular 1o the wall without designing for the concentated out-of-plane
bending moment that could be induced at the beam-wall joint.

Strength is the primary concern, although Section 2.2 of the code touches only
lightly on this aspect. When designing for shength, the design loads are applied
to the strucnre a¡rd the structu¡e is first analysed for the intemal member and
section forces (analysis of structure). Affer obtaining the internal member a¡d
section forces, each seotion is then analysed for the purpose of reinforcement
detailing (analysis of sections). A common practice is to carry out the analysis of
structure by means of linear elâstic analysis and the aaalysis of sections by means
of inelastic or plastic aaalysis, With this design practice, the various sections of
the structu¡e would yield at more or less the same time when the stnrcture is
loaded to the point of collapse. Such design practice is acceptable ûom the (a) A uuss with an unintentional mechanism
structural safety point of view but the resulting design may not be the most (note that all nodes except A and H can move \4'ithout intemal restraint)
economical. Nevefheless, it has the major advantages that the desìgn procedures
are relatively simple and that computer softwa¡e for línear elastic analysis of
structure and inelastic/plastic analysis of sections are readily available.

If the analysis of stucture is caníed out by means of inelastic or plastic analysis

so as to take advanlage of the redisnibution of internal forces from the sections building
that have already yielded to t̡e other unyielded sections, a more economical
design may be obtained. The method of redisfiibution of moments (see Section
5.2 ofthe code) is an abridged version of such practice. However, care is needed
to check and ensure that all the yielded sections have suffrcient ductility to
withstand the plastic deformation so caused until the structu€ collapses.
Ductility check is in general not easy because a rather rigorous limited ductility
elasto-plastic analysis method is required. If such rigorous analysis method is not
I
resorted to, it is betler to slay with linea¡ elastio analysis of stucture and apply
I

only those rediskibutions of internal forces that are explicitly permitted \Mith clear floating
guidelines given in Chapter 5 ofthe code. upwards

Stability is the ability of the structure to prevent bodily movement and buckling overtuming
of any part of the strucfu¡e. Bodily movemeDt is the displacement of part or
whole of the stmctue without internal restraint (i.e. without straining any part of
the structure) and without proper extemal restraint (i.e. without being properþ
affixed or rest¡ained by extemal means). It is basically a first order effecl.
Bodily movement can occur due to forrration of one or more unintentional
mechanisms or due to movement of the whole structu¡e (such as sinking, floating,
sliding or overtuming) relative to the foundation, as illust¡ated in Figure 2.1. @) Movement relative to foundation
Unintentional mechanisms ôould be fonned in many ways, as shown in Figure
2.2. For instance, if all the shear walls of a building are arranged to be parallel to
each other, the building will tend to move in the direction perpendicular to the Figure 2.1 Examples of bodily movement

d
L
,-

structural elements involved have initial curvature and./or eccentricity, or be very

sudden with no prior waming if the structu¡al elements are perfectly straight and
in line with the applied loading. In any case, buckling is a kind ofb¡ittle failure
and is therefore extremely dangerous. Buckling occurs mainly in slender
stmctu¡al elements. Fom.rnately, concrete sÍuctures a¡e mostly quite bulky (ar
+ least compared to steel structures) and so buckling failure of concrete structures
lateral is uncommon. However, it could happen ifthe structu¡e is not detailed properly.
movement For instance, if the analysis of stn¡cture is carried out using a two-dimensional
Plan of building analysis method with all possible out-of-plane actions neglected and at the end no
restraint against the out-of-plane movement of the structure is provided (many
(a) All shear walls parallel io each other inexperienced engineers are lempted to forget about the possible out-of-plane
actions bpcause there is no oul-of-plane loading acting on the two-dimensional
model ofthe structure), out-of-plane buckling could occur. Excessive reliance on
computer analysis rnay also cause the same problem because many inexperienced
engineers just concentate on structual elements subjected 1o loading and are not

aø
awa¡e of the factthat sometimes structural elements have to be added to provide
restraint rather than to cany loading.
rotation
Robustness is the faii-proof ability of the structu¡e against disproportionate
Plan of Plan of collapse when subjected to local damage due to accidental overload or impact.
building building The aim of providing robustness is to avoid the situalion whereby damage to a
small part of the st¡uctu¡e or failu¡e of certain individual elements would iead to
@) All shear walls intersecting at one point collapse ofa large partor any majorpart ofthe structure. To provide robustress,
(a) at the global scale, the general afiangement ofthe structu¡e must be desigaed
to avoid any inherent weakness that could lead to progressive collapse, and (b) at
the local scale, every structural element should be designed to have at least some
nominal lateral resistance and be effectively tied to other parts ofthe sEucture,

(c) Canfilever bem contrected to shea¡ wall
Fì,gore2.2 Some examples of rmintentional mechanisms
Buckling is the loss of stability due to geometic nonlinearity arising from the
deflection of part or whole of the structu€. Unlike bodily movement, buckling
causes straining of the structure and is thus dependent on the stiftress of the
stuctral elements involved (mainly the lateral stiftess of struts and out-of-
plane stiffness of walls). It is basically a second orde¡ effect (also referred to as
P-A effecf by some ongineers). The loss of søbility could be gradual if the
Desigaing the structure at the global scale to avoid inherent weal<ness that could
lead 1o progressive collapse (or, in other words, desigring the structu¡e to have a
good structural integrity) requires carefuljudgement by an experienced engineer,
No simple structural analysis can replace such engineering judgement. Perhaps
some kind of collapse mecha¡ism analysis should be carried out to trace the
sequenre of events that could happen during collapse in order to assess the
structual integrity of the st¡ucture, which is required by Clause 22.23(b) of úe
code. Performing such collapse mechanism analysis quantitatively is not going
to be easy as the computations involved are exceedingly complicated and there is
still no computer softwa¡e available in the market, although resea¡ch has pointed
out the possibility of applying nonlinea¡ finite element analysis for this þurpose.
Anyhow, at the least, conceptual and qualitative collapse mechanism analysis
should be ca¡ried out to identi$, any key structwal elements, the failure of which
would cause the collapse of mo¡e tha¡ a limited pofion close to the element in
question. As a general rule, as far as practicable, the aEangement of the structufe
should be designed such that only a small number of structwal elements are
indispensable (these are called key elements) and the indispensable elements (the
key elements) must be designed to resist accidental loading and/or protected to
prevent removal by accident. Another general rule is to provide as much
redundancy as possible so that when a certain parî of the strucfirre fails, the
loading that it is carrying may be redistributed for eventual transmission to the
foundation without causing pro gressive collapse.
At the local scale, every stucnrral element should be designed to have at least For easy reference, a flowchart of the ¡obustness design procedure, similar to the
some nominal resista¡ce in all directions. For instance, a beam subjected mainly one given in BSSI 10: Part 1: 7997, is presented in Figarc 2.3.
to vertícal downward loading and sagging moment should also be designed to be
capable of resisting certain nominal vefical upward loading and hogging moment
because tåe direction of accidental toading is in general very unpredictable and
can act in any direction. When there a¡e elements that a¡e effective in resisting
load in only one direction, such as guy wires, stay cables and suspension cables,
particular ca¡e is needed to ensure that when the loading direction reverses, tbe
structure still has sufficient ¡esista¡ce to avoid initiation ofprogressive collapse.
ln addition, every structu¡al element should be properly connected to other parts
of the structue. For connecting precast components to the structure, monolithic
co¡nections are generally preferred. If practicable, the reinforcing bars to be
anchored at tbe connections should be mechanically connected or weided
together instead ofjust lapped side by side to improve the structwal integrity.

Following tle events of September 11 in 2001, much resea¡ch has been carried
out into the ¡obustness of strucfu¡es particularly with respect to disproportionate
collapse ofhighrise buildings. The two principal reports relating to this topic are
"Safety in Tall Buildings" published by the UK Institution of Sfructural
Ensure tïat stucture cu
Engineers and "láorld Tmde Cente¡ Building Performance Study" published by resist notional ho¡iæntal
the USA Federal Emergency Management Agency. Their recommendations load (see 2.3.1.4)
essentially follow the guidelines given in BS81 10 and Eu¡ocode 2, namely:
. IdentiÛing any key elements in each sfiuctrue, for which their failu¡e would
lead to the collapse ofa greater portion ofthe building than that local to the
element in question. Where these key elements cannot be designed out, the Provide periphcral, intemal,
horizonøl md vefical ties
design of these elements will lzke their importance into account using (see 6.4.1)
appropriate measu¡es.
. Ensuring the stucfires will be capable of safely resisting a notional
horizontal design ultimate load at each floor level.
. Ensuring the structurÊs will be provided with effective horizontal ties a¡ound
their peripheries, internally a¡rd to their vertical elements.
o Ensuring that the vertical elements of the structue will be provided with lies
such that úre removal of a vertical element, other tha¡r a key element, will
result in any collapse being limited to the portion of the structure in close
proximity to thal element.
These guidelines have been incorporated into the code ofpractice. Where for any which cannot be tied, be bridged ?
reæon it is not
For very special and important buildings, it may be Decessary 1o allow for the feasible !o
effects of paficula¡ h¿zards or fo¡ an unusually high probability of the structure introduce ties
surviving an accident even though damaged. the following
In such case, safety factors higher
procedue
than those given in the code may be required. Design such elements
should be
as key elemeils
adopted:
In Hong Kong, structures are normally designed with sufñcient sοbility and (see 2.3.1,4)

robustness so that the effects of impact and impassive loading should be small.
The danger against sabotage and t€Íorist attacks is not considered high in
general. Useful information on protective system against these loarlings can be
found in CEB and fib publications. If building structur€s are required to be
designed against these extreme actions, specialist literafure should be consulted
and the design canied out accordingly. Figure 2.3 Flowchart ofrobustrcss desigD procedure
(the clause numbers refe¡ to those in the code)

ll
2.2.3 Serviceability limir srare (SLS)
SLS design is concemed mainly wilh the ñrnctionality of the structure (depending
on the tlpe and intended usage of the structure, ñnctionality here may include
finess for purpose, user comfort, appearance and other special requirements such
as water-tightress of the structure). To satisff the SLS requirements, the
craching, deflection and vibration of the structure under normal usage are
analysed and checked against the acceptable limits stþulated in the code of
practice and ifapplicable additional requirements set by the client. Traditionally,
the fatigue resistance, durability arld fre resistøtce of the structure are also
evaluated and checked as a part ofthe SLS design, although these perfomrance
requirements should have been designed for in a separate design routine (fatigue
resistance, durabilþ and fire ¡esistance are not really subsets of serviceability).
The analysis of stmctu¡e for evaluation of cracking, deflection and vibration is
usually carried out by means of linear elastic analysis because the structure is
expected to ¡emain elastic rmder nounal usage. In many cases, where tle
cracking, deflection or vibralion are not expectecl to cause concern, deemed-to-
satisþ desip rules fo¡ dimensìoning and detailing of tle sructural elements (e.g.
maximum spar/depth,ratios, minimum reinforcement areas and maximum steel
bar spacings) have been provided so as to simplifr the design process.
Cracking affects the appeannce, durability and water-tightress of the structure.
Broadly speaking, there a¡e two categories of cracks: non-sfucfital cracks (also
called restraint-induced cracks) and skuctu¡al cracks (also called load-induced
cracks). Non-strucfiral cracks are due to sedimentation, sbrinkage or thermal
movenent ofthe concrete, which ifrestained, would induce tensile stresses large
enough to cause cracking. Structural c¡acks are due to tensile sbesses developed
in the concrete resulting fiom the applied loads. Non-structwal cracks may be
contolled by proper concrete mix desip, careful plaming of the constuction
sequetrce, proper contol of the temperature and moistre conditions during
cwing, provision of movement joints and provision of crack distribution
¡einforcement. On the other hand, stuctural cracks may be controlled by limiting
the tensile shesses developed in the concrete, proper reinforcement detailing and
provision of crack distribution reinforcement.
The necessity to limit the deflection a¡ises ftom the problem that excessive
deflection oftle süuctu¡e could lead to drainage difficulties, damage ofûe non-
structures such as finishes, partitions, glazing, cladding and services (these
normally could not accommodate too much deflection), and movement or
vibration that could affect the user comfort ofthe structure.
Vibration is of concern because it may cause discomfort to the occupants and./or
interference with proper function of sensitive equipment housed in the building.
The¡e a¡e two major sou¡ces of vibration: wind-induped vibration and machine-
induced vibration. Wind-induced vibration may be limited by contolling the
lateral deflection of the building under the equivalent static wind load or by
controlling the peak accele¡ation ofthe building under the dynamic wind load.
On the other hand, machine-induced vib¡ation should better be contolled by
isolatine the source ofvibration.
2.3 Loads
2.3.1 Design loads
The loads are defined in terms of cha¡acteristic loads, which should have a
reasonably high probability ofnot being exceeded under normal usage during the
design working life ofthe building. The cha¡acteristic dead load, imposed load
and wind load are given in the Code of P¡aclice for Dead and Imposed Loads for
Buildings and the Code of Practice on Wind Effects.
The design load, i.e. the load value to be used in the design calculations, is to be
taken as the cha¡acteristic load multiplied by an appropriate partial safety factor
(i.e. load safety factor), which is dependent on the type ofloading and limit state.
A set of desigrr loads for robustness design has, for the fi¡st time, been given in
the Hong Kong codes. The design loads for robustness (no partial safety factor is
to be applied) comprise of:
(a) A notional ultimate horizontal load at each and every floor level equal to
l{d"r,rc 1.5% of the characteristic dead weight of the stuctt¡re between mid-height
'\¡o1r),,, ofthe storey below and mid-height ofthe storey above.
(b) A notionat ultimate load of 34 kN/m' on each key structu¡al element, acting
in any direction (vertical o¡ horizontal, upward or downward, and inwa¡d or
outward). The projected a¡ea of the key element is to be used for the
determination of the ultimate load.
(c) A notional ultimate load of 34 kò{/# on each building component
connected to a key structural element acting in any direction. The projected
a¡ea of the building component is to be used for the determination of the
ultimate load but the reaction to be transmitted to the key structural element
should be tlre maximurn that might reasonably be transmitted having regard
to the stength ofthe component and the strength ofthe connection.
Following the progressive collapse of Ronan Point in 1968 (a UK residential
building that suffered extensive damage following a gas explosion in one of the
apartments), regulatory requirements were introduced in llK to provide (in
buildings above 5 storeys) structural resistance with the aim of limiting damage
caused by an accident or misuse so that it is not disproportionate to the cause.
These rules are included in both the UK Building Regulations and the.B¡itish
Standard code of practice. The value of 34 kN/m' originates from resea¡ch
carried oul by the Building Resea¡ch Establisbment following the gas explosion
at Ronan Point. Measurements ofpeak p¡essuÌe following such incidents we¡e
assessed and the value of 34 kN/m' was found to be the peak pressure that was
exceeded by only 2% of incidents. It is therefore of the orde¡ of a 1 in 50 event.
This was considered to be a reasonable value to use in design.
The notional ultimate horizontal load in (a) above should not in general govern
the design of buildings in Hong Kong, because of the relatively large wind loads
to be designed for. What needs to be done during the design process is just to
make sure that the design ultimate wind load 1o be applied at each floor level is
not smaller than this notional ultimate horizontal load for robustness.
|
fuat6¡ uuhwe{x àL > \t/h'* L uJi,wl* {¿''i t"4
 In most cases, the notional ullimate loads in (b) and (c) above should have little
effect on the design ofthe building structure, but still design check is required to
confum compliance with these two robustness requirements. Perhaps the greatest
effect is on the provision of out-of-plane resistance to each shea¡ wall for
¡obustness. The projected area of a shear wall in the oul-of-plane direction is
quite large. This, multiplied by the notional ultimate load of 34 kN/m', could
lead to a fairly large loading. Even then, the robusfrress requirement would
govern the design of the shear wall only when the thickness of the wall is
relatively small.
2.3.2 Loads for ultimate limit state
The major loads to be considered for IJLS design are dead load, imposed load
and wind load, which are to be combined into at least th¡ee different load
combinations: (dead load + imposed load), (dead load + wind load), (dead load +
imposed load + wind load), as per Table 2.1 ofthe code. It should be noted that
although "lable 2.1 resembles the table in Clause 2.4.3 of BS8l10: Paú 1: 1997,
the load safety factors given in Tzble 2.1 a¡e not the sarre as those given in the
BritishStanda¡d. hBSSll0,thebeneficialloadsafetyfactorforimposedloadin
the (dead load + imposed load + wind load) combination has been set as 1.2.
This is actually wrong because the imposed load rnay be present or absent, and
when the presence of imposed load is beneficial, the absence of imposed load
would produce a more c¡itical condition. Therefore, the beneficial load safety
factor should have been taken as 0. In the new code, this mistake is corrected and
the beneficial load safety factor for imposed load is set equal to 0
Clause2.3.2.4 of the code suggests that for ULS design, the creep, shrinkage ald
temperature effects need only be considered where they are significant, for
example, for the verification of ultimate limit states of stabilþ where second
o¡de¡ effects are ofimportance; in most other cases, they need not be considered
at ULS, provided that the ductility and rotational capacity of the struqtural
elements a¡e sufficient. However, in actual practice, it is not easy to check
whelher the ductility and rotational capacity ofthe structural elements are really
sufficient because there is still no established analytical method for doing so.
The creep, shrinkage and ΀mperature effects a¡e generally more sigaificant in
relatively long and./or tall slructures, in which case, the resulting movement can
amount to 100 mm or more. It should be advisable 1o always inolude the creep,
shriokage and temperature effects in the ULS design of long (say, longer than
100 m) and/or tall (say, taller than 1 00 m) structures.
2.3.3 Loads for serviceabilify limit state
The major loads to be considered for SLS design are dead load, imposed load,
wind load, differential settlement offoundation (ifany), and creep, sbrinkage and
temperatu¡e effects (if any). For SLS design, the load safety factors for all types
ofloading a¡e taken as 1.0. I¡
most cases, ifthe design and detailing ofthe
structural elements a¡e carried out in acco¡da¡rce with the deemed-to-satis! rules
given in the code, no fu¡ther checks on SLS are required.
2.4 Materials
The stength of a material is defined in terms of its characteristic strength, which
should h¿ve a reasonably hìgh probabilþ ofbeing achieved with proper quality
cont¡ol and site supervisíon assumed. The desigr strengtb, i.e. the strength value
to be used in the design calculations, is to be taken as the cha¡acteristic sirength
divided by an appropriate partial safety factor (i.e. material safety factor), which
is dependent on the va¡iability of the st€ngth pa¡ameter, possible difference
between the in-situ strength and the laboratory measured strength value, resulting
uncertainty in the estimated sectional resistance a¡d limit state.
In the older 1985 edition of BSBI10, the material safety factor 7. for steel
rei¡forcement is taken as 1.15, while in the newer 1997 edition of BSBI10, the
mate¡ial safety factor ,,m for steel reinforcement is changed to 1.05. Both these
two material safety factors have been used in Hong Kong. However, beariug in
mind that the sou¡ces of reinforcement in Hong Kong are quite diversified wilh
varying degrees of quality control, the reduction of the material safety factor for
rei¡forcement from 1.15 to 1.05 is not considered advisable, at least at this stage.
Hence, the material safcty factor fo¡ reinforcement of 1.15 in the 1985 edition of
BSSI l0 is adopted.
2.5 Alalysis and verification
As stated before, the analysis for the purpose of reinforcement detailing and
verification of compliance with all the limit states consists of two stages:
r a¡alysis ofstucture; and
r analysis ofsections.
Guidelines for the analysis of stuchue and analysis of sections are given in
Chapters 5 to 7 of the code. However, ltrese rules are limited to design of
süuctures r¡nder static loads or equivalent quasi-static loads of wind. For
dynamic analysis, specialist guidance should be sought. Where non-linear
effects, such as those due to geometric nonlineæity or material nonlitrearity, are
significant, specialist guidance should also be sought.
New and alternative meúhods
New and altemative design methods not found in the code may also be used
provided it can be demonsl¡ated that the basic performance requirements in
Section 2.1 ofthe code are complied u,ith and that the overall standards ofthe
design are not inferior compared to the general standards stipulated in the code.
For such pu{pose, the design engineer msy have to verif, the accuracy of lhe
theoretical analysis by model and prototy?e tests and provide evidence that all the
performance requirements have actually been achieved by ñeld monitoring
anÜor testing.
 3 MATERIALS
3.1 Concrete
3.1.1 General
This section applies only to concrete made ûom the locally available rock
aqgrcgate (mainly granite rock aggregate).
Unlike the B¡itish Standard BS8l10, which covers only norrnal-stength concrete
up to grade C60, high-strength concrete up to grade Cl00 is cove¡ed in the new
code. This represents a great advancement as the usage ofhigh-stength concrete
is increasing and more design guidance is desperately needed.
The properties of concrete are dependent on the raw malerials (mainly the type of
aggregate used), the curing conditions and the local concreting practices. Hence,
the characteristics ofconcrete made in different places are trot quite the same and
the mate¡ial properties of concrete given in other codes are not necessarily
applicable to Hong Kong. To resolve this problem, the local universities have for
many years conducted resea¡ch on the properties of the local concrele. It has
been found that generally speaking the local concrete made with granite rock
aggregate has lower elastic modulus, lower strength wilh the same mix
proportion and larger shrinkage compared to most other concretes made
elsewhere. These characteristics of the local concrete have alreadv been
incorporated in the new code.
The ¡ecommendations made in this section on the properties of the local concrete
and on the usage of high-strengfh concrete are based mainly on the ¡esea¡ch
studies carried out at The University ofHong Kong.
3.1.2 Elastic deformation
'io''' år The elastic modulus of concrete is given in the new code in botb an equation
forrn and a lable form. The equation for the evaluation of elastic modulus is:
;)ú tít'f \
ti*{-'}r¡*,r <i¿nfu , t"'*J--I-'-¡'
e¡ù t: ¿lb1 E'=3'46'[h +3'21
LA y\r1) where E" is the shof-term elastic modulus of the concrete (in kN/mm2) and 71"
---fs the cube compressive strength (in Ninrmj). It is applicable to concrete with
Ko zl\Ç Subeshengthrangingfrom20to l00N/mm'.
The elastic modulus values evaluated as per the above equation agree closely
wilh the corresponding values given in the previous Hong Kong Building Code -
Buildings and Lands Departuen! Code of Practice for the St¡uctural Use of
Concrete: 1987 within the cube strength range of 20 - 45 N/mm2 and the Hong
Kong Highways Code - Highways Departrnent, Structures Design Manual for
16
Highways and Railways: 1997 within the cube strength range of 20 - 60 N/rnm'
Hence, up to a cube strength of 60 N/mm', there is little change in the elastic
modulus. The major change in the new code is the extension to cover cube
shengths from 60 to 100 N/mm¿.
It should be noted that when the mean elaslic modulus is required, the mean cube
strength should be used in the above equation to evaluate the elastic modulus and
when the characteristic elastic modulus is required, the characteristic cube
slrength should be used instead.
The elastic modulus of the local concrete as given in the new code is based on
extensive lests carried out at The University of Hong Kong, which have been
published in the foliowing paper:
Kwan A.K.H., Zheng W. and Lee P K.K., "Elastic modulus of normal- and high-
strenglh conc¡ete in Hong Kong", Transactions, Hong Kong Institution of
Engineers, Vol.8, N0.2, 2001, pp10-15.
31 3 Creep and shrinkage
The methods of estimating the creep and shrinkage of conc¡ete in the new code
are the same as those given in the Hong Kong Highways Code - Highways
Department, Structures Design Manual for Highways and Railways: 1997, which
are acnrally based on Appendix C of 855400: Part 4: 1990, According to
BS5400: Pa¡t 4, the creep and shrinkage ofconcrete may be evaluated using the
following equations:
þ"(t) = Ky K^ K" K. K j ,)' ,
ê",(1) = KL K, K. Kj
in which /" is the creep coefficient, a"" is the shrinkage strain, and the "K"
coefficients are as defined in BS5400: Part 4. However, over thË years, it had
been found from field measurements aad laboratory tests that the shrinkage ofthe
local concrete is generally much larger than that predicted by the 855400: Part 4.
The most probable cause ofthe much larger shrinkage ofthe local concrele may
be attributed to the properties of the granile rock aggregate being used in Hong
Kong. In order to allow for the larger shrinkage ofthe local concrete, the Hong
Kong Highways Code incorporates a modification factor c. into the formula for
shrinkage, as depicted below:
8",(t) = c, KL K. K" Ki
The modification factor c, has been assigned a value of 4.0 in the Hong Kong
Highways Code. This same modification factor is adopted in the new code
The modification factor of4 0 implies that the sb¡inkage ofthe local concrete is
about four times that of similar concrete in UK. Such implication has a¡oused the
controversy whether the sbrinkage ofconcrete in one place can be fou¡ times that
 of simila¡ concrete in another place, despite the fact that the shrinkage of concrete The stess-strain curve in fhe new code differs from that of BS8ll0/8S5400 in
is heavily influenced by the properties of the constituent materials. Some the following two wa¡rs:
gover¡ment departmetrts and universities are conducting long-terrr shrinkage (a) The initial gradient of fhe curve (i.e. .0" ) has been shanged to match with the
lests of the local concrete. Hopefrrlly, in the near future, when the tests a¡e 2Ço rt/y¡ty1¡>
completed this contoversy could be settled with a more reliable shrinkage model (b) at the extreme
developed specifically for the ooncrete in Hong Kong. Engineers a¡e advised to comprËssion fibre when the cotrcrete fails) with the inc¡ease ofthe concrete
keep an eye on the publication ofthe tost results and the new shrinkage model. strength l*
to beyond 60 N/¡¡m' has been taken into accoutrt so as to
extend the applicability ofthe curve to high-stength concrete.

3.1.4 Stress-stain relationships for design

However, although Table 3.2 has been referred to for the evaluation ofthe initial
gradient E , the method of applying the mate¡ial safety factor 7. has not been
The stress-strain curve of concrete given in the new code is compared to that of
BS81l0/8S5400 in Figure 3.1. speciñed. ln theory, the material safety factor should be applied before
evaluating the initial gradient using either Table 3,2 o¡ Equation 3.1 ofthe code.
Since after application of the material safety factor, the concrete stength
becomes l^/T^ nd is no longer a whole number, it should be better 1o use
oo Equation 3.1, whicìh after the material safety factor is applied, becomes:
Parabolic
curye
o
H
z":3.46.,fÇifi +t.zt
ît)

Afer changing the infial gra.dient E", the value of s0 rceds to be adjusted.
tt- Mathematically, for the initial part of the stress-stain cu¡ve to remaiu a parabolic
.ð. = 5.5 ]s curve, the value of eo needs to be set equal to 2 oo f E r. Hence, for mathematical
I /m compatibility, the value of ao should have been adjusted to:
in kN/mm2)
r-;- St¡ain
tf | 34u",ly^)
eo =2.4 x l}a
Il
6* = 0.0035 ùo -- --------;-
Lc

(a) Stress-stain curve in BS8l10/855400

which is in general larger than t$at ofBS8110/855400.

The value ofthe ultimate concrete sfrain eoo has to be ¡educed when the concrete

oo
strength 7* is higher than 60 N/mm2 because.high-sttn$ .o".t9_ggl9f4]L
has a shess-süain cuwe that would, u11s¡ ¡s¿shing the pealq dcali¡e-loarilapidly
dcali¡e-loarilapi
@gh-strength corcr€te is more brittle). To
allow fo¡ such effect, tbe code has provided the following formula for the
v) evaluation of e", :

e- = 0.0035 for /"" < 60 N/nm2

e* :0.0035 - 0.00006!fC';6õ) for f^ > 60N/mm2
ð"(s.'99".r3!9"_-t-z)
S,t¡ain This formula is based on the research study carried out ¿t The Universify ofHong
Kong which has been published in the following paper:
(b) Strgss-stain cuwe in the new code Ho J.C.M., Kwan A.K.H. and Pam H.J., "Ultimate concrete stain and equivalent
rectangular stess block for design of high-stength concrete beoms", Structural
Figure 3.1 Stess-stain curves of concrete Engineer, Vol.80, No. 16, 2002, pp26-32'
 3.2 Reinforcing sfeel
Ïhe-gedCS¡i$Qdand üreir mechanical properties given in the code are rhe s3me
as those in RSRI 10. Paficular attention should be paid to ùe requirements for
4.1
mechanical couplers, which a¡e becoming quite commonly used in Hong Kong.
4.1 .1
Prestressing tendons
The stress-st¡ain relation and other mechanical properties given in the code are
basically the same as those in BS8l10. It should be noted that rhe 0 5%o proof
stress fi is used inplace ofthe yield sûess d (whichmay nor be well defined)
to define the ultimate strength ofprestressing tendons.
3.4 Prestressing deviccs
The requirements for prestessing devices are given in ihis section in a very brief
ma¡ner because no general rules can yet be explicitly and exhaustively set. It is
better to either employ proprietary products with good perforrrance records or
employ only those new products with sufficient field trials caried out to
demonstrate their compliance with all ttre perforrnauce requirements in Section
2.1 ofthe code.
New m¡terials
New materials not covered in the code may also be used provided it can be
demonstated tlnt the basic perforrnance requirements in Section 2.1 ofthe code
are complied with and that the overall standa¡ds ofthe materials a¡e not inferior
compared to the general stand.ards stipulated in the code. For such purpose, the
design engineer has to provide sufficient information, including manufacturing
data, track record ofprevious applications, testing and proposed quaiity controls,
to allow independent lhird party evaluation.
,l
å1
f
rl
j
4 1.2
il
DURABILITY AND FIRE RESISTANCE
Objecfives
Durability
The purpose of duability design is to snsu¡e that the structure will perform its
intended firnctions satisfactorily for a sufficiently long period oftime or at least
during its design working life without requiring excessive maintenance.
Durability design requires consideration right at the begiruring of the design
stage, although some engineers tend to leave this important issue to the stage of
writing up the specification.
Duability design is not just a material problem. Crack contol is at least as
important. No matter how good the materials are, if the structure cracks
extensively to allow direct ingress of the aggressive ingredients into dre core of
thestucture,thestn¡ctu¡ewillnotlastlong. Therea¡etwolypesofcracks: non-
stuctural cracks and structur¿l cracks. They could be caused by use of an
inappropriate coDcrete mix, inadequate lemperature/moisture contol during
curing, restraint against shrinkage/temperature movement and excessive tensile
st¡ains induced by the applied load. Hence, as the code says, it is dependent upon
the integration ofevery aspect ofdesign, materials and constmction
The guidelines for durability design provided in the code are applicable only to
stuctures in normal environment and with a design working life of 50 years. For
more severe environment and,/or desiga working life longer than 50 years,
additional protection measures, such as the incorporation of cathodic protection,
protective coatings a¡rd corrosion inhibitors etc, which have not been included in
the code, may be necessary.
As part ofthe durability design, it is strongly advised to cany out life-cycle cost
analysis of the structure, taking into account the initial cost of coDstruction, the
futu¡e cost ofmaintenance and the consequential loss ofutility due to disruptions
caused by the maintenance works required. Life-cycle cost analysis is not yet
coÍrmon in Hong Kong but is already well established in many olher places as a
useful lool for deciding the level ofprotection to be provided.
Fi¡e resista¡rce
The purpose of fire resistance design is to ensu¡e that every structural element
would possess an appropriate degree of resistance to flame penekation, heat
transmission and collapse during fire attack.
From literature review, it would appear that the most commoD approach to fire
engineering is to use the prescriptive method. Each country's respective code
gives requirements for minimum cover and minimum member size based on

research on the characterisdc behaviour of that particuiar country's consüuction
materials, wilh respect to concrete and rei¡forcement strengths and types of
aggregate etc. Hong Kong already has guidance on cover requirements as
published in the Code ofPractice for Fire Resisting Constuction. The guidance
is based on the local condilions,
As with the prescriptive approach, the fire engineering calculation method in each
counfy's code is based on calibrated testing specific to that country. It is
therefo¡e questionable whether the procedures and formulae specified in either
the Eurocode or BS81 10 are applicable to Hong Kong. In addition, the accuracy
ofthe data on which the calculations are based would also be called into question
as iÎ represents a notional fue development rate used fo¡ the standa¡dised testing
of malerials. Ir practice, a real fire may follow a significantly different pattern
depending on the quantity and type ofcombustible mate¡ials present. Therefore,
in the absence of any clear benefit, it is considered not appropriate to include the
calculation method in the Hong Kong Code of Practice until further Hong Kong
speciñc research has been carried out.
4.2 Requirements for durability
4.2.1 General
A reinforced concrete slructure may deteriorate because of deterioration of the
concrete itself or because of cor¡osion of the steel reinforcing ba¡s inside the
concrete. Common causes of detenoratíon of concrete include alkali-aggregate
reaction, chemical attack, fieezing and thawing action and mechanical abrasion.
In most cases, it is the corrosion ofthe steel ¡einforcing ba¡s that is more likely to
be the major problem.
The cor¡osion of steel is an elecûo-chemical reaction. When ürere exists a
difference in elec¡ical potential along the steel bar in concrete, an elecEo-
chemical cell is set up: there form anodic and cathodic regions, connected by an
electrolyte in the form of pore water in the ha¡dened cement paste. The
positively charged ferrous ions Fe* at the anode pass into the pore solution while
the negatively charged free electron e- pass tluough the steel into the cathode
where they are absorbed by the constituents of the electolyte a¡rd combine with
water and oxygen to form hyclroxyl ions (OHf . The hydroxyl ions travel tlrough
the electol;te and combine with the ferrous ious to fo¡rr fenous hydroxide,
which is convefed by fudher oxidation 1o rust. The electro-chemical reactions
involved a¡e as follows:
Anodic reactions:
Fe -+ Fe.-+2e-
Fe- + 2 (OH)- --r Fe(OH)2 (fenous hydroxide)
a Fe(OHþ +ZHzO + Oz + 4 Fe(OH)3 (ferric hydroxide)
Carhodic reactioni
4 e- + Oz+ 2H2O -+ 4 (OHf
22
i;
T-
Both water and oxygen (or air) are needed for the process to continue. The¡e is
no corrosion in dry concrete (because the¡e is no \ryater) nor in concrete fully
immersed in water (because there is little oxygen or air). The optimum relative
humidity for corrosion is 70 to 80%,
The differences in electro-chemioal potential can arise f¡om diffe¡ences in the
environment oftåe concrete, for example when a part ofit is wet and another part
is dry. A simila¡ situation can a¡ise when there is a substantial difference in the
thickness ofcover to tle stee-.
Fortunately, even wittr a conlinuous supply of oxygen and water, the steel in
concrete does not necessarily corrode. The concrete cover p¡otects the steel from
conosion not just by hindering the ingress of deleterious fluids but also by means
of passivation. Steel embedded in hydrating cement paste rapidly forms a thin
passivating layer of oxide, which strongly adheres to the underlying steel and
gives it complete protection from reaction with oxygen and water, i.e., from
corrosion. This state of the steel is known as passivation. Maintenance of
passivation is conditional on an adequately high pH or in other words high
alkalinity of the pore water in contact wilh the passivating layer. Basicall¡
concrete is alkaline because of the presence of lime, i.e. Ca(OH)2, which is
liberated as a byproduct during cement hydration. The pH ofthe pore water in
ha¡dened Pofland cement paste is initially a¡ound 12.6 to 13.5, which is high
enough to offer passivation protection to the steel.
However, there is everywhere carbon dioxide, i.e. COz, in the air. Ca¡bon
dioxide reacts with moistrue to form ca¡bonic acid, which tåen reacts with the
lime in the pore water of concrete to form calcium carbonate, a neutral product.
This is called ca¡bonation. As a result, the alkalinity of the concrete gradually
drops and once thg pH is reduced to below around 10, the passivation protection
to the steel will be gone and if there is oxygen and water, the steel will start to
corrode.
Apart from carbon dioxide, chloride ions may also destroy the protective
passivity layer on the surface of embedded steel thereby causing de-passivation.
Chlo¡ide ions activate the surface of the steel to form an anode, the passivated
surface being the cafhode. The reactions involved a¡e as follows:
Fe" +2 Cl- -+ FeClz (fenous chloride)
FeCIz +2 HzO + Fe(OÐz + 2HCI
Thus, the ciloride ion Cl- is regenerated so that the rust contains no chloride,
although ferrous chloride is formed at the inlermediate stage.
Rust has a lower density than steel, Hence, ¿s corrosion takes place, the volume
increases a¡rd since the expansion is restained by the surrounding concrete,
bursting st€sses are induced, resulting in cracking, spalling or delamination of
the conc¡ete cover. This makes it easier for aggressive agents to seep tbroughthe
concrete towards the steel, with a subsequent increase i¡ the rate of corrosion. In
other words, once corrosion starts, it will accelerate.
i. 2l
ù
tr
 't--
With the exception of mechanical damage, all the adverse influences on the
durability of concrete involve the tnnsport of fluids or ions through the concrete.
There a¡e four fluids/ions principally relevant to the durability of concrete: water,
oxygen, carbon dioxide and chloride ions. They can move through concrete in
different ways:
o permeation(flowunderpressure gradient);
c dillusìon (flow under concenhation gradient);
o sorpfion (flow under capillary suction).
As sorption is normally insignificant the major factors influencing the durability
are the permeability ard diffiisivity of the concrete.
Permeation and diff:sion are dependent on the size, porosity and connectivity of
the pores. Since the same geometric factors influence both permeation a¡d
diffirsion, permeability and diffusivþ are inter-related and it may be said that
when the permeabilþ is high the diffirsivity is also likely to be high and vice
versa. Test results have shown that they are roughly proportional to each other
and consequenfly they are often loosely ûeated as equivalent (the common
practice ofreferring 1o the permeability ofconc¡ete to carbon dioxide o¡ chloride
ions is actually wrong because it is the concent¡ation gradient, not the pressure
gradient, that drives tho carbon dioxide or chloride ions through the concrete).
However, unlike permeation, the diffirsion of gases or ions tirough concrete or
any porous medium is sensitive to the relative humidity or the degree of
saturation of the pores. The difñ¡sion of gases, such as oxygen and carbon
dioxide, through dry pores is much faster than through wet pores or tlìrough the
po¡e water. In contast, the diffusion of ions, such as chlorides and sulphates,
takes place only tbrough the po¡e waler a¡rd thus ionic diffusion is possible only
when the pores are saturated or at least partially satu¡ated.
There are several different types ofpores in concrete: gel pores, capillary pores
and ai¡ voids in the hardened cement paste arid pores in the rock aggegale, each
of different size and therefore conEibuting differently to the overall permeability/
diffirsivity ofthe concrete. Arnong these, the gel pores, because oftheir minute
size (9 nm in diameter), have basically no effect. The permeability/diffusivity is
dependent also on the connectivity ofthe pores. Since the air voids are usually
isolated (except in honeycombs), they also have little effect. Relatively, the
capillary pores, which a¡e much larger in size compared to the gel pores and are
generally inter-connected in the form of capillaries, have the grcatest effect on
perme abi lityldiffi rs ivity.
Since the capillary pores are essentially the original space occupied by the free
water in the concrete mix remaining unfilled by gel products, the amount of
capillary pores in the concrete is dependent mainly on the waler/cementitious
ratio and the degree of hydration (in tum, degree of hydration is dependent on
curing). A sufficiently low water/cementitious ratio together with good curing
will produce a concrete with a relatively low permeability/diffirsivity for high
durability. The addition of supplementary cenentitious materials (í.e, pozzolanic
materials), such as pulverized fuel ash and condensed silica fume, can also help
to reduce the perrneability/diffusivþ ofthe concrete.
On the specification ofconcrete for durability, the British Sta¡d¡¡d BS81 10: Part
l: 1997 refers extensively 1o another British Standard 855328, In this regard, it
is noteworthy that EN206 has replaced the British Standa¡d 855328, which has
now been withdrawn. However, as 8N206 generally deals with conc¡ete mixes
based on cylinder strengths - the norm in Europe, the British St¿¡rdards Institution
has produced an accompanying code of practice 858500, which interprets the
recommendations of EN206 in terms of cube stengths. Therefore, where
¡eference is made to 8N206, the accompanying Brilish Standard 8S8500 should
also be refered to.
4 2.2 Desigrfor durability
Apart from specifying the right materials to be used and appropriate concrete
covers to be provided, the other issues to be considered include:
(a) designing the structu¡e to minimize uptake of waler or exposure to moisture;
(b) designing the structu¡e to avoid non-structural cracking;
(c) designing the structue to minimize stmctural cracking;
(d) adding, where necessary, crack distribution reinfo¡cement to control c¡ack
widths;
(e) providing, if necessary, addilional protection measures such as cathodic
protection, protective coating to the concrete structure and/or protective
coating to the steel reinforcing bars etc.
Since many processes of deterioration of a reinforced concrete structure occur
only in the presence of water or moisture, it is important to design the structûe to
avoid ponding and ¡undown of water. In other words, as far as practicable, the
slructure should be desigrred to have good drainage everywhere in the structu¡e
including both extemal and intemal areas. In this regard, allowance should be
made when necessary to compensate for the effect of long-term deflection of the
structue by applying pre-cambering or by increasing the drainage fall. If
pondingirundown of water is somehow unavoidable, such as in bathroom o¡
kitchen areas, considerations should be given to the possibility of providing
waterproofing. At locations likely to remain damp for long periods of time, e.g,
at confined places where moistu¡e tends to be trapped or near drainage pipes
which may have fi¡tu¡e leakage problems, it may be prudent to provide extra
concrete cover to the steel reinforcement.
As mentioned before, non-structural cracks a¡e due to sedimentation, shrinkage
or thermal movement of the concrete, which if restained, would induce tensile
stresses large enough to cause cracking- Non-structu¡al c¡acks ca¡r appear during
the plastic, curing and long-term stages. Those appearing during the plastic stage
include plastic settlement cracks (due to extemal restraint of the sedimentation
movement of the ftesh concrete mix by the reinforcing bars and the formwork)
and plastic shrinkage cracks (due to selÊrestaint of the drying shrinkage
movement of the top surface of the fresh concrete because of rapid evaporation).
Plastic cracking can generally be avoided by improving the mix design of the
conc¡ete (basically reducing the water content and increasing the fine powder
content) so as to ¡educe sedimentation and by shielding the top surface of the
fresh concrete from direct sunshine and wind so as to prevent rapid evaporation
25
Non-stuctural cracks appearing during the curing stage üe mainly eady thermal
cracks. They are due to internal or extemal restaints against the themral
movement ofthc concrete as the tempetature ofthe concrete changes due to the
heat generated ûom the chernical reactions of the cementitious materials.
Intemal restaint (also called self-restraint) is the major cause of cracking in
massive concrete strucfl¡res while extemal rest¡aint is the major c¿use in concrete
str¡ctures cast against rigid movement restaints. If inæmal rest¡aint is the major
cause, then insulation would help to reduce thennal cracking' However, if
external restraint is the major cause, no insul¿tion should be applied as insulation
would actually aggravate the problem. Many engineers do not distinguish
between the two types of restraints and speci$ insulation 1o be applied in all
cases. This practice is wrong and is the root cause of many early thennal
cracking problems in Hong Kong. In any case, regardless of whether intemal
aûd/or external r€staints exist, internal cooling by airlwaterlliquid nitogen
would help to mitigate the early thermal cracking problem, but most contractors
do not like this because ofthe touble and cost involved, especially ifthe cooling
is not separately priced for as a bill of quantity item by itself' Apaf from
applyrng intemal cooling, the thermal cracking p'roblem may also be mitigaæd to
some extent by cmploying a concrete mix (one with a substantial portion of the
cement replaced by pulverized fi¡el ash or ground granulated blasÚrmace slag)
that would generate less heat during ouring. A detailed discussion on whether or
not insulation should be applied is available in thc following article:
Kwan A.K.H. and Ng I.Y.T., "Avoidance of early the¡mal cracking in conc¡ete
stuctues: 1o insulate or not io insulate?", Hong Kong Engineer, February' 2004'
ppl5-ló.
Non-stnrctural cracks appearing du¡ing the long-term stage include thermal
movement cr¿cks and sbrinkage movement cracks. They a¡e due to the thermal/
shrinkage movement of the conc¡ete stn¡cttllc being restrained by rigid walls o¡
supports. Such kind of cracking is a common problem for long podium structures
with several tower blocks on top supPofted by rigid core walls, which forrr parts
of the podium structtrres and restrain the thermal/sh¡inkage movement of the
podium decks. One way of alleviating the problem is Ûo provide movement
joints. If shrinkage movement is the major cause of cracking, then the provision
of late-cast stips and the addition of shrinkage reducing agent to the concrete
mix should also be considered as possible miligation measures.
On the other hand, stnrctural cracks are due to lensile stresses developed in the
concrete resulting from the applied loads, They are almost unavoidable, unless
the structr¡re is to be oompletely redesigned or prestressed, which is usually not
practicable. Nevertreless, the crack widths can be contolled by limiting the
tensile stesses developed in the concrete and by proper reinforcement detailing.
Regarding the addition of crack distibution reinfo¡cement for conholling crack
widths, it should be noted that it does not stop cracking ofthe concrete; it only
distibutes the c¡acks so that there will be more cracks formed each wilh a
smaller crack width.
For the additional protection measures, which have not been covered in the code,
specialist lilerature should be consulæd.
26
4.2.3 Exposureconditions
The exposure conditions are classified in the new code as:
Exposure condition I - mild (based on environnent)
Exposure condition 2 - moderate (based on environment)
Exposure condition 3 - severe (based on environment)
Exposure condition 4 - very severe (based on environment)
Exposwe condition 5 - abrasive (based on mechanical wear and tear)
Tbe above classification sptem is based on the guidance given in BS81l0 and
8M06. However, the principles of classification have been modified to reflect
the local conditions (e.g. freezeltbaw exposure and de-icing salts exposurc are not
of concem in Hong Kong whereas conditions of high humidity and marine
environment are). Morcover, the numbe¡ of exposure conditions based on
environment has been reduced for sirrplification. In BSBll0, there a¡e five
exposure conditions classiñed according to the environmental conditions: mild,
modcrate, severe, very severe and extreme, but in the new code, there are only
four exposure conditions classified according to the envi¡onmental conditions:
mild, moderate, severe and very severe. Therefore, the mild, moderate, severe
and very severc exposure conditions in fhe new code are not quite the same as the
corresponding exposwe conditions in 858110. For example, the mild condition
(exposure condition 1) in the new code covers slighfly more lhan just the mild
condition in BSEI l0 or the condition XCl in EN206, as the exposure condition 1
in the new codc applies also to concrete su¡f¡ces inside buildings protected ûom
the effects of condensation or humidþ (a normal condition of Hong Kong
buildings), which is simila¡ to the condition XC2 in EN206. Consequently, the
durability requircrnents oft[e respective exposure conditions in the new code a¡e
slightly dífferent fiom those in BSE I 1 0.
4.2.4 Cover
fi
The conc¡ete cover to the steel reinforcement serves the following functions:
û
r . to protectthe steel against corrosion;
ú
. to protectthe steel against fire;
. to p¡ovide suffrcient embedment for safe tansmission ofbond forces;
fi
¡ to allow the coa¡se aggregate particles ofthe concrete mix to pass through
the gap between the nould and the steel reinforcement;
t r to allow for unevenness ofthe surfaces ürat the conc¡ete will be cast against
|.Ì
and fulure changes in dimensions due to su¡face teatments such as bush
hammering.
Because ofthe multi-fimctions ofthe concrete covet, corrosion protection is only
one ofthe considerations in specifying the required concrete cover.
As in BS8110, the nominal cover concept is used. Nominal cover is the design
depth of concrete cover to all steel reinforcement, including links. It is the
dinension used in design and indicated on the drawings. The actual cover to all
reinforcement should not be less than the nomi¡al cover minus 5 mm' Such an
approach is adopted to ensure tlat during construction, an adequate mi¡imum
cover lo reinforcement, which is paramot¡nt for corrosion protection, can be
achieved. The nomi¡al cover specified in the new code is therefore not the
minimum cover as stipulated in Building (Construction) Regulations o¡ Code of
Practice for Sfr¡ctural Use of Concrete: 198'1. There is a 5 mm difference
between nominal cover and rninimum cover.
The required nominal covers for corrosion protection under dífferent exposr¡re
condilions are specified in Table 4.2 of the code. Along with the required
nominal cover, the lowest grade of conc¡ete, the maximum free water/cement
ratio and the minimum cement content are also givea in tle table. It should be
noted that the nominal cover, lowest grade of concrete, maximum water/cement
¡atio and minimum cement content together form a "package". They all have to
be specified and checked for compliance at the same time.
Taking into account the slight diffe¡ence in the classification of exposwe
conditions, lhe nominal cover, maximum water/cement ratio and minimum
cement content etc speoified in Table 4.2 ofthe code a¡e generally in linc with the
¡ecommendations given in EN206, 858500 (Îhe British Standard that
accompanies EN206 and adapts it for specification in terms of cube strengfhs)
and the report "Speci$ing concrete to BSEN206-I/8S8500" published by the
UK Quarry Products Association.
In Hong Kong, concrete suppliers tend to use higher cement content and lower
wate¡/cement ratio than required, Consequently, the concrete sEength achieved is
usually higher than the specified strength. Given the relatively small difference
in cost, such an approach is prudent to ensr¡re the provision ofan adequate safety
margin to the concrete strength. As a result, a higher concrete strength is usually
available that can be beneficial for protecting the steel reinforcement against
corrosion.
The addition of supplementary cementitious materials (also oalled mineral
admixtures, such as pulverized fuel asb, ground granulated blastfirmace slag and
condensed silica fr¡me etc) to concrete mixes is becoming increasingly popular.
Addition of these supplementary cementitious m¿terials can improve the overall
perfonnance ofthe concrete and should be encouraged. Although not explicitly
stated, any such supplementary cementitious materials added may be counted
towards the cement content fo¡ sadsrying the maximr¡m water/cement ratio and
minimum cement content requirements stipulated in Table 4.2. In actual fact, it
has been stated in Clause 4.2.6.3 of the code that when pulverized fuel ash is
used, the total content of cement plus pulverized fuel ash should be at least as
great as the values given in Tables 4.2 and, 4.4 and tbat in these tables, the word
"cement" in "cement contenf' and "water/cement ratio" means the total content
of cement plus pulverized fuel ash, or in other words, the total cementitious
materials content. This point should be notpd when checking compliance with
Tables 4.2 and 4,4.
The addilion ofplasticizer or superplasticizer to concrete mixes is also becoming
quite popular. Additiou of these admixtu¡es can improve the workability of the
concrete mix and allow the wate¡/cement ratio or cement content of the concrete
mix to be reduced for the same workability requirement. The minimum cement
28
content requirement is to ensure proper consolidatíon of the concrete mix.
However, with the addition of plasticizer or superplasticizer, a good workability
can be achieved at relatively low cement content. Hence, the minimum oement
content should be dependent on whether plasticizer or superplasticizer has been
added. Pending more research studies to provide experimental suppor! the
minimum workabilþ of the coDcrete mix should be specified instead.
Although the exposure condition 4 (very severe environment¿I condition) in lhe
new code embraces also marine environment, for the specification of marine
concrete (i.e. concrete fo¡ use in marine environment), the Recommended
Specification for Reinforced Cotrcrete in Ma¡ine Environorent in the Port Works
Design Manual: Pa¡t 1: 2002 may also be considered. In this recommended
specification, the requirements for nominal cover, lowest grade of conc¡ets,
maximum \À,ater to total cementitíous mate¡ials ratio and cemenlitious materials
content are 75 mm, C45, 0.38 and 380-450 kg/m3 respectively. Except for the
required lowest grade of concrete, these requirements are similar to or slightìy
more stringent than the respective requirements in the new code. In fact, based
on the author's own experience, the restriction of the water to total cementitious
materials ratio at not higher than 0.38 would lead to a concrete grade ofat least
C50, which is actually in line with the new code.
4.2.5 Concrete materials and mixes
Unde¡ cert¿in specific conditions, the code allows the lowest conc¡ete grade to be
reduced by not more than 5 provided the maximr¡n w¿te¡/cement ratio and
minimum cement content requirements are proven to have been met and allorvs
tlre minimum cement cont€nt to be reduced by not more lhat l0% provided the
corresponding wate¡/cement ratio is ¡educed by not Iess than the percentage
reduction in the cement cont€nt. Tighter quality cont¡ol and a more systematic
checking regime are required when these reductions a¡e to be applied. These
allowable reductions are derived mainly ûom BS81l0. In Hong Kong, there
should be no parlicular difficulties in achieving the required strengûr grade and
minimun cemsnt cont€nt. Hence, in practice, there is seldom the necessity to
make use ofthese allowable reductions.
'When
the maximum aggl.eg le size is smalle¡ than or larger than 20 mm, the code
requires the minimum cement content to be adjusted upwards or downwa¡ds,
respectively. As mentioned before, this is to ensure proper consolidation ofthe
concÌÊte mix. However, since the workability of the concrete mix can always be
improved by adding plasticizer/superplasticizer, there is no longer the necessity
to impose any minimum cement cotrtent requirement. Pending more research
studies to provide experimental support, the minimum workability of the concrete
mix should be specified instead. Such perfonnance speciñcation would allow
more flexibility on the side ofthe concrete p¡oducer and easier control on the side
ofthe engiueer responsible for sile supewision.
In the new code, the addition ofpulverized fueI ash (PFA) up to 35% ofthe total
cementitious materials content is allowed. Although thp P¡actice Note PNAP9O
limits the amount of PFA ta 25%, a number of specifications written in Hong
29
 i
-?1
Kong have higher permissible levels of PFA. Both the Housing Departrnent and
MTRC specifications have been allowing the addition of PFA up to 35% fot quite
some time. As concretes produced in accordance with these specifications have
been performing satisfactorily, it is considered appropriate to increase the
permissible level of PFA 1o 35%.
On one hand, the addition of PFA can reduce the pemreability/diffusivity of the
concrete by converting the soluble lime (a by-product of the chemical reection
between cement and water) in the concrete to further gel products. On the other
ha¡d, it may also reduce the alkalinity of the concrete, which is needed for
maintaining the passivation p,rotection to the steel ¡einforcement, Ovemll, the
durability performancc of concrete made with PFA is about the same as that of
concrete made wilhout PFA, provided the concr€te with PFA is of the same
strength grade as úe concrete wilhout PFA. However, concrete wilh PFA added
generally develops stength at a slowe¡ rate and requires a longer period of
curing. More attention to the curing and ¡emoval of forrrwork is required with
conc¡ete mixes containing PFA.
4.2.6 Mix proportions
As a basic principle, the total cementitious content is limitcd to not more than 550
kg/m' in orde¡ to maintain a high dinensional stability (i.e. small changes in
dimension due to early temperatue rise during curing, drying shrinkage at the
long-lerm stage and creep under sustained loading).
For normal concretc, i.e. f- s 60 N/mm2, which is relatively easy to produce
without the u5s 6f ¿ high total cementitious contcnt, the above maximum limit of
550 kg/m3 should be suictly applied and should not be exceeded unless tùere is a
proven need and special considerations have been given to the risks of early
thermal and drying shrinkage cracking and to the possible adverse effects of
cfEep.
For high-stenglh conc¡ete, i.e, f- > 60 N/mm2, which is generally more diffrcult
to produce and requires the
imposition of the above ive.
Nevertheless, if thc lotal cial
considerations should be given to the risks ofearþ fhermal and drying shrinkage
cracking ard to the possible adverse effects of creep. Moreover, under normal
circumstances, the ccment content (note:
^not
tbe total cementitious content)
should be limiæd to not more than 450 kg/m'.
Regardless of whether the total cementitious content of the concret€ mix is on the
it
high side or not, is recommended that for massive concretÊ structures'
temperah:re rise evaluation tcst should be ca¡ried out to determine the
tÊmperature rise of tlrc concrete, for long concret€ stn¡ctures, drying shrinkage
test should be canied out to detcmrine the ultimate shrinkage stain of the
concrcte, and for concrete struÊtures likely to be affccted by creep effects, creep
test should be ca¡ried out to determine the creep coefücient ofthe concrete.
4.2.7 l:Ùlix constituents
Since the presence of chloride in concrete may cause de-passivation of the steel
reinforcement and may affect the sulphate resistance of the concrete, it is
necessary to control the chloride content in the conc¡ete mix. No admixtu¡e
containing any chloride should be used. The total chloride content in the
concrete mix, expressed as a percentage of chloride ion by mass of the total
cementitious materials content (including both cement and PFA), should not
exceed 0.1, 0.2 and 0.35 for prestressed concrete or steam-cu¡ed concrete,
cotrcrete made witl¡ sulphate rcsisting cemen! and concrete containing steel or
other mctals, respectively. These values are slightly more stingent than those
I
permitted.in BS81l0: Part (same as those in 8S5328) and EN206 and a¡e
specifically developed to cater for the aggressivc environment in Hong Kong.
Though not common, delete¡ious chemical re¿ctions between the aggregate
particles and the surrounding hardened cement paste, leading to disruption and
cracking of the concrete, have been observed in Hong Kong. The most common
chemical ¡eaction is that between tle active silica constituents of the aggegatc
and the alkalis in the cement, Such reaction is called alkalïsilica reaction.
Another type of deleterious aggr€gate reaction is that between some dolomite
limestone aggregates and the alkalis in the cement' This is called ¿Ifrøli
carbonate reaclion. Any aggregare that may be susceptible to alkali-silica
reaction or alkali-carbonate reaction (collectively known as alkallaggregate
rcaction) should not be used. The úsk of alkali-aggregate reaction may be
reduced by:
. limiting the amount of alkalis in the concrete mix, and
o adding pozzolanic materirls such as pulverized fuel ash a¡d condensed silica
ñ¡me, which react with the lime in the concretc and thereby consume part of
the alkalis.
The guidelines given in the code are based on the Practice Note PNAPl80.
4.3 Reqnirements for fire reslst¡nce
Tbe Code ofPractice for Fire Resisting Constnrction is refer¡ed to regarding the
nominal cove¡ and minimum dimensions of members required for fue resistance.
Because of is low permeability to r¡/atel and vapour, a high-strength concrete
cover
mix.
demo
st¡uclufes.
 r--

In SLS design, a linear elastic analysis would normally suffice. Three methods of (b) minimum a:cial load combined with coexistent bending moment;
evaluating the soction stiffness ofthe members are given in the code. They are G) maximum bending moment combined with coexistent axial load; and
all regarded as acceptable but a consistent method should be applied to all (d) any other coexistent combinations of axial load and benrling moment which
members of the structwe. This is because if the method used to evaluate the will be more critical than the above cases.
section stiffness tends to overestimate the section stiffness, it would overestimate
the section stiffuess of all members so that the relative stifüress of the members
would remain more or less the same. In the analysis of structure fo¡ intemal 5.1.4 Imperfections and second o¡der effects
force distibution, it is the ¡elative stiffiress of the members tlat matters, not the
absolute values of the section stiffness of the members. However, if the analysis The second order effects are not expected to be signiÍcant at the SLS. Hence,
of structure is to determine the maximu¡n deflection, then the¡e will be a small these only need to be considered at the llLS. The requirement and method of
difference in the deflection results when different methods a¡e used to evaluate incorporating_.¡lhe second order e in the analysis ofstructure are stipulated in
the section stiffness. Section 5.fofthe code and will be explained later in due course.
(,w
5.1.3 Load cases and combinations 5.2 Analysis of structure

The load combinations to be considered in the design ofbeams and slabs and in
the design of columns and walls have been given in the code for the relatively
simple cases only,
Fo¡ more complicated cases, such as beams subjected to momenl, shea¡ and
torsion, colums subjected 1o axial load and biaxial bending, core walls subjected
to axial load, biaxial bending and torsion, and ûames with bracing members etc,
no general guidance can be given. The design engineer will have to exercise
his/her own judgement on the load combinations to be considered in the design so
that no critical loading case is omitted.
It should be noted that the load combinations for the design of beams and slabs
specified in the new code are not the same as those given in BS81 l0 or Eurocode
2. In BSSl 10, the load combinations 1o be considered are:
(a) all spans loaded with the maximum design load; and
(b) alternate spans loaded with the maximum design load and all other spars
Ioaded with the minimum design load.
In Ewocode 2, the load combinations to be considered are:
(a) alternate spans loaded with the maximum design load and all other spans
loaded \Mith the minimum design load; and
5.2.1 Idealisation of the structue
Rules for the classification of structural elements into the following types have
been provided:
o beams/deep beams (based on CIRIA Guide No.2);
r slabs/one way slabs (same as in Ewocode 2);
r ¡iþþsd^vaffle slabs nottreated as discrete elements (same as in Eurocode 2);
. coh¡mns/walls (same as in Eurocode 2).
However, there a¡e no rules forthe classification of strucfural elements into core
walls, coupling beams, arches and shells etc.
Apart from the above, the provisions given in Clause 5.2.1 of the code are the
same as those given in Clause 5.3.1 of Eurocode 2.
Fo¡ T and L beams, a formula for evaluating the effective flange width årff over
which unifomr conditions ofstess can be assumed. has been orovided as follows:
2

(b) any two adjacent spans loaded with the maximum design load and all othe¡ t.n=lb*,,+n*
i=l
, spans loaded wilÌr tÏe minimum design load.
_ - - In the new code, the load combinations to be considered are:
IWYVVWVv !ßVl;
¡:ri:rn@)
^ .-
all spans loaded with the maximurn design load; where ö"o., :
0,2bi + 0.1l, < 0.21Þ and ö.5.¡ < å,. The effective flange width is
,,
,) O) alternate spans loaded with the maximum design load and all other spans to allow for shea¡ lag in the flange, which causes decreasing bending stress away
r r':i^'!b\ : l-$i:+
l\r Ioaded with the minimum design load; and from the web. Note that the effective flange width D.n is dependent on the type
,i, .. ,r r.,, (c) any two adjacent spans loaded with the maximum design load and all other of loading, span length and suppof conditions because the distance between
sÀ"Ë-2.'r*:-"+ spans loaded with the minimum design load. points of zero moment /. is dependent on these parameters.

Particular attention should also be paid to the load combinations forltre design of
columns a¡d walls specified in the new code. In both BSSi 10 and Eurocode 2,
A formula for evaluating the effective span / of a beam o¡ slab has also been
given in the code as follows:
there is no recommendation on the load combinations to be conside¡ed, but in the
new code, the following load combinations are recommended:
I : I^ t a,'r a, Z.(V.n v<,/ l{a.//\
(a) maximum axial load combined with coexistent bending momenl;
 0-= t* + k +k> fr,=
'
.
lt
ht¡vr i'Li=
'/
Z!
,,.C^ø\F¿
Y\'
)t
Z,
where /, is the clear distance between the faces ofthe supports, and. a, and a, in which ¡ia.*o is the design suppoÍ reaction The value oî ÀMuo is estimated
are each equal to nio.(h|2, S*/2) zt the respective end of the span. However, it by assuming that ¡;d,*e is uniformly distributed across the width .s* of the
should be noted that drese effective spans are provided mainly for membe¡
support, as illustrated in Figure 5.2.
analysis (i.e. for analysing the bending moment and shear force along the beam or
slab when subjected to vertical load only). For ûame analysis, the code staGs
that these assumed effective spans should be used only where appropriate. It is
suggested herein that for frame analysis, the effective span of a beam o¡ slab
spanning over two adjacent suppofs should be taken as the distance between the
centrelines ofthe two adjacent supports,

For any support, unless rotational ¡estraint has been properly provided (and
structurally designed and detailed as such), the support should be considered as
not providing any rotational ¡estraint and t¡eated as a hinge support.

Where a beam or slab is monolithic with its suppofts to provide rotational

restraints, the critical design moment at the support may be taken as that at the
Supporting element
face ofa rectangular suppo¡t, or at 0.2 / inside the face ofa ci¡cular support of
freated as a hinge support
diameter /, but should not be taken as less than 0.65 ofthe full fixed end
moment, as shown in Figure 5.1.

r-t\ l,ln/e,r -i\

\titlv,Á l' \
-z/:s'
Full fìxed
end moment 7;
Reaction assumed to be
uniformlv distributed

Note: Critical design moment Frgtrre 5.2 Reduçtion of design suppoÍ moment
should not be taken as less tha¡r
0.65 of ñ¡ll fixed end moment

5 .2.2 Analysis of sections fo¡ ultimate limit states

Figure 5.1 Critical design moment ât a monolithic suppoft
For analysis of sections at ULS, the "plane sectìons remain plane" assumption
may be adopted. Based on this assumption, the bending strain (ofboth the steel
and concrete) within a section va¡ies linearly across the section (note however
. Where a beam or slab is spanning continuously over a support, which may be that the sÎress-strain relationships are not linear). Basically, the bending strain of
X-/ti---.'ù-Tcorsidered as not providing any rotational restraint, the desígn support moment, a longitudinal fibre (whether steel or concrete) is directly proportional to the
calculated on the basis of a span equal to the cenl¡e-to-centre distance between distance ofthe fibre from the neutral axis
supports, may be reduced by an amount as follows:
^À,IEó
The stength of a section under both short- and long-term loading may be
assessed using the stress-strain curves derived from the design strengfhs of the
materials as given in Sections 3 1, 3 2 and 3 3 ofthe code'

37
 r-
5.2.1 Anaþsis of sections for sewiceability limit states

For analysis of sections at SLS, the "plane sections remain plane" assumption
may be adopted, In addition, it may be assumed that both the steel and concrete
remain elastic so that their stress-strain relationships a¡e linea¡. Since the
þs¡ding stain ofa longitudinal fibre is directly proportional to the distance ofthe
ñbre from the neutral æ<is and the stess-strain relationships are linear, the
bending st¡ess within a section should vary linearly across the section arid should
be directly proportional 1o the distance from the neutal axis.

5.2.4

\,
'/,â\ 5.2.5 iding lateral restraint
íÐ
"*WVt(, (,*l/^tÅt
'" '-'-'0 ' providinglateralrestrainlonlytheverticalloadsneed NOP
be analysed as a series of suÞframes, each for the Monolithic frame Sub-frames

dùa-<rrJl^þ
-ro " P*ryì" of analYsing
il carriedoutonalevel-
- o,, level being conside¡e
\--f_J*ì
ì1 n7 ä shown in Figure 5.3. Figure 5.3 Simplification into subframes for analysis on level-by-level basis
"" generally be assumed to be fxed unless a pinned end is clearly more reasonable.

As an alternative, a monolithic ftame not providing lateral restaint may also be

i
fì
ri
analysed as a series of simplified sub-frames, each fo¡ the purpose of analysing :,:

an individual be"m only, so that tbe analysis can be ca¡ried out on a beam-by- q
beam basis, Each sub-frame consists ofthe beam being considered, the columns
ir
attached to the ends of the beam and the olher beams on either side, if any, as È
shown in Figure 5.4. Each end of the columns and beams remote ûom the beam
being considered may generally be assumcd to be fxed unless a pinned end is
clearly more rcasonable. The stiffness of the other beams on eilher side of the
beam being considered should be taken as half their actual values if they are
taken to be fixed at thei¡ oute¡ ends.

Øø'Y Þw\ As anothe¡ altemative, a monolithic frame not providing latcral restraint may also
,---ã- ,{----- -
A ' ba öl æ be analysed as a series of continuous beams, each for the purpose of analysing tbe
beams at one level only, so that the analysis can be caried out on a level-bylevel
basis. Each continuous beam consists of the beams at the level being considered
spanning over supports not providing any rotational reshaint, i.e. hinge supports,
as shown ín Figure 5.5. Where the bearns at one level bave been analysed as a
V
WWutir,ut,.;
rL",r\rtV,ñU Monolithic frame

actual stiftiess. It should be noted that the critical load combinations for the
design of columns æe in general different from those fo¡ the desigr ofbeams. Figure 5.4 Simplification inlo sub-fiames for analysis on beam-by-beam basis

38 39
-

Continuous beam

Sub-frames for determining moments

in column GK at joint G and Frame subjected to Sway-frame for
Monolithic frame joint K, respectively horizontal loads only analysing the effects
-" r,-\r. ofhorizontal loads
(wa/
"-: Figure 5.6 Simplification into a sway-fràme
-]-. " , ,1^ À_, Figure
'i t*
5.5 Simplification into contínuous
i , ^ n
beams and sub-frames

itù\+'"i. bÞ4u)<-4_17 UTr*.ç.) A.^)._

5,2.6 Frames providing lateral stability
For a monolithic frame providing lateral stability to the structure as a whole (or in
other words, helping to resist the lateral loads), the horizontal loads acting on the
ûame and the resulting sway need also to be considered. In addition, if the
columns of the frame are slender, the addition¡l moments arising from fhe P-Á
effect should be imposed.
In a tall building, one o! more shear/core walls should h¿ve been incorporated to
help resist the lateral loads and to provide partitions and enclosures for the
utilities, lifrs and staircases etc. With the co-existence of shear/core walls and
frames, complex wall-ûame interaction \ilould occur and it is trot easy 1o directly
determine the amount of horizontal loads that could act on the frame part of the
building. Wall-frame interaction occurs mainly because the deflection mode of a
wall and the deflection mode of a fr¿me are very different. For example, when a

The analysis of the frame may be simplified by considering the effects of the
W uniform dvall) is subiected
lncfease wltlr helsht as a c
vertical loads and the effects of the horizontal loads separately, and then subjected to a hõrizontal
superimposing the effccts of the vertical loads and the effects of the hoúzontal ràcI-
loads together to obtain the individual member forces. For lhe analysis of the dWw
tame subjected to vertical loads onl¡ tho same procedures as fot a monolithic
frame not providing lateral restraint may be applied (e.g. simplifring the frame
be transferred mosdy to the frame part solþat thewall part would be subjected to
into sub-frames or continuous beams). For the analysis of tle frame subjected 1o
horizontal loads only, the frame may be simplified into a sway-frame by littlc slorey shea¡ whereas towa¡d thêJbottom, the storey shear would be
t¡ansfened mostly to rhe wall part so that the tame part would be subjected to
incorporating hinges at the points of conûaflexure of all the beam and column
members (the point of contaflexure of a beam or column may be assumed to be
little storey shear. For more detailed explanations and in-depth analysis, the
toç91-e"! grrhr.senfrr_ofJbe-rn9rnÞer if the membe¡ is rigidly connected at both
fotlowing book is ¡ecommended (the whole of Chapter I 1 of the book is on this
topic):
ends to the other pafs of th
Stafford Smith B. and Coull 4., Tall Building Structures: Analysis and Design,
"qr_gr,r-rþ-e_L:s_-4-Crglnd*fl _o_9lg! John rüiley & Sons, Inc., 1991, 537pp,
5.ó,
 Because of wall-frame interaction, the horizontal loads are fi)istributed to the In addition to the above, for members in a ûame providing lateral stability, the
wall and frame parts of the building in direct proportion toleir relative lateral redistribution of moments is limited to l0% (i.e. Pb > 90%), and the design
stiffness. There are, in fact, no simple melhods for analysing the wall-ftame ullimate resistance moment at any sectíon should be at least 909io of the
interaction and the distribution of horizontal loads to the various pafs of the maximum moment of the section before redistribution.
building structwe. The only realistic way of analysing the lateral behaviour of a
composite wall-f¡ame structure is by means of computer analysis using either the
fiame analogy or the finite element method. 5.3 Second order effects with axial loads
.r-
À¡øry'W.nu'sä"n"", although the code allows the analysis of a frame Second order effects should be taken into accouût whe¡ever the lateral deflection
-
, Q -- $oads onty to be canied out by simplifying the frame of the stucture could affect the load distribution and equilibrium of any part of
UM^d- fvÚwa')author does not recommend this overly simplified method the structure. They are more likely to occu¡ in slender elements subjected to axial
; there is any wall in the building. Nevertheless, the practice of considering the loads such as columns, walls and piles subjected to vertical loads, arches
' " -. effects of the vertical loads and the effects of the horizontal loads separately, and subjected to both axial compression and bending moment, and plates Glates
. Ì i. ,- | ...,, .{ " 1¡s¡¡ adding the effects of the vertical loads and the effects of the horizontal loads include slabs) and shells subjected to both in-plane compression and out-of-plane
, i' '. I *i rrrl.J together to obtain the individual member forces is still applicable, despite the bending.
'r complexity due to wall-ûame interaction.
.) "r,
. : t,,, -
To take into account ûre second order effects, the deformed state ofthe súucture
should be used for the analysis ofstructure. Such analysis with the deformation
5-2.7 Slabs of the structure taken into account may be carried out in an iterative man¡er. At
the beginning, the deformation of the structwe may be neglected and the analysis
Provisions for the analysis ofslabs a¡e given in Chapter 6 ofthe code and will be of structure ca¡ried out with all the likely imperfections (such as initial cuwature
explained later in due course. and eccenticity) incorporated to determíne an initial estimate of the deformation
of the structu¡e. Then, using the i¡ritial estimate of the deformation of the
slructure so obtained, the analysis of stuclure is ca¡ried out again for a better
5.2.8 Corbels and nibs estimate of the deformation of the stucture. If the better estimate of the
deformation ofthe structu¡e tums out to be signiûcantly larger than the previous
Provisions for the analysis ofcorbels and nibs are given in Chaper 6 ofthe code estimate, ûre analysis of structrue is repeated each time with the deformation of
and will be explained later in due course' stucture updated until the deformation results converge to constant values.

It should be noted that after incorporating the second order effects, columns
5.2.9 Redistribution of moments originally designed fo¡ axial compression and uniaxial bending only might
become subjected to axial compression and biaxial bending and walls originally
similar to those designed for in-plane loads only might become subjected also to oul-of-plane
concrete grade bending.
set and that the
a smaller value
when a higher stength concrete is used has been incorporated. 5.4 Shear w¡Ils

Fo¡ redistribution of moments to be applied 1o an intemal moment dist¡ibution Shear walls are defined as plain or reinforced concrete walls conbibuting to the
obtained by elastic analysis, the foltowing conditions must be satisfied: lateral stability ofthe structu¡e. Care should be taken to differentiate shea¡ walls
Condition I Equilibrium between internal and external forces is maintained' from infill panels in frames. Both shear walls and infiIl panels can be plain or
Condition 2. Where the design ultimate resislance moment has been reduced, reinforced and thus they crrnot be differentiated by just looking at whether the
the section subjected to largest moment meets the following: wall panels are plain or ¡einforced. Iafill panels a¡e those wall panels in frarnes,
xld < (P6 - 0.4) for f" < 40 N/mm2 which serve as removable partitions and thus ca¡not be treated as contributing to
xld<(po -0.5) for40N/mmz< f" s70N/mm2 the lateral stability ofthe structure. Although, theoretically, infill panels can be
redi{ribution structurally connected to the fiames to act as structural elements contibuting to
in which'" - -momenlater lateral stabilig, since tenants in Hong Kong iike to repartition the floor areas by
monent before redistribr:t ion
knocking down the infill panels, it is better not to rely on infill panels to provide
condition 3. A1 any section, the design ultimate resistance moment is at leasl
lateral stability.
70% of maximum moment before ¡edistribution
- -
The lateral loads acting on each shear/core wall should be obtained ûom a global 6 ULTIMATE LIMIT STATES
analysis of the structu¡e, taking into account wall-frame interaction, coupling
effects and eccentricities of the applied loads. The term "shear cent¡e" has been
24)." .; ., ,.. è\r'r'\/
*^
refer¡ed to in the code, but in realþ, due to warping restrai¡ts and coupling 6.1
't>
Agu,.w )t ::'l
tutn^.9 ly on 6.1.1
no need to determine the position of the shear centre any more.
As mentioned several times before, although shear walls are generally designed
to ca¡ry in-plane loads oniy, they rray be subjected to oul-of-plane loads due to
the application of accidental loading (a requirement for robustness design) and
probably also due to the second order effects arising from oul-of-plane deflection
ofthe walls and the axial loads actins on the walls.
6.12
5.5 Transferstructures
Transfer structures a¡e horizontal elements, whioh tansfer vertical loads from
ve¡tical elements above laterally to other vertical elements below at diffe¡ent
positions. They are Decessa¡y wherever there are discontinuities in the vertical
elements. ln the process oftransfening the vertical loads laterally, the ta¡sfer
structures are themselves subjected to very large vefical shear loads and thus
have to be quite deep orlhick.
Because of the læge stiffrress, a transfe¡ structwe is quite sensitive to the
differential axial shortening of the vertical elements supporling it. Hence, the
vertical loads should be tansfened ttuough the transfer stucture 1o the vertical
elements below as uaiformly distributed as possible. 'Where the interaction
between the tansfer stucture and the vertical elements below could be
significant, the transfer structure should be analysed together with the vertical
elements below as an integral structue. If there a¡e shea¡ walls above the
transfer structüe, the local effects ofthe shear walls should also be considered.
When horizontal loads are t¡ansfened through the tansfer structu¡e, the vertical
elements above and below would be subjected to horizontal shear loads and
would therefore deflect laterally causing the development of large concentrated
moments at the joints between the vertical elements and the tansfer stucnre.
Such concentrated moments should be considered in the design. Moreover, the
possible second order effects on the vertical elements due to sidesway of the
tansfer structure should also be considered.
5.6 Precast elements
There is no additional guidance given in the code for the design of precast
elements. Basically, prccast elements should be ilesigned to the sane standard as
the in-situ cast elements. Ifthe construction sequence and schedule could have
significant effects on the behaviour of the stuctüe, such effects should be firlly
allowed for in the analvsis of structure.
44
Members in flexure
General
This section deals with the design of beams and various types of slabs. The
general requírements for the design ofbeams are given in Clause 6.1.2, These
requirements are applicable also to slabs. The additional requirements for slabs
are given in Clauses 6.1.f, 6.1.4 and 6.1.5.
Beams
The provisions given in the code are fo¡ the design of beams of nonnal
proportions only. Deep beams (spar/overall depth ra.lio <2 for simply-supported
beams and span/overall depth ratio 3 2.5 fot continuous beams) have been
excluded entirely. For the design of deep beams, specialist literature should be
consulted.
To maintain lateral stability (i.e. to avoid lateral buckling of the compression
zone of the beam), the following slenderness limits need to be imposed:
o for simply-supported a¡d continuous beams, the clear distance between
lateral restraints should not exceed the minimum of (ó0 b" or 250 b"z / d );
. for cantilever beams with lateral ¡esl¡aint only at support, the distance
between the lateral restraint and the free end should not exceed the
minimum of QS b, or 100b.2 t d);
in which à" is the breadth ofthe compression zone and d is the effective depth.
eiven in the following'
Assumptions:
o The strain disfibùtion may be evaluated based on the "plane sections remain
plane after bending" assumption.
. The stresses in the concrete in compression may be derived using the stress-
stain cr¡rve given in Section 3.1 of the code wirh ¡, : 1.5. Àltematively,
the simplified stress block given in Figure 6.1 may be used.
¡ The tensile strenglh ofconcrete is negligible.
. The stresses in the steel ¡einforcement may be derived using the stress-strain
curve given in Section 3.2 ofthe code with y^ : 1.15.
. The lever arm is not to be taken as greater than 0.95 times the effective
depth.
. The effect of any axial load may be ignored if it does not exceed 0.1 f-
times the cross-sectional a¡ea.
 -
BS8l10: Part I provides a limit to the neuûal axis depth to ensure a minimum
level of flexural ductility (i.e. that large sùains are developed in the tension
reinforcement at ULS). For redistribution of moments up to 1002, the limit given
, âu-l ffi
T^
is xld < 0,5 . For redistribution of moments greater than l0% (i.e. Bo < 90Vo)' a
lower limit of xld < (þo - 0.a) is applied to ensure that sufhciently large strains
are developed in the tension reinforcement to provide the plastic hinges wilh
enough rotational capacities for meeting the plastic deformation demand arising
from the redistibution of moments. These limits in BS8ll0 are, however,
applicable only 1o relaÍively low strength concrete. For ooncrete of strength
higher than 40 N/mmr, which possesses progressively lower ductility as the
Neutral
concrete sfienglh incteases, in order to maintain a consistent level of flexu¡al
axts
ductitity, the tension to balanced steel ratio has to be lowered. Hence, the
pemrissible neutral æ<is depth has to be reduced accordingly. The University of
Hong Kong has canied out some ¡esearch on this a¡d its recomrnendations for
higher strength concrete a¡e those formulas given above for redistribution of
moments up to l0%. For redistibution of moments greater than 10%' the
formulas a¡e modified 1o incorporate p, in a way similar to that of 858110.
Section SΡess

More background information on the derivation ofthese forrnulas can be found in

Note: e,u = 0.0035 for l" <60N the following paper:
Ho J.C.M., Kwan A.K.H. a¡rd Pam H.J., "Minimum flexural ductilþ design of
e"" = 0.0035 - 0.00006 vt[1" 60-) for f," > 60 Nimm2
high-strength concrete beams", Magazine of Concrete Research, Vol.56, No.1,
2004,pp13-22.

Figure 6.1 Simplified st¡ess block for concrete at ULS Design formulas for rectangular beam s :
Two dimensionless factors, namely: ,< and K', are employed in the formulas.
They are defined respectively as:
In 858110: Part l, the ultimate concrete strain e"" for the flexural design of
normal strength concrete is taken as a oonstant value of 0 0035. Research by x = 4_:and .
both the UK Concrete Society and The University of Hong Kong has shown that bd, f^
the ductility of concrete decreases with higher concrete strength. Hence, the M'
ultimate concrete strain needs to be reduced as the concrete stength increases.
The limitation on strain in Figure 6.1 is based on the research by The University
^ --bd, fo,
ofHong Kong, as published in the following paper: in which M is the required moment of¡esistance as calculated from the design
Ho J.C.M., Kwan A.K.H and Pam H,J., "Ultimate concrete st¡ain a¡rd equivalent loads and M' is the maximum achievable moment of resistance of the section
rectangular sfess block for design of high-stength concrete beams", Sûuctural with no compre the
Engineer, Vol.80, No. 1 6, 2002, pp26-12. upper limit, If K"
compression rei are
Limitations: derived. From this figure,
r 'Where
the redisl¡ibution of moments does not exceed 1002. I
for /1" S 40 N/mm2, rld <0.5 C=0.674-x09bx' z=d-045x
¡ñ
for 40 < f" I 70 N/mm2 , xld < 0.4
a-
for 70 < f, s 100 N/nm2 ,
xld < 0.33 and no moment redistribution M'= Cz = 0.67
r
"" x0 9bx x(d -0.4sx) = o'4s21ra'),r - 0.45xld)bd' f^
. Whe¡e the redist¡ibution of moments exceeds l0%.
for fu < 40 N/mm2, xld<(pr -0.4)
r'= Y . =0a02(xtd)(1-0'45x/d)
for 40 < /"" < 70 N/mmz, xld < (po - 0.5) bd'f."
moment after redist¡ibution
rr-r Wi¡rLr-r Dr = substituting the respective upper limit value ofx/d into the above equalion, the
moment before redistribution values of K' wrder differenl situations are obtained as:
 Where Îìe redistibution of moments does not exceed 10%, tests on the shear capacity ofhigh-strength concrete beams have been carried out
for f" < 40 N/mm2, K' :0.156 at The University of Hong Kong. The test results revealed that the shear strength
for 40< {" <70N/mm', K' =0.132 ofa reinforced concrete bea¡n without stimrps increases with increasing concrete
compressive strength but the rate ofincrease in shea¡ strength gradually dec¡eases
fo¡ 70< 7* <l00Nimm', K' =0.l13andnomomentredistibution as the conc¡ete strength increases until the shear stength stops increasing at a
. lVhere the redistibution of moments exceeds 100/o. cube strenglh of about 80 N/mm'. Based on this, the limiting value of design
for f* <40N/mm2, K' :0.402 (po - 0.4) - 0.18 (pb - 0.4)2 shea¡ stress has been increased to 0 8¿; with 4" set at 80 Nimm2. This
for 40 < l* < 70 N/mm2, K'=0.402(þ, -0.5)- 0.18 (/b - 0.5)2 actually gives a limiting value of design shea¡ sÍess of 7.16 N/mm2, but has been
rounded down to 7.0 N/mm'. However, at such high design shear stress level,
very heavy shea¡ reinforcement would be required. Moreover, the compression
struts formed inside the beam would be subjected to fairly high compressive
0.67 !t!- forces. To avoid brittle failure, it should be prudent to provide generous lateral
, ,yl confinement to ûre compression struts. Particular attention should also be paid to
the use of good quality aggregate because the shea¡ suength of high-strength
concrete is quite sensitive to the quality of aggregale used. The afore-mentioned
09r î f
"" x0.9bx
tests carried out at The University ofHong Kong have been published in:
C = 0.67 Islam M.S., Pam H.J. a¡d Kwa¡r A.K.H., "Shear capacity of high-stength

Neutral
axls
I concrete beams witl point of inflection within shea¡ spaa", Proceedings,
Institution of Civil Engineers, Structu¡es and Buildings, Vol.128, February, 1 998,
pp91-99.
z= d -O.45x

Shear reinlorcement:

r<- The shea¡ reinforcement to be provided is dependent mainly on the direction and
the magnitude of v relative to v" (Îhe design concrete shea¡ stress without shear
reinforcement) and v, (the nominal shear strength to be provided by minimum
shear reinforcement). v" is as given in Table 6.3 of the code, while v, is to
Figure 6.2 Evaluation of M' arñ K' ensure that when shear cracking occws, the tension originally canied by the
concrete before cracking can be transferred to the shea¡ reinforcement without
causing it to yield. In BS81 l0: Part 1, v, is set at a constant value of 0.4 N/mm2,
fo r fl anged bea ms.
De s i gn fo rmu las
Ifthe neutral axis lies within the flange, the design may be carried out following However, as the concrete strength increases, the shea¡ force at which cracking
the same procedures as for rectangular beams. Otherwise, the design should be occurs inc¡eases. Hence, the tension, which has to be tansmitted to the shear
carried out by direct application oftle assumptions given in Clause 6.1.2.4(a). reinforcement when shear cracking occurs, also increases a¡rd the a¡ea of shear
reinforcement has to be increased to prevent it tom yielding at the time of shear
cracking. The required minimum shea¡ reinforcemeni is proportional to the
tensile strength of concrete, which is assumed to be proportional to its
eiven in the follorarhg:
compressive slrength to the power 2/3, Therefore, for concrete strength higher
Shear stress:
tha¡ 40 N/mrn', the minimum shear reinforcement has to be increased by a factor
The design shear stess v at any section should be calculated from: of (¡",t+O)'u . The amount ofshear reinforcement to be provided is given by:
'Where v < 0.5 v" throughout the beam:
v
b"d no requirement fol structu¡al elements of minor impofance;
minimwn links for slructu¡al elements of importance.
where Z is design shear load and å" is tle breadth ofthe seclion (or the breadth
Where 0.5v" S v < (v"+v,):
ofthe web or rib below the flange in case ofa flanged beam).
minimum links to be provided.

I¡ no case should v exceed 0.E/* or 7.0 N/mm2, whichever is the lesser,

Where (v.-v,) I v <min. (0 S.m or 7.0N/mm2):
regardless ofthe amount and type ofshea¡ reinforcement provided. It sÞould be - v. ) I 0.87 f y, if only links provided;
1"" > å"s" (v
noted that BS8l10: Part 1 limits the design shear süess to only 5.0 Nimm'. Some up to 50% ofthe shear resistance to be provided by t}le shear reinforcement
may be in the form ofbent-up bars.

-F-
Concrete shear stress:
The desip. ooncrete shea¡ stress vc is derived ûom the expression:
/ r nn z¡4¡g\"0 ( f-..\''t l'
'""T )rÆll'"-lf
v-=0191
" '-l
t
b,d ) \d ) \2s) y_
in which the value of f should not be taken as greater than 80 N/mm2 (note that
BSSI l0: Pa¡t I limits the value of /* at not greater than 40 Nimmz). The term
,c. is the a¡ea of longitudinal tension rei¡fo¡cemen! which continues for
distance of at least d beyond the section being considered. At supports, only the
part of longitudinal tonsion reinforcement, which meets the curtailment and
anchorage requirements at the section being considered, should be counted
towards ,4, .
Shear resistance under diferent conditions'.
The provisions given in lhe new code fo¡ the detemrination of shear resistance
underdiffe¡entconditionsaresimila¡tothose inBS8ll0:Partl: l99Texceptthat
the material safety factor of the shea¡ rcinforcement is set at 1.15 and when f,
40 N/mm2, both the design concrete shea¡ stress v" and the limiting value of
design shear stess v are higher (increase in shea¡ stength until f" reaches 80
N/mm2 is allowed).
6. 1.3 Solid slabs supported by beams or walls
The provisions for the design ofbe"-s are also applicable to solid slabs.
Slabs subjected to distibuted and concenfated loads may be analysed using an
appropriate elastic analysis method or alternatively a plastic analysis method.
For elastic analysis, an amllical method may be used if the slab has simple
geometry but the ûnite element method may have to be used if the slab has
complex geomeüy or is subjected to a complicated anangement of loading. For
plastic analysis, both Johansen's yietd line method ånd Hillerborg's strip method
are considered acceptable.
For stabs subjected to concenl¡ated loads, there will be concentrated moment
occurring at each loading point. Due to stess concentration, the concentrated
momenl could be many times larger than the average moment actiDg on the
section undemeath the loading point parallel to the nearest support. To capture
such concenhated momen! if the finite element method of analysis is employed,
a very fine mesh is required. However, in reality, because of the large magnitude
of the concentated momeft, the section at the loading point would yield at an
early stâge and the moment would be redist¡ibuted to other parts of the slab,
resulting in a less concentated and more uniform diskibution of moment. As
slabs are generally quite lightly reinfo¡ced and should trerefore possess a fai¡
amount of flexural ductilit¡ such redis&ibution of moments could be very
subslantial. To lake advantage ofthe redistribution ofmoments, which can help
to reduce the maximum moment to be designed for, it may be assumed that the
r
slab moment is resisted by an effective width over which the total moment (the
loul moment may be obtained by integrating the moment per unit width
determined by a rigorous elastic analysis along the effective width) is uniformly
distibuted. For a slab simply supported on two opposite edges, the code has
provided guidelines for the evaluation oflhe effective width. These guidelines
.
are the same as those given in BSS l l0: Part L: 1997 For any other slab with a
more complicated geometry or support aÍangement, no general guideline could
be given and rigorous limited ductility elasto-plastic analysis will Ìnve to be
a ca¡ried out ifadvantage is to be taken ofthe possible redistibution ofmoments 1o
reduce the maximum moment to be designed for.
Apart ûorn the above redisüibutiot of moments ac¡oss the effective width, the
redist¡ibution of moments from the suppods to the mid-span sections may also be
carried out as per Clause 5.2.9 ofthe code. Ifthe analysis ofthe slab is carried
out for the single load case of all spans loaded in accordance with Clause
6.1.3.2(c),1he redist¡ibution of moments is limited to 20%.
For the elastic analysis of rectangular fivo-way spanning slabs subjected to
>
unifonnly distibuted loads, the code has provided formulas fot the di¡ect
evaluation of the design maximum moments. These formulas a¡e identical to
those given in BS81 10: Part 1: 1997,
Regarding the shea¡ design ofsolid slabs, the shea¡ ¡esistance ofa solid slab is to
be evaluated as ifthe solid slab is a beam (i.e. the values of v" and v. are to be
evaluated in the same way) but the shear reinforcement may be designed
according to the following more relaxed rules:
Where y y":(
no requirement for shea¡ reinforcement.
Where v. < v <(v"*v,):
minimum links to be Provided.
Where (v"+v,) S v < min. (0 8JÂ or 7.0 N/mm2):
,4* > å" s" (v - v. ) /0.E7 fy , if only links provided;
the shear reinforcement may be provided in the form of links and/or bent-up
bars in any combination.
6.1.4 Ribbed slabs
The¡e are two types of ribbed slabs: one type with the topping considered to
$ contribute to stuctural strength and the other type wit¡ the opping not
.i!l
considered to contibute to süuctual strength. Each ofthem has to meet certain
requirements as stipulated in Clause 6.1.4.1 of fhc code' Where the topping is
considered to conûibute to structural strength, its thickness should not be less
than the minimum thickness given in Table 6.9.
The continuous span of a ribbed slab may be analysed as a one-way spanning
solid slab as per Clause 6.1.3.2 ofthe codc. Ifthe ribbed slab has equal structual
properties in two perpendicular directions (see Clause 6'1.4 2) nd its flanges and
r--
t-
ribs have sufficient torsional stiffiress (see Clause 5.2.1.1), it may be analysed as
a lwo-way spanning solid slab in accordance with Clause 6.1.3.3 or as a flat slab
in accordance with Clause 6.1.5, whichever is more appropriate.
Ifit is impracticable to provide sufticient reinfo¡cement to develop the full design
moments at the supports, the desígn moments at the suPports may be redistributed
to the mid-span sections. Clause 6.1.4,2 of the code allows 100% of the momenls
at the supports to be redistributed to the mid-span sections so that effectively the
slabsaredesignedasaseriesofsimplysupportedspans. Ifthisisdone,sufFrcient
reinforcement should be provided over the supports to control cracking' The
code recommends that such reinforcement should have an area of not less than
25% ofthzt in the middle ofthe adjoining spans and should extend at least 15%
ofthe spans into the adjoining spans. Nevertheless, lhe auüro¡ is ofthe view that
in order to avoid excessive cracking, the redisnibution of moments should be
limiæd to say only 50%; 100% redisribution of moments is a bit too much and
may adversely affect the sewiceability ofthe structu¡e.
Regarding the design resistance moment, the provisions given in Clause 6.1'2.4
for determining the design resistance moment of beams may also be applied to
ribbed slabs. When determining the design resistance moment of ribbed slabs,
the sEength conEibution of the forming blocks, which remain part of the
completed structure, may b€ taken into account.
Regarding the design shea¡ resistance, the provisions given in Clause 6.1.2'5 for
determining the design shea¡ resistance of flanged beams may also be applied to
ribbed slabs, except thåt the breadth ofthe ¡ib may be increased to allow for the
contribution of the forming blocks. The she¿¡ reinforcement may be desiga as in
the case ofsolid slabs.
6.1.5 Flat slabs
The provisions given in tbe code, which are identical to those given in BS8l10:
Part l: 1997, are for application to flat slabs supported by a generally rectangular
anangement of columns with a longor to shorter spans ratio not exceeding 2.0.
Fo¡ such kind of flat slabs, the equivalent fiame method may be used for the
analysis, as provided for in the code. Fo¡ flat slabs wíth more complicated
arrangement of columns, other methods, e.g. the finiæ element method, may have
to be used, but in such cases the desigr engineer will have 1o make his/he¡ own
judgement as to whether the provisions in this section are applicable'
In any direction, the effcctive dimension of a column head /r may be taken as:
: It.n
/n min. (/n. or (*. ) Ìro
where lno is the actual dimension msasured 40 mm below the soffrt of the slab or
drop, and /r,o is given by:
/n* = l" +2(dh-40)
of
in which l. is the dimension of column in same direction and dn is the depth
the column head.
Drops are provided mainly to increase the punching shear resistance ofthe slab.
Their possible influeuces on the distribution of moments within the slab should
be ignoted "nless 1ìe smaller dimension of the drop is at least 1/3 of the smaller
dimension of the surrounding panels.
For flat slabs consisting of ¿ series of rectangular panels, i.e, supported by a
generally rectangular arrangement of colurnns, the equivalent frame method of
dividing the stn¡cture longitudinally and transversely into frames comprising of
columns and sEips ofslab may be used for the analysis. Although for analysing
the equivalent frames, simplified methods of firfher dividing each frame into
sub-frarres as mentioned in Clause 6.1,5.2(d) of the code are permitted, it is
considered better to employ a more rigorous analysis method based on either
grillage analysis or ñnite element analysis.
For equivalent frame analysis, the slab panels should be assumed to be divided
into column strips and middle strips, as depicted in Figure 6.9 of the code. After
obtaining the desiga momeDfs ftom the analysis, the design moments should be
apportioned between the column strips and middle strips as per the ratios given in
Table 6.10 ofthe code.
Rules for dealing with the positive and negative design moments in internal and
edge panels and for dealing with openings in panels are given in Clauses 6.1.5.3,
6.1.5.4 and 6. 1.5.5 ofthe code, respectively. These are the sane as those given in
BS8ll0: Partl:1997.
The critical consideration for shear is the punching shea¡ a¡ound the columns. To
design for punching shear at a column, the design effective shea¡ force l/on needs
first to be detennined, as per the code using Equation 6.40 for an internal column
or Equation 6,41 for a comer or edge column. It should be noted that the desigrr
effecdve she¿r force Z.ñ is trot the same as the design shea¡ transferred to the
column and actually comprises of two components:
¡ the shear force component arising fr,om the design shear ( transferred to ûre
column; and
. the shea¡ force component arising from the design moment M, transfened
1o the column.
The above two componetrts a¡e more apparent ifEquation 6.40 and Equation 6.41
ars rewritten into the following forms:
You=Yr+1.5Y! fo¡ an intemal column
r.p
: It
1.25V. + 1.5:l- for a comer or edge column
The maximum design shear stess at a column so evaluated using Equation ó.40
or Equation 6.41 above and Equation 6.42 on a perimeter bounding the column
face or the column head should not exceed 0 e.,[
or 7.0 N/mm2, whichever is
the lesse¡ (note üre increase in maximum design shear stress compared to
BSBll0: Part l: 1997).
 r-_

The nominal design shear sfess v appropriate to a particular perimeter being Moments and þrces Ìn columns:
considered is calculated as:
The moments and forces in the columns due to the design loads shall be
evaluated by analysis ofstucfire as per Chapter 5 ofthe code. However, ifthe
v=--uaY coluon being considered is a slender column, the moment due to the design loads
should be added with the deflection induced momeûÎ (also referred to as
where u is the effective len$h of the outer perimeter of the punching failure additional moment induced by deflection) to obtain the design moment at ULS'
zone. Based on the value of v retative to v",lhe shea¡ ¡einforcement may be Then for any column, regardless of whether it is a short or slender column, the
designed according to the following rules (which are simila¡ but not exactly the design moment at ULS should not be tâken as smaller than the desþ ultimate
same as those fo¡ solid slabs): axial load multiplied by a minimum eccentricity e"" , which may be set equal to
Where v ( v": the minimum of (0.05 times overall dimension of column in the plane being
considered or 20 mm).
no requirement fo¡ shear reinforcement.
'Where 3v 32v":
v" Deflection induced moments at ULS:
shear reinforcemetrt may be provided in the form of links, as per Equation The deflection of a rectangular or ci¡cula¡ column at ULS may be taken to be:
6.44when vo I v I l'6vo andEquation6.45 when 1.6v.< v 32vo'
a" = p,Kh
the shea¡ reinfo¡cemer¡t provided should not^be less than that required to
provide â nomin¿l shear strength of 0.4 N/om'. in which:
2v" < v and a reinforcing system is provided: ¡\2
Where
justification should be provided to demonstr¿te thc valiclity ofthe design' p,=# b)
N--N
x= lv_ <1
6.2 Member¡ rxialþ loaded wlth or wlthout flexure -,lvr¡
/, is to allow for the slendemess of the column. Its variation with /./å has been
6.2.1 Columns t¿bulaæd in Table 6.13 ofthe code. On ûre otler hand, X is a reduction factor
that corrects the deflection to allow for the influence of æ<ial load on the size of
Short and slender columns: compression zone in the column section (note that K < I when ,V > lfou and K
columns a¡e classiûed inûo short columns and slender columns, which have to be : I when il I iV", ). The deflection i¡duces an additional moment given by:
dealt with in different ways. A column may be considered as short when both
Io/h alrd lo/b arc less than 15 if or 10 if the column is
the columu is braced Mto= Na,
rmbraced. otherwise, it should be considered as slende¡. The effective height l. The distibution of îhe above addition¿l moment along the height of the column is
of a column in a given plane is equal to p lo wtrere l" is the olear height of the dependent on wheúcr the column is braced and on the end conditions.
column. If bolh ends of the column are restrained (i.e. are connected to othe¡
members), the clea¡ height /" of the column is the clea¡ height between the end
For a braced column, if both ends are not providing rotational resFaint, the
disl¡ibution of additional moment is like a ll2 cycle sine curve with a maximum
restaints. In such case, the clear height lo should not exceed 60b. Ifone end is value at mid-height. If one end is providing rotational restraìnt while the othe¡
un¡esüain to any other members)' the clea¡ end is not. the dist¡ibution of additional moment is like a 314 cycle sine curve
height of e from tbe clear su¡face of the with a maximum value at approximately 0'4 times column height from the end
restained clear height /o should not exceed
with no rotational resþaint. If both ends are providing rotational restraint, the
min. ( tlpes of end conditions are defined' The first tϡee distribution of additional moment is a full cycle sine curve with a maximum
arc ap and unbraced columns whereas the fourlh one (Aee value at mid-height. The distribution of additional moment and the resulting
end c only to rmbraced columns' Depending on the design moment envelopes for different end conditions a¡e shown in Figure 6.3.
oombination of end conditions at top aûd bottom, the value of the factor nay p For an unbraced column, if one end is providing rotational restraint while the
be obtained from Table 6.ll for braced columns and from Table 6.12 for
unbraced columns. From these tables, it can be seen that for braced columns, the
values of p
are generally smaller than or equal 1o 1.0 while for unbraced
columns, úe values of ate generally larger than or equal to l'2' Hence, the
p
distibution of additio¡al moment and the resulting design moment envelopes for
slendemess ratios (i,e, l^ f h and /e /b ) of unbraced columns tend to be larger. different end conditions are shown in Figure 6,4.
 r
tress v appropriate to a particular perimeter being Moments andforces in columns:
'? o..
The moments and forces in the columns due to the design loads shall be
r'- evaluated by analysis of stucture as per Chapter 5 of the code. However, if the
¿"t -' colurnn being considered is a slender column, the momeut due to the design Ioads
should be added with the deflection induced moment (also referred to as
'the outer perimeter of the punching failure additional bv deflection) to obtain the desien moment at ULS
tivè.to v", the shear reinforcemenl may be Then for any resardless of whether it is a short or slende¡ column, the
6 rules a¡e similar but not exactly the design at ULS should not be taken as smalle¡ than the desien ultimate
axial by a minimum eccentricity e,i", which may be set equal to
the of (0.05 times ove¡all dimension of coh¡mn in the plane being
for shear reinforcement' or 20 mm).
-r.dt

.: v 32v,: Deflection induced momenls at ULS:

-.sfiéa¡ ¡einforcement may be provided in ú¡e form of linksì per Equation
The deflection ofa rectangular or circular column aÎ ULS may be taken to be:
6.44when v" 5 v I l6r" andEquation6.45when 1.6v" < v
the shea¡ reinforcement provided should not^be less than that
4u: þ,ffi
provide a nominal shear strength of 0.4 N/mm'.
$y'here 2 v. < v and a reinforcing system is provided:
justification should be provided to demonstate the validity ofthe design-
| (r.\'
/'^ 2ooo[óJ
¡: N*-N .t
6.2 Members axially loaded with or without flexure N* - Non

fl is to allow for the slendemess of the column. Its varíation wirh \lb
has been
6.2.1 Columns in Table 6.13 ofthe code. On the other hand. K is a reduction factor
the deflection to allow for the influence of axial load on the size of
Short and slender columns ion zone in the column section (note that K < 1 when lr' > No.' and K
Columns are classified info short columns and slender ich have to be :1 when No., ). The deflection induces an additional moment given by:
dealt with in different ways A eolumn may be conside short when both
I,* / h amd 1," / b ate less tha¡r 15 if the column is bracedþt 10 if the column is M^¿¿ =N
unbraced. Otherwise, it should be considered as slendey' The effective height i. The distibution of additional moment along the height of the column is
of a column in a given plane is equal to p l" where y'. is the clear height of the dependent on whether is braced a¡d on the end conditíons.
column. If both ends of the column a¡e restrairy¡d (i.e. are corurected to other
members), the clear height /" of the column is,the clear height between the end For a braced column, if are not providing rotational reshaint, the
dist¡ibution of additional a l/2 cycle sine curve with a maxrmum
restraints. In such case, the clear height /" not exceed 60á Ifone end is
value at mid-height. If one end is rotational restraint while the othe¡
un¡estrained li.e. is a free end not to any olher members), the clear
end is not, the distibution of additional is like a 314 cvcle sine curve
beisht of the column is the vefical from the clear surface of the
with a maximum value at approximately 0.4 times column height from the end
rest¡ained end to the fiee end. In such dse, the clear height /" should not exceed
witl no rotational restaint. If both ends are providing rotational restraint, the
min. (60 å or 700 b'z lh) Four types o/end conditions are defined. The fi¡st th¡ee distibution of additional moment is a full cycle sine cuwe with a maximum
a¡e applicable to both braced and columns whereas the fourth one lfree value at mid-height. The dist¡ibution of additional moment and the resulting
end condition) is applicable to unb¡aced columns. Depending on the design moment envelopes for different end conditions are shown in Figure 6'3'
combination of end condítions a¡rd bottom, the value of the factor / may For an unbraced column, if one end is providing rotational restaint while the
be obtained from Table 6.11 for b¡aced columns and from Table 6.12 for other end is not (i.e. a free end), the distibution of additional moment is like a
unbraced columns. From these tables, it can be seen that for braced columns, the ll4 cycle sine cuwe with a maximr¡m value at the end with rotational resl¡aint. If
values of p arc generally smaller than or equal to 1.0 while for unbraced both ends are providing rotational restraint, then the dist¡ibution of additional
columns, the values of p are generally larger than or equal to 1.2. Hence, the moment is like a 712 cycle sine cu¡ve with maximum values at the ends. The
slendemess ratios (i.e. I h ar,d /." /b ) of unbraced columns tend to be larger. disfibution of additional moment and the resulting design moment envelopes for
I
"* different end conditions a¡e shown in Figure 6.4'
...-1_.

-
Dcsign
End conditiom Initial nì0rÌcn1s Additional moments
lions In¡t¡al moments Additional Design of column (from analysis) nr0menls cnvelope
dltlÍn (from analysis) mom€r¡ts
mommts
-
envelope
-ì Mt¿t --i+
i+ I

->l ê;tu-

-ì.- l+

Stiffer end joint

of stiffriess /tr I -l<- <-

Less stiffjoint -> N2

of stiffrtess /r2 tvÃz=tutù
->l , o5M,a kl

Figure 6.4 Desigrr moment envelopes - unbraced columns

Fo¡ unbraced columns at t¡e same storey level, which are conshained to deflect
sideways by the same amormt, an average ultimate deflection ¿"," may be applied
I
to all the columns. The average deflection du." ñâY be evaluated u, ¡o" io
M1
Smalle¡
which n is the number of columns.
moment
In theory, all slender colußns should be designed for biaxial bending. However,
Clause 6-2.1.3(c) allows slender columns bent about a single axis to be designed
Figure 6.3 Design mometrt envelopes braced columns only for nniaxial bending provide d hlb <3 and l"f h <20 .
-

56
ffieEääWY"-'

-
on of column sections lor ULS: reinforced concrete wall supportíng the beams and slabs which are simply
m the analysis of column sections for ULS, the sa.me assumptions should be supported on the wall, the effective height /. should be assessed as though the
made as for the analysis ofbeam sections. For a short column subjected only to wall is a plain wall. Except for short braced walls loaded almost symmetrically,
axial load, the ultimate axial load lr' may be evaluated by: the eccenticity e." of the axial load in the direction at right angle to the wall
N: 0 4f,,A" + 0'tsfyA," (i.e. in the out-of-plane direction) should be taken as not less than 0.05 times the
thickness of wall or 20 mm, whichever is lesser (similar 1o that for columns).
For a short braced column supporting an approximately symmetrical arrangement
of beams, the ultimate axial load N may be evaluated by:
Clause 6.2.2.2(d) of the code suggests that when a horizontal force is resisted by
N :035fc,A"+0.67LA* the in-plane action of several walls, the proportion allocated to each wall should
be in proportion to its relative stiffness. However, the term "¡elative stif[ness" is
Fo¡ a column to be designed for biaxial bending, Clause 6 2.I.4(d) of the code
provides a simple method, which is applicable only if the column section is
difficult to define and for many years has been loosely interpreted by different
engineers in totally dif[e¡ent ways. ln fac! ftom the theory ofsl¡ucture point of
rectangular and symmetrically reinforced. In the mo¡e general case of an
view, this poshrlation that the horizontal shea¡ should be apportioned to each wall
arbitarily shaped column section, such as a circular section, an L-shaped section,
in proportion to its relative stiffness is wong. The distribution of the horizontal
a polygonal-shaped section and a section with a hole etc, a more rigorous and
generally applicable method will h¿ve to be used. Fully computerized design load to the walls is dependent on many factors including spatial distibution of
the walls, torsional stiJfness of the building and coupling between the walls, as
methods for columns subjected to axial load and biaxial bending may be found in
illustrated in Figure 6.5, and wall-frame interaction, as explained before.
the followin g publications :
Kwan K.H. and Liauw T,C, "Computerized ultimate strcngth analysis of
¡einforced concrete sections subjected to axial compression and biaxial bendìng",
Computers and Stuctures, Vol.2 1, No.6, 1985, ppl I 19- 1 127.
Horizontal load Horizontal load
Kwan K.H. and Liauw T.C., "Computer aided desigl of reinfo¡ced concrete
members subjected to axial compression and biaxial bending", Structwal I I

Engineer, Part B, Vol.63B, No.2, June, 1 985, pp34-40.

Liauw T.C. and Kwan K.H., "Computerized design of reinforced concrete
columns by load factor method", Journal, Hong Kong Institution of Engineers,
February, 1988, pp23-30; also in: Traasactions, Hong Kong lnstitution of È È >
Engineers, Vol.1, 1989, pp3-10.

62.2 Walls
(a) Case I þ) Case 2
General:
The overall stability of a multi-storey building should not, in any direction,
depend on the out-of-plaae rigidity ofthe walls alone. Lateral supports to walls
should each be designed to resist, in the out-of-plane direction, the sum of the Horizontal load Horizontal Ioad
design ultimate horizontal load acting on the wall (due to the applied loads) and a I I

horizontal load equal to 2.5Yo of the design ultimate vertical load (for preventing
lateral buckling). From the design point of view, walls a¡e classified into
reinforced walls and plain walls. A reinforced wall is a conc¡ete wall containing
at least the minimum qua¡tities of reinforcement specified in Clauses 9.6.1 to
9.6.4 of the code. A plain wall is a concrete wall containing no reinforcement or
insufficient reinfo¡cement to satisfl the minimum quantities of reinforcement
iÈn Ø; Ë
Ì
*ø
=n ut Ø= ¡

specified in Clauses 9.6.I to 9.6.4 of the code; for a plain wall, any reinforcement
is ignored when considering the strength ofthe wall.
(c) Case 3 (d) Case 4
Des i gnof reinforc e d w a lls :
For a reinforced wall constructed monolithically with the adjacenl construction,
the effective height /. should be assesssd as though the wall is a column Fieu¡e 6.5 Distribution ofhorizonlal load to walls
subjected to bending at right angle to the wall (i.e. oufof-plane bending). For a

58
 I
-
Fou¡ different cases are shown in Figure 6.5. In Case l, a horizontal load is
applied to the geomefiic cen1re of a building wilh two shea¡ walls of diffe¡ent
lateral stifûress each al one side of the building. Although the stiffuess of the two
shear walls are different, because of the necessity to satisfu the torsional
equilibrium condition, the loads distributed 1o the two shear walls are the same,
i.e. each of the two shea¡ walls is allocaæd one half of the applied horizontal
load. In Case 2, a horizontal load is applied to the centre of a building with a
core wall and lwo shea¡ walls of different stiffness. In this paficular case,
because of the high torsional stiftess of the core wall, the building would deflect
Iaterally with little nvisting. As a result, the horizontal load is distributed to the
core wall and the shear walls in proportion to their lateral stiffness. In Case 3, a
horizontal load is applied to the cenfe of a building with a symmetrical
anangement of shear walls. Because of symmetry, the building would deflect
laterally with no lwisting. As a resul! the horizonl¿l load is distributed to the
walls in proportion to their lateral stiffness. In Case 4, a horizontal load is
applied to the centre of a building with foru shear walls, two of which are solid
walls while the other two have window openings and a¡e in effect coupled shear
walls. The problem in this particular case is that it is very difficult to determine
the lateral stiftress ofthe two coupled shear walls, which do not deflect like solid
cantileverwalls. Becauseofthedifferenceinthedeflectionmodesofthecoupled
shear wails and the solìd shear walls. the horizontal load is not distributed to the
walls in proportion to thei¡ lateral stiffness.
I¡ view of the afore-mentioned situations that may a¡ise in many multistorey
buildings, it ís not really appropriate to suggest that the ho¡izontal shear should
be apportioned to each wall in proportion to its relative stiffness. îhe distribution
of horizontal loads to the shear/core walls should be analysed by a three-
dimensional finite element analysis prograrn with the shear walls modelled as
plane stress or shell elements taking into account the spatial distribution of the
shear walls, torsional rigidity of the core walls, coupling effect between shea¡
walls and wall-frame interaction.
Regarding the design of the walls, reinforced walls a¡e classified into stocky
reinforced walls (i.e. reinforced walls with l,f h < 75 if braced or with l./ft < 10
if unbraced) and slender reinforced walls (i.e. reinforced walls other tlran stocky
reinforced walls). They are analogous to short and slender columns. The design
ofstocky reinforced walls and slender reinforced walls are governed by Clauses
6.2.2.2(Ð ar:,d, 6.2.2.2(g), respectively.
Design of plainwalls:
The aulhor does not recommend the incorporation of any plain walls in multi-
storey buildings, especially tall buildings, the main reason being that in multi-
storey buildings, basically all the walls are key elements that âre indispensable
and need to be designed to withsta¡d a design ultimate load of 34 N/m' in any
direction including the outof-$lane di¡ection. Such a high-intensity design
ultimate load would demand more than nominal reinforcement to be provided in
the walls. Hence, it is inappropriate !o design any walls in multi-storey buildings
as plain walls. At the most, in a multi-storey building, plain walls should be used
for non-stuctu¡al partition walls only.
6.3 Torsio¡ and combiDed effects
It is said in the code that in "normal" slab-and-beam or framed construction,
specific considerations of torsional effects a¡e not usually necessary There is,
however, the question of what is normal and what is not normal. The decision on
whether to consider lorsional effects should be based on whether the torsional
stiffness of the membe¡ being considered would have significant effect on the
load distribution of the structu¡e. If the torsional stiffness of the member could
have certain effects on the load distribution, the torsional stiffness ofthe member
should be taken into account in the analysis; otherwise the lorsional stiffness of
the membe¡ may be ignored. For example, when the membe¡ being considered
does not lie in a vertical plane (i.e.Îhe member is cu¡ved horizontally or has a
horizontal bend along its length), the torsional sdffness is likely to have some
effects on the load distribution and therefo¡e should be taken into âccounl in the
analysis. As anothe¡ example, when a horizontal member is connected at an
angle to another horizontal member in such a way that the flexural rotation of a
member would become the torsional rotation of the other member, the to¡sional
stiffness ofthe member subjected to to¡sional rotation should be considered in the
analysis. Throughouttheyears,theauthorhasseenquiteanumberoftorsionally
cracked beams in Hong Kong, which were most probably due to the belief of
some design engineers that "specific consideralions of torsional effects a¡e not
usually necessary" The autho¡ is of the opinion that design engineers should
better have a second thought and always satisf themselves that the torsional
effects are really negligible and are not going to cause torsional cracking before
negìecting any torsional effects.
In any case, when the torsional resistance of any member is relied on to carry
loading, the torsional stiffness, torsional reinforcement and torsional cracking of
the member should all be carefrrlly considered. However, the author does not
recommend the reliance on the torsional resistance of any membe¡ to carry
loading, unless there is no other altemative.
The main difficulty in the evaiuation of the torsional consta¡t C of a concrete
section is that it is dependent on the extent of cracking of the concrete section. in
the case of a concrete section subjected to flexure, it can be assumed that all the
concrete wilhin the tension zone as demarcated by the position ofthe neutral axis
has cracked, but in the case of a concrete section subjected to torsion, no similar
strategy can be applied. To avoid such difficulty, it is assumed in the code that
the torsional constant C may be taken as half the St. Venant value (the value
evaluated for the whole section assuming there is no cracking), The rorsional
constant C is given by:
c=
lØn^,,'n^*)
in which / is a coefficient depending on the shape ofthe section,
å,," is the smaller dimension of the section,
å,"" is the larger dimension ofthe section, and
h^*) is the St. Venant torsional constant.
the bracketed term þ h-t^'
-
The above should clarify the second paÉgraph in Clause 6 32 ofthe code (where If the combined shear stress due to shea¡ a¡rd torsion exceed the value of v,.,n in
the term "St. Venant to¡sionai constant" should be replaced by the term "torsional Table 6.17, torsional reinforcement should be provided but in no case should the
constanf' and Equation 6.64 therein should be amended accordingly). Further combined shea¡ stress due to shea¡ and to¡sion exceed the value of v," in Table
explanations can be found in the following book: 6.17. The values of v,,in ild v" given in Table 6.17 zte based on those given in
A.H. Alten, Reinforced Concrete Design to BS81l0 - Simpiy Explained, E. & Table 2.3 of BSS l 10: Part 2: 1985, but have been modified for increased shear
F.N. Spon Ltd, London and New York, I 988, 239pp.
sfength when concrete of strength grade higher than C40 are used.
The torsional stiffness of a non-rectangular section may be obtained by dividing
The stipulations for provision oftorsional reinforcement in the code a¡e tÏe same
the seclion inlo a number of rectangles and summing up the torsional stiffrress of
as those in 858110: Parl 2: 1985 except that the material safety factor of the
these rectangles- The division of the section should be arranged so as to
torsional reinforcement has been set at 1.15 instead of 1.05 in the British
maximize the calculated torsional stiffness, as illustrated in Figure 6.6 for the
Standa¡d. Basically, the torsional reinforcement should consist of rectangular
case ofa¡r l-section. closed liriks togetrer with longitudinal reinforcement and is additional to any
requirements for shear or bending.
b,4 _L
hl 6,4 Desþn for robustness against disproportionate collapse
T I Clawe 2.2.2.3 of the code has demanded, for the purpose of achieving stuctüal
À h-h-ht integrity, that the stuctural elemenfs should be effectively tied together using the
following types of ties:
I . peripheral ties;
h6
o intemal lies;
F_ å¡_____-l F_ åu-_-N T r horizontal ties to columns and walls;
r vertical ties to vertical load-bearing elements.
The ties should be provided with tension reinforcement effectively anchored to
c = [pr b *3 h+ prh þ,- b *)+ &h61 Q f b *)]t2 cz= W 4 h,3 b,+ p j hb1 b 6+ p 6b.r (h - h,- h6tlt 2
the stuctural elements that are being tied together. Reinforcement provided for
rt

C = maximum of (Cr or G) other pu4roses may be regarded as forming part of, or the whole of, these ties.

Perípheral ties'
Figure 6 6 Calculation of lorsional constrant for an I-section
At each floo¡ and roof level, an effectively continuous peripheral tie capable of
resisling a tensile force of ,{ should be provided within 1.2 m ofthe edge ofthe
To evaluate the torsional shear stress in the section, the to¡sion carried by each building or wifhin the perimeter wall. 4 (i" kN) is the lesser of (20+4ns) or 60'
rectangular component needs to be calculated first. It may be assumed that the where zs is the number of storeys.
torsional mom.trt carried by each rectangular component is proportional to its
torsional stiffness, or in other words, given by the following equation: Internal ties:
At each floor and roof level, intemal ties in two directions, approximately at right
h.,"3h^* angles, should be provided. They should be effectively continuous throughout
r'-1x\1;-r;_l
¿ !í'm¡n fr8x / their lengfhs and should be anchored to the perþheral ties at each end. They
may, in whole or in part, be spread evenly in the slabs or may be grouped at
in which I¡ is the torsional moment acting on a rectarìgulã component and I is be¡-s, walls or other appropriate positions, but at spacings not greater than 1.5 i.
the torsion acting on the whole section. Having obtained the torsional moment where /, is the greater of the distances between the centres of the columns,
acting on each rectangular comPonen! the torsional shear stress v, in any
frames or \ryalls supporting any two adjacent floor spans in the direction ofthe tie
rectangular component may be determined as: under consideration. In each direction, the ties should be capable ofresisting a
tensile force (in kN/rn width) equal to the greater of:
vr: Trx t^ ^\
À..'(å.* -h-^ß) \etr9r/*'.*¡ o, ¡
7.5 5
The above equation is derived by assuming a plastic shear stess distribution cha¡acteristic dead and imposed floor loads.
where (G,. + ø ) is the sum ofthe
circulating round the rectangle
Horìzontal îíes lo columns and walls:
Each extemal column and every mete length of each extemal wall carrying
vertical loads should be anchored o¡ tied horizontally into the structu¡e ¿t each
floor and roof level with a horizontal tie capable of resisting a tensile force equal
to the greater of:
I not less than one-halfofthe depth at the root ofthe projection
. 2.0 4 orif4 ifless,where /, isthefloortoceilingheight(inm); or
. 3Yoo¡ttr" ioll design ultimate vertical load carried by the column or walls.
For a comer column, honzontal ties should be provided in two directions
approximately at right angles. Each horizontal tie should be capable of resisting
the above tensile force,
Vertical ties to vertical load-bearing eleuents:
Each column a¡rd wall carrying vertical loads should be tied continuously from
the lowest to the highest level. The tie should be capable of resisting a tensile
force equal to the maximum design ultimate dead and imposed load received by
the column or wall ftom any one storey.
There are only t$'o other intemational codes giving explicit design ¡equirements
for the provision ofties in concrete buildings. They are BS81l0: Part l: 1997
and Eurocode 2. The requirements of these two codes are not the same.
Generally, the tie forces given in BS8l10 are based on the number of sloreys
while those ofEruocode 2 a¡e based on span lengtl. Hence, direct comparison
between these two codes is rathe¡ difEcult. However, the much smaller tie force
requirements for intemal ties and horizontal ties to cohunns in Eurocode 2
compared 1o that stipulated in BS8l10 does not appear to have beenjustified. It
isthereforeconsideredmoreappropriatetofollowBSS1l0. Thelierequirements
given in the new code a¡e the same as those outlined in 858110'
Other than the provision ofties, the necessþ to provide bridging elements should
also be considered. According to Clause 6.4.2 of the code, a design check should
be carried out to evaluate, fo¡ each storey in tum, the scena¡io oflosing any one
ofthe vertical load-bearing elements. The design requirement is that the loss of
any vertical load-bearing element would not cause the collapse of a signifrcant
part ofthe stucture. The vertical load-bearing elements 1o be considered lost in
tum are the columns and the walls. In the case of a column, the whole column
would have to be considered lost. I¡ the case of a wall, only lhe Iength ofthe
wall between adjacent lateral supports or between a lateral support and a free
edge needs to be considered lost.
According to Clauses 2.2.2.3 and' 6.4,1.7 of the code, all vertical load-bearing
elements should be provided with vertical ties. If, somehow, it is inappropriate or
impossible 1o provide effective vertical ties to all the vertical load-bearing
elements, then the verfical load-bearing elements not provided with ve¡tical ties
should be considered to be removed in tum and the vertical loads normally
carried by each such element in question should be t¡a¡sfened through bridging
elemenls 1o the adjâcent vertical load-bearing elements. Catenary action may be
employed to transfet the vertical loads but allowance should be made fo¡ the
horizontal ¡eactions lecessary for equilibrium.
Corbels and nibs
A co¡bel is a short cantilever Projection satisfying the following conditions:
. a"ld <l
. the depth ofthe projection at outer edge ofcontact area ofsupported load is
Corbels may be designed using the strut-and-Îie model, wilh the concrele and
reinforcemeff assumed to act as struts and lies, respectively. A1l reinforcement
should be properly anchored. At the front face, the ¡einforcement should be
ancho¡ed either by welding to a ûansverse bar or by bending back to form a loop
Shea¡ reinforcement should be provided in the form of horizont¿l links
distdbuted in the upper two-thirds ofthe effective depth at the root ofthe corbel
Where a continuous nib is less than 300 mm deep, it should be designed as a
short cantilever slab The tensìon reinforcement should project ûom the
supporting member across the top ofthe nib to a point as near to the front face of
the nib as the cover requirement permits. It should be anchored at the front face
either by welding 1o a t¡ansverse bar or by bending through 1800 to form a loop.
6.6 Stairc¡ses
The provisions given in the new code for the design of staircases a¡e the same as
those in BS81l0: Pa¡t 1: 1997. Generally, staircases should be designed in
accordance withthe requirements for beams and slabs given in Clauses 6.1.2 and
6 1.3 of the code. The specific issues related to the design of stai¡cases a¡e the
depth of section and permissible spar/depth ratio. For a staircase, the depth of
section should be taken as the minimum thickness perpendicular to the soffit of
tìe staircase. P¡ovided the stair flight occupies af lreasl 60%o of the span, the
permissible span/depth ratio calculated in accordance with Clause 7.3.4 (tltts
clause is for beams and slabs) may be increasedby l5%.
6.1 Foundations
For the design of pad footings and pile caps, the "rigid footing or cap"
assumption may be adopted for the determination of bearing sfesses or pile loads
(commonly known as rigid footing or cap design). If the flexibility of the footing
or cap is to be considerqd for the purpose ofreducing the bearing stresses or pile
loads ¿t the comers of the footing or cap (commonly known as flexible footing or
cap design), a rigorous and accurate elastic analysis method, such as the finile
element method, should be used.
Design ofpadfootings:
For the design of an isolated footing, the critical section may be taken as that at
the face of the column or wall supported. While calculating the design ultimate
moment, it should be noted that no redistibution of moments is allowed lf the
width ofthe section is relatively large, the tension reinfo¡cement for resistíng the
ullimate moment should be distributed with two-thirds of the total lequired This shear force is to be distributed over a width extending one effective depth on
concentating within a zone near the cente of the coh¡mn supported; otherwise either side of each pile centre or as limited by the actual dimension of the pile
the tension réinforcement should be uniformly distributed, as stipulated in clause cap.
67.2-2 of tlrc code. 'While designing for shear sfrength, it should be noted that Effecrive width:2 x (1200 + 1000) = 4400 mm
two shear conditions need to be considered: (1) shear along a vertical section 7500 x 103
extending across the full width of the footing; and (2) punching shear around Shea¡ shess = = 1.42 N/mm2
ffiä
concentmted loads.
Shear enhancement may only be applied to the stip of pile cap within I '5 x pile
diameter at each side of a pile.
Design of pile caps;
ttre ienãing design of a theorY if Since 1.5 x pile diameter = 1.5 x 900 = 1350 mm, the whole effective width is
the pile cap is considere the truss within the strip where shear enhancement may be applied.
analãgy ifihe pile cap is Pile caP, Shear strength v, : 0.45 N/mm2
the c¡itical sections slioul te¡ ofthe
Shear span 4" = 1500 - 500 - 0.3 x 900:730 mm
pile inside the face ofthe pile. The shear span a" should be taken as the distance :
Effective depth d 1200 mm
from the face ofthe column to the critical section for the shear being considered.
where due conside¡atiou has been given to the shear disl¡ibution across the
2d 2x1200
section. shea¡ enhancement (see Clause 6.1'2. the a!
-:-:5.¿t 730
pile cap of width equal to 3.0 times the pile In E¡hanced shear strength = 3.29 x 0.45: 1.48 N/mm2
considêration ofthe shear distribution across be
Since the shea¡ st¡ess is smaller than the enhanced shear strength, no shea¡
reinforcement is required.
Punching shear checking is also to be caried out (omitted he¡ein for brevity).

r 000 3000 1000

of the piles is greater than 3 0 times the pile diameter,
ile cap due to the concentrated pile loads should also be
requirements for flat slabs (see Clause 6 1'5'7)'

()
20% of
pile
dimeter
sketch ofthe pile cap layout is shown in Figure 6.7.
1000
T
Load on each pile = i5000/4 = 3750 kN
I
I
Bending moment at face of columt= 2 x 3750 x (1,5 - 0 5) : 7500 kN-m
M
-'-.
7500x10"
---5ooox l2oo'zx = n fì'ìO
bd2 L 35 ()
zld = 0.5+ ^lo.z5-+ = 0.966
'l]0.9
'lakez=0,95d= 0.95 x 1200:1140mm
Section Section
M Effeotive depth of
7500x 10" : 16439 mm2 pile cap = 1200 mm
for fo¡
0.87 f ,z 0.87x460x1140 bending shear
design dEsign
Use 34 Q25 bars; l, provided = 16690 t^'> 16439 mm2, O'K'
Shea¡ force a1 c¡itical section = 2 x 3'l 50 = 7500 kN Figure 6.7 Design of a pile caP

ot
 6.8 Beam-column joints

Since there have been no specific requirements for the desígn of beam-column
joints in the previous Hong Kong codes a¡d in the Brilish Standards, most design
engineers in Hong Kong are not familia¡ with how beam-column joints could be
better designed. Therefore, some basic tainíng for the Hong Kong engineers on
the design of beam-column joints may be needed.

The code states tïat beam-column joints should be designed to satis$ the
following criteria:
¡ at serviceability limit stâte, a beam-column joint shall perform at least as
vvell as the members that it joins; and
r at ultimate limit state, a beam-esl1¡¡1¡ joint shall have a design sbength column column
sufficient to resist the most adverse load combinations sustained bv the
adjoining members. Figure 6 8 Cracking and spalling at beam-column interface

To be more specific, the basic requirements of a beam-column joint are;

o at serviceability limit state, the beam-column joint should not crack The usual practice in the past ofnot putting in anyjoint shear orjoint confining
seriously and should not show any signs of spalling; and reinforcement (commonly known as empty joint design) is no longer considered
. at ultimate limit state, the beam-column joint should be able to maintain its satisfactory. However, putting all the reinforcement required by the new code
structural integrity so as to allow the full strength potentials of the adjoining together, the ioint could become quite congested with reinforcing bars. Fixing of
members 1o be developed. the beam reinforcemenl, the column reinforcement and the joint shear and
confining reinforcement at the joint would demand higher ski1l than before a¡rd
,.4
At the veftical interface between the beam atd the column, stress concentration render placing and compaction of tle conc¡ete in the joint much more difficult.
t occu¡s and vertical cracks often appear at the tension side of the interface, as Design engineers should be awa¡e of such practical difficulties arrd should pay
.li shown in Figure 6.8. If the tension reinforcement of the beam has not been particular attention to the various site supervision problems that may arise.
properly anchored into the join! the tension reinforcement tends to be pulled out Ho*"v"., all this trouble is necessary for the beam-column joints to maintain
fi of the joint and this could ¡esult in the formation of fairly wide cracks at the their structural integrity at the ultimate limit state.
ü beam'column interface (the crack width is actually the same as the bond-slip of
the tension reinforcement). Moreover, if the tension reinforcement is bent into The provisions in Section 6.8 of the code for the design of beam-column joints
r the column in such a way that the bearing stresses induced at the inside of the are largely based on the New Zealand Standard NZS3101 : Part I: 1995 Part2 of
Ï
:,ti
bend are only resisted by a thin concrete cover, tåe conc¡ete cover could spall off,
as depicted in Figure 6.8. Hence, proper detailing of the reinforcement going into
this standard (N2S3101: Pat2:1995), which is a comment¿ry handbook, has
provided some background info¡mation and explanations of the rationale behind
the joint is crucial for the high performance at serviceability limit state of a beam- the various design requirements for beam-cohunn joints. However, NZS3101 is
ffi
'u column joint. This is however easier said than done because there is often the not easy to follow partly because of the complicated behaviour of beam-column
r.l problem that the joint does not have sufficient space 1o accommodate the required joints and partly because there are many differences between New Zealand and
anchorage length of the reinforcing bars coming in from the beams. Hong Kong in the way the design loads and strengths are specified. For easier
¡efe¡ence and understanding, a brief introduction to the fi:ndamental theory of
Apart from cracking and spalling nea¡ the beam-column interface, the joint itself beam-column joints is presented in the following'
might also fail due to diagonal crushing or splitting. To avoid such kinds of
failure, it is necessary to: Joint Jorces:
. limit the joint shea¡ stress as per Equation 6.71 in Clause 6.8.1 3 of the code, There a¡e two types of beam-column joints, intemal beam-column joints and
. external beam-column joints, as shown in Figure 6.9. An intemal beam-column
provide horizontal joint shear reinfo¡cement as per Equation 6.72 in Clause joint is a joint that is connected 1o an intemal column and two beams at its two
6.8.1.5 ofthe code,
sides in the same vertical plane of the joint. An extemal beam-column joinl is a
. provide vertical joint shea¡ reinforcement as per Equation 6.73 in Clause joint that is connected to an external s de in the
6.8.1.6 ofthe code, and samevertical plane of the joint. Each to fo¡ces
. provide confining reinforcement to the length of column fonning the beam- ùom the membe¡s connected to it (i.e. strated in
column joint as per Clause 6.8.1.7 of the code. Fisure 6.10
 --
For a three-dimensional beam-column joint with beams framing into the joint i-n
two different vertical planes (i,e. not all the bea¡ns are in the same vertical plane),
the forces acting on the beam-column joint need only be considered in each
Column vertical plane at a time. ln other words, such a th¡ee-dimensional beam-column
joint has to be considered twice, each time as a two-dimensional beam-column
joint in a vefical plane oontaining the beams in the same vertical plane.
NZS3101 states t}rat the forces in each vertical plane may be conside¡ed
independently, but in actual fac! the horizontal joint shear foroe in one ve¡tical
plane could affect the joint shear reinforcemeût design in the other vertical plane
through the factor C, (see Clauses 11.3.5 and 11.3.6 of NZS3I01: Pa¡t I and
Clauses 6.8.1.5 and 6.8.1.6 ofthe new code).

The forces acting on a beam-column joint in each vertical plane are:

. axial force N, horizontal shear force Z" and bending moment M" from the
column above the joinl;
. axial force N', horizontal shea¡ force Z.' and bending moment M.' from the
colurrn below the joint;
Intemal beam- External beam-
. vertical shea¡ force Z, and bending moment M, from the beam at one side
column joint column joint ofthejoint; and
r vertical shear force /o' and bending moment M"' from the bea.n at the other
side ofthe joint, ifanY.
Figwe 6.9 Intemal and external beam-column joints

Horizontal and vertical joint sheor Jorces :

The bending moment Mo may be considered as comprising of a couple of tension
force T a¡rd compression force C, while the bending moment Mo' (if any) may
be considered as comprising of a couple of tension force Z' and compression
4ìy" force c', as illustrated in Figure 6'11' From axial equilibrium, T=C and
T'= C' . To ensu¡e that the joint would maintain structural integrity even when
the longitudinal reinforcing bars in the beams have yielded, the commentary of
NZS3101 suggested that both T and T' should be evaluated with the tensile
st¡esses in the reinforcing ba¡s assumed to be 25% higher than the yield st¡ess. In
other words,

T=C=I.254tly
T':C':r.25/4\,fy

Because oflhese horizontal forces induced by the bending moments acting fiom
the beams, at mid-height of the beam-column joint, there is a horizontal joint
shear force {, Biven by:

Vi -- T +C'-V.

VJ- At the cenfie of the beam-column joint, there is also a vertical joint shear force.
M. The vefical joint shear force (" may be derived from similar considerations.
ç/
For simplicity, the vertical joint shear force may be approximated as follows:

Figure 6.10 Member forces acting on a beam-column joint vþ: (h,/h,)vh

 -_-

plaaes containing the beams). If the horizontal joint shear forces in the trv,;r

horizontal directions oflhejoint are V,^ and. Vu,Íhen C, would be given by:
V
a = -----JL
Column ' V +V

Maximum allowable joint shear slress:

l¿\ M" To safeguard the core concrete (i.e. the diagonal compression stut in the
'l'
Plane l/" concrete) from crushing failu¡e, a¡r upper limit needs to be set to the joint shear
stress. The new code has imposed the following limit:
C'+
t. __-JL < (r)51
I
v;l
Beam 16 D)n'

in which å, is the effective joint width and å. is the depth of the column in the
direction of the horizontal shear being considered. In case the above limit is
vJ----- exceeded, regardless ofhow much shea¡ reinforcement is to be added to the joint,

M"'v the joint size (mainiy the column size) will have to be enlarged until the above
joint shear stress is kept within limit.

ConJinement.
Coh¡rnn In addition to the horizontal and vertical joint shear reinforcement, horizontal
transverse confining reinforcement also has to be provided, The horizontal
transverse confining reinforcement in beam-column joints shall not be less than
Figure 6.1I Intemal forces in a beam-column joint thatrequired by Clause 9.5 2, with the exception ofjoints connecting beams at all
fou¡ column faces in which case the transverse confining reinforcement may be
reduced to one-half of that required by Clause 9 5.2. In no case shall the link
Horizontal and vertical joint shear reinforcemenl: spacing ofthe fansve¡se confining reinforcement exceed l0 times the diameter
The horizontal and vertical joint shear forces are to be resisted by a truss of the smallest column bar or 200 mm, whichever is the smalle¡.
mechanism comprising of a diagonal compression strut in the concrcte and the
horizontal and vefical joint shear reinforcement. To ¡esist the horizontal joint Additíonal comments:
shear force, account may be taken of the contribution of the diagonal The new code has not provided any guidelines for the design ofbeam-wall joints
compression strut 10 shear resistance and the increase in concrete shear stength but actually in tall buildings, beam-wall joints are at least as important as beam-
due to the axial compression on the column. To resist the vertical joint shear column joints. In coupled shear walls, the joints bet\¡r'een the coupling beams a¡rd
force, account may be taken ofthe cont¡ibution ofthe diagonal compression strut the shear walls are often subjected to large vertical shear forces and concentrated
to shea¡ resistance and the counterbalancing effect of the axial compression on moments. As for beam-column joints, the reinforcing ba¡s from the coupling
the column. Taking into account tle contibution of the diagonal compression beams should be generously anchored into the walls (note that all the reinforcing
stmt and the eflecls of the axial compression on the column, the new code has ba¡s in a coupling beam. including those near the longitudinal axis, are subjected
provided the following formulas for evaluating the a¡ea of horizontal joint shear to tension). Moreover, since beam-wall joints could fail by shear sliding along
reinforcement l,n and the area ofverticaljoi¡t shea¡ reinforcement lj" : the beam-wall interfaces, the shea¡ sliding strength of each inte¡face should be
checked. For mo¡e information, the following Papers are referred 1o:
Vì C,N.I Kwan A.K.H. a¡d Zhao 2.2., "Tesling of coupling beams with equal end
"ih
": o.B7 lyr
-'-5_---r-
n
A"f^
I

rotations maintained and local joint deformation allowed", Proceedings, I C.8.,

)
Structures and Buíldings, Vol 152, February, 2002'pp67'78.
.
/=L:- o.4v; -ctN' 0.4(hblh,Yji-c jN' Kwan A.K.H. and Zhao Z.Z, "Cyclic behaviour of deep reinforced concrete
0.87 fw 0.87 fY coupling beams", Proceedings, I.C 8., Structures and Buildings, VoI.152, August,
2002,pp283-293.
lt1i, V,i and ll' of V,n, Vn and N
are the values Zhao Z.Z. and Kwan A.K.H., 'Nonlinea¡ behavior of deep reinforced concrete
state and allocating the beneficial eflect of the axial
C, is the factor for coupling beams", Slructural Engineering and Mechanics, Vol.15, No.2, 2003,
ppl 8 l-1 98.
compression to the two horizontal directions of the joint (or the two vefical

72
llor ked examples for design of beam-column joints. a¡rd ils tensile stess taken to be 25% higher than the yield stress. Based on this,
the tensile and compressive forces at the right side may be evaluated as:
T : C : 1.2sA\tí = 1.25 x2413 x 460 I 1000 = 1387 lù{
t On the otber hand, at the left side, the hogging moment of 400 kN-m would lead
to the following tensile and compressive forces:
Column size
Y - 5004oo*1.0"
fiø
'V"
900 x 900 K= bd'f""
:
x 640' x 45
=0.043

tr
3þ32
zld 0.s +
10.25
-ù = 0.es0

z = 0.950 x 640 = 608 mm

Mr'( Vr1 T, : c, : M - 4ooxlo6,tooo

=658kN
z 608

Neglecting I," which is usually quite small, the horizontal joint shear force rj¡
Beam size Beam size may be determined as:
700 x 500 700 x 500 Y; : r -r' : 1387-658 = 729 kN
(effective depth:640) (effective depth = 640)
The vertical joint shea¡ force V, may be determined as :

f, = 45 MPa \: : (h, th.)v; : (Tooleoo) x 72e = 567lll

fyt:460lvlPa The joint shear stress is checked below:

l!- < l-zsf- ( l'zsf^= 11.25 N/mm2), o.K.

b,h. Yr"lr::,
Figure 6,12 Design of an intemal beam-column joint = = 0.90 N/mm2
900 x 900

Consider the internal beam-columnjoint in Figure 6.12. The¡e are two possible
scena¡ros:
(a) the two beams at the two sides of the joint are both subjected to hogging
moments; and
(b) one beam at one side ofthejoint is subjected to hogging moment while the
other beam at the othe¡ side ofthe joint is subjected to sagging moment.

In Scenario (a), the worst loading case occu¡s when the hogging moment at one Ast =2413 mmz
side is maximum and the hogging moment at the other side is minimum. The
joint shear reinfo¡cement is to be designed so that structu¡al integrity ofthe joint
will be maintained until plastic hinges a¡e formed in the beams. However, if
plastic hinges are formed in both beams at the two sides of the joint, there will be
nohorizontaljointshea¡. Maximumhorizontaljointshearactuallyariseswhena
plastic hinge has been formed in the beam subjected to larger hogging moment
and no plastic hinge has been formed yet in the beam subjected to smaller
hogging moment, as depicted in Figure 6.13.

As an example, let the column be subjected to an axial load of 5000 kN, the beam
on the right be subjected to a hogging moment of 650 kN-m and the beam on the
left be subjected to a hogging moment of400 kN-m (all these a¡e the loads at the
ultimate limit state). A plastic hinge would be formed at the right side first.
Hence, at the right side, the top reinforcement may be assumed to have yielded Figure ó.1 3 Scena¡io (a) - both beams subj ected 1o hogging moments

75
 _-

Thenextstepistoevaluate C,. Ifthejointisatwo-dimensionaljoint(i.e there The vertical joint shear force V,i may be de termined as:

a¡e beams in one vertical plane only), then C, : 1.0. Ifthe joint is a three- h.)t/; : (700/900) x 1952 = 1518 kN
v,i = (h, /
dimensional joint (i.e. there are beams in two vertical planes), then the horizontal
joint shear in the two vertical planes yj* and V.y should each be determined and The joint shear st¡ess is checked below:

cj evaluated as vrl(Vn + z¡, ). Assuming in this particular case that the joint is
a two-dimensional one, the value of C, is evaluated as C, :1.0' Having
l¡ : '::?':::
b,h. 900
900 x
: 2.4r N/mm2 s o.zsr." ( o,zsr".: n.2s N/mm2), o K.

evalualed C,, the area ofhorizontaljoint shear reinforcement l,n is calculated as: The next step is to evaluate C, . Assuming in this particular case that the joint is
a two-dimensional one, the value of C, is evaluated as C, = 1.0. Having
evaluated C,,theareaofhorizontaljointshearreinforcement l,n iscalculatedas:

Use 4 legs of $16 bars (2 closed horizontal rectangular links equally spaced
v,l (05-g¿Ll - 1e52xl0'[o.r-l'0x5000xtor'l= ruzo..,
n = 0.87f,,1
vertically, i.e. one at l/3 height and the other at 213 height ofthe joint). ' A,l^/ 0.87x460 1 900x900v45 )
l" provided: 804 mm2 > 661 mm2, O.K. Use 6 legs of $20 bars (3 closed horizontal rectangular links equally spaced
The area of vertical joint shear reinforcement l" is calculated as: vertically, i.e. to be located at 1i4 height, 2l4height and 3/4 height ofthejoint)
1,, provided = 1885 mm2 > 1770 mm2, O.K.
..i,
a =-=
0.4v: -c,N' 0.4x 567 x l0r - 1.0x 5000 x 101
<0 The a¡ea ofverticaljoint shear reinforcement 1," is calculated as:
0.97 f * 0.E7 x 460
9.4¡t1 _C N. 0 4x 1518x 10r -l 0x5000x10'
Hence, no vefical joint shear reinforcement is required. <0
o*.7
'."
f
/ ra
087x460

In Scenario (bl, the worst loading case occu¡s when the hogging moment at one Hence, no vertical joint shear reinforcement is needed (note however that vertical
side is maximum and the sagging moment at the other side is also maximum' joint shear reinforcement may be needed when the axial load N' is small)
The joint shear reinforcement is to be designed so that stucû¡al integrity ofthe
joint will be maintained until plastic hinges are formed in the beams. Maximum
horizontaljoint shea¡ arises when a hogging plastic hinge has been formed in the
beam subjected to hogging momenl and a sagging plastic hinge has been formed
in the beam subjected to sagging moment, as depicted in Figure 6. l4 À¡
.,+- . -
As an example, let the column be subjected to an axial load of 5000 kN, the beam Sagging plastic /f\tø' Hogging plastic
on the right be subjected to a hogging moment of 650 kN-m and the beam on the hinge formed IÌ" hinge formed
left be subjected to a sagging moment of 250 lòI-m (all these are the loads at the
A,t:2413 mmz
ultimate limit state). At the right side, the top reinforcement may be assumed to
have yielded and its tensile stress laken ø be 25Vo higher than the yield stress.
C'+ 25A,t [\

Based on this, the tensile and compressive forces at the right side may be vt'
1
I
evaluated as:
l25A,2lr+ +C
T = C = r2sA,tû = 1.25 x2413 x 460 / 1000: 1387 kN

On the other hand, at the left side, the bottom reinforcement may be assumed to
have yielded and its tensile sfiess taken lo be 25V:o higher than the yield stress.
Based on this, the tensite and compressive forces at úie left side may be evaluated
as:

T' = C' : 1.2sA,,îy = 1.25 x982x 460 / 1000:565 kl'{

Neglecting I/. which is usually quite small, the horizontal joint shear force (i
may be determined as:
v,i : r *c' : 1387 +5ó5 = 1952 kN Figure 6.14 Scenarío (b)-one beam subjectedto
hogging moment and other beam subjected to sagging moment

SERYICEABILITY LIMIT STATES
7.1 General
limit state rnay be satisfied by:
The design requirements of sewiceability
¡ buildings", whichwe all should stive fo¡. Ahigh-perfomrance concrete building
following the deemed-to-satisf, rules in dimensioning ard detailing of the
structüal elements; or
. analysing the loading eflects and verirying that tåe deformalions and crack
widths etc are within acceptable limits (this is often referred to as direct
calculation in the code)
When analysing the structure for satisfying the serviceability limit
of the structure should be modelled.
requirements, the most likely behaviour
During analysis of stucture in ultimate limit state design, simplifoing
assumptions that a¡e on the safe side but not necessarily realistic are often made.
For instance, in ultimate limit state desiga, it is often assumed that the torsional
stifûress of frarne members, the out-of-plane stiffness of shea¡ walls and the
stifftress ofnon-structures are negligible. Such assumptions are conservative fo¡
ultimate limit state design but ate not necessarily so when applied to
serviceabilily limit state design. The existence in reality ofthe stíffness that has
been neglected couÌd significantþ affect the internal load distribution within the
stucture thereby causing unexpected cracking of certain structu¡al componeils
whose loading effects have been underestimated. Hence' during analysis of
sÎructure in serviceability limit state design, the structural model should be as
close to the reality as possible; no stiffiress of a¡y structual or non-structu¡al
component should be neglected if its presence couid affect the intemal load
distribution If the structural parameters of certain components, especially those
of non-structural components, are difficult to be determined accurately, then at
least the possible ranges shouJd be estimated and fhe analysis ofstructwe ca¡ried
out repeatedly using different combinations of possible values of the structwal
parameters so as to allow for the worst possible scenario in the design.
after decades of massive buitding development in many cities over the world' the
concrete. The autlio¡ has been serving as an independent consultant on concrete
materials and struchues for many years. In all the problematic cases that the
author has investigated, although the developers and the owners/tenants were
normal and should be acceptable is not entirely defensíble. In many cases, the
extent of cracking and the crack widths could have been contolled to be much
smaller if proper considerations had been given during the serviceability limit
state design.
The author has been advocating the idea of designing "high-perfomance concrete
is one that has, not just a high safety standard, br¡t also all round high
performance in ter¡ns of sen'iceability, functionality, .druability, structu¡al
¡¡ssr¡r¿¡nce rneasures,reparability, environmental friendliness and sustainability.
All in all, the autlpr holds the strong view that apart Êom safety, the other
perfonnance attributes such as those mentioned above, especially sewiceability,
should also be carefirlly aftended to in the st¡uctual design, as promulgated in the
state
following paper.
Kwan A.K.H., Au F.T.K. and Lee P.K.K., 'Iligh-performance concrete buildings
for the new millennium", Progress iD Structural Engineering a-nd Materials,
VoI.5, No.4, October-December, 2003, pp263-213.
When assessing the loads in the analysis of structu¡e for serviceability limit state
design, a distinction should be made between characteristic loads arrd qrpected
loads. Chancteristic loads are conservativs estimates of the loads that could
occu¡ with a relatively long retum period while expected loads a¡e the likely
loads that could occur every now and then wilh a relatively high tequency of
occlurence. Generally, for estimating tbe maximum response of the st¡ucture
under nonnal conditions (the code calls this ttre limit state calculation or
calculation to satisff a particular limit state), the characteristic loads should be
used, but for estimating the likely or average rcsponse of the stucture r¡nder
nonnal condilions (Îhe code calls this the best estimate calculation), the expected
loads should be used instead. To clariS why'lrnder nonnal conditions" is
emphasized above, the aulhor would like to sfiess that the estimation of
maximum response under abnormal or exteme conditions is part of ultimate limit
state design, not serviceability limit state design. For dead loads, the
characteristic and expected values a¡e the same and thus it will be sufficient to
just use the cha¡acteristic values in the analysis. For imposed loads, however,
both the cha¡acteristic and expected values may have to be considered separately.
When calculating deflection, it is necessary to assess how much of the applied
Ioad is permanent and how much is transitory. Dead load is 100% permanent.
But the propofion ofthe characteristic imposed load that should be considered as
state
pennanent is dependent on the usage of the building. The code suggests that for
still
gof normal domestic or ofüce occupancy, 25Yo of the characteristic imposed load
should be considered as permanent and for stuctu¡es used for storage, at lsast
75% should be considered as pennanent when the upper limit to the deflection is
being assessed. This suggestion is based on 858110: Pa¡t 2: 1985.
Wheu analysing the structure for serviceability limit state desiga, the elastic
modulus of ooncrete should be taken as the value given in Table 3.2 or evaluated
using Equation 3.1. The foot¡ote of Table 3.2 says that where the mean or
characteristic value of elastic modulus is required, the appropriate mean or
cha¡acteristic st¡eneth of the concrete should be used for the evaluation. For
79
- -
.ture under normal conditions, the
'çtìc concrete strenglh should be
'rf the structure under no¡mal
7.1 e expecled concrete strength
're is any uncertainty in the
e estimated and different
idea of the reliabiliw of
The code states, in Clause 7.2.1, that "cracking is normal in ¡einforced concrete
slructu¡es". This statement is actually exûacted from Eurocode 2 (there is no
similar statement in BS8110). However, our clients, the future tenants and the
general public are not neccssarily in agreement with this statement. Young
engineers may also misinte{pret this statement as suggesting that crack control is
unimportant Therefore, the author would like to elaborate and qualifl, this
statement as "minor cracking in reinforced concrete structures is normal but
extensive cracking should be avoided".

'e design, a linear The crack width limits are si Table 7.1 of the code Basicallv. for
rting the section prestressed members with lendons, thefrequent load combinatíon should
code. They are be considered and the k width should be limited to 0.2 mm. and for
be applied to all reinforced members prestressed members with unbonded tendons, the quasi-
rUOn is used to permanent load should be considered and the crack width should be
-u¡ate 10 eva the section limited to 0 róm (except water retaining structures for which a lower limit of
,.ron (the entire section, 0.2 mm to be applied).
calculation shows the 1o be
Lods of analysis are used in combinations to be considered, namely, frequent load combination and
rembers can be t¿ken into -permanent load combination, are adopted f¡om Eurocode 2 According to
:ifüress of the highly stressed parts Eurocode 0: Basis of Structural Design (also called EN1990), a frequent load
.c cracked tansformed section (the compression area combination comprises of the permanent actions and the frequent values of the
combined rvitlÌ the rei¡forcement on the basis of modula¡ va¡iable actions, whereas a quasi-permanent load combination comprises of the
pemanent actions and the quasi-permanent values of the va¡iable actions. The
frequent value of a variable action is the value that the total period of time for
which it will be exceeded is a small fraction of the reference period, while the
acking quasi-permanent value of a va¡iable action is the value that the total period of
time for which it will be exceeded is a large fraction ofthe referencs period (see
Clause 1.5.3 of Eurocode 0). They are both given as reduced values of the
General characteristic values; the frequent value is given as p, times the
c value (ry, < 1) while the quasi-permanent value is given as r¿,
There are three main reasons for controlling cracking of a
llmes characteristic value (yz, < 1). These terms are not entirely compatible
structure:
. with the used in the B¡itish Søndards. The author would like to
ensure proper functioning;
suggest frequent value of a variable load may be interpreted as the
o avoid impairing dwability;
expected high unde¡ normal condition (that could occur frequently and
o avoid causing unacceptable appearance. might be a small fraction of time), whereas the quasi-permanenl
value of a variable mav be intemreted as the expecled averase load under
Depending on the location and the lighting a crack of width greater normal condition (that be in place most ofthe time and could be exceeded
than 0.3 mm can be quite conspicuous even standing at a distance of for a substantial f¡action ime) In other words, they are in reality expected
several metes away. Hence, for aesthetic there is a need to contol the loads under normal condition h different frequencies of occurrence
crack width at not greater than 0.3 mm. ,É the cracking is extensive and at the
same time the crack widths are ¡elativel (> 0.3 mm), the tenants or owners
may feel uneasy about the safety
Annex Al of Eurocode 0 has the values of y, and ty,
building and start complaining (even
loads in buildings as in Table 7 1 values are not in line with what have
though from the stuctural s point of view, the building is still very
been suggested in Clause 7.1.3 3 ofthe code (this clause is actually the same as
safe). Hence, in order to safeguard ourselves from futr¡re complaints and to avoid
the conesponding clause in BS 8 1 1 0: Part 2: 1985). Basically, the values given in
imparting the bad impression to the public that the building has not been properly
Eu¡ocode 0 are marginally higher than those suggested in BSB 1 1 0: Part 2. As the
designed, it should be prudent to pay particular attention to crack conüol Failure
design and calculation methods in the new code follow those in BSBI 10, it is
to conl¡ol cracking would not in general cause any immediate safety problem but
considered more appropriate to adopt the values given in 858110 because the
would in the longer term create mistrust between the general public and the
design/calculation methods and the design values together form an integrated
structural engineering profession.
package and should be compalible with each other. Adjusting the values of g,
.t
 r

estimating the maximum response of the structu¡e under normal conditions, the
elastic modulus conesponding to the characterisfic concrete sùength should be
used, but for estimating the average rcsponse of thc stucture under normal
conditions, the elastic modulus corresponding to the expected concrele strength
(i.e. the mean stengtl¡) should be used instead. Ifthere is any uncertainty in the
elastic modulus, the possible raage of values should be estimated and diffe¡ent
values within the range tried in the analysis to obtain an idea ofthe reliability of
the calculation.

Regarding the rrethod of analysis foi sewiceability limit state design, a linear
elastic analysis method may be used. Three metlods of evaluating the section
stiftess of the members have been given in Clause 5.1.2 of the code. They are
all regarded as acceptable but a consistent method should be applied to all
membe¡s of t¡e structure. Wherc a síngle value of section stiffness is used to
cha¡acterise a member, it should be sufliciently accurate to evaluâte the section
stifûress based on the uncracked concrete section (the entire concrete section,
ignoring tåe reinforcement) even though the calculation shows the members to be
cracked. Where mo¡e sophisticated methods of analysis are used in which
variations in properties ove¡ the length of members can be taken into accoun! it
might be more appropriate to calculate the slifrress of the highly stessed parts of
the members based on the cracked ta¡sformed section (the compression area of
the concrete section combined with the ¡einforcement on the basis of modular
ratio).

7.2 Cracking

7.2.1 General

There are tl¡ee main reasons fo¡ contolling cracking of a reinforced concrete
sEucture:
. ensureprop€rfirnctioning;
¡ avoid impairing dì¡ability;
o avoid causing unacceptable appearance.

Depending on the loc¡tion and the lighting condition, a cack of width greaûer
than 0.3 mrn can be quite conspicuous even to someone standing at a distance of
several met¡es away. Hence, for aesthetis reason, there is a need to co¡tol the
crack width at not greater than 0.3 mm. If the cracking is extensive and at the
same time the c¡ack widths are relatively large (> 0,3 mm), the tenants or owners
may feel uneasy about the safety of the building and start complaining (even
though from the stuctural engineer's point of view, the building is still very
safe). Hence, in order to safeguard ourselves ûom futu¡e complaints and to avoid
imparting the bad impression to the public that the building has not been properly
designed, it shor¡ld be prudent to pay particular attention to crack control. Failure
to contol cracking would not in general cause any immediate safety problem but
would in the longer term create mistrust between the general public and the
structural engineerinþ profession.
IFr--*

and r¿, provided by Eurocode 0 slightly so that they become compatible with The formula for evaluating the su¡face c¡ack width has been given as follows:
BS8ll0, the values of r¿, and v/2 tobe used in the new code may be obtained as
surface crack wiOth: _jG9e_
inTable7.2. o- -..,,\
, *r(
\ å-' ./
Table 7.1 Values of y, and t¡/2 fbr imposed loads grven in Eurocode 0 in which ø". is the distance from the point being considered to the surface ofthe
Building area over which imposed load is applied V Vz nearest ¡ongitudinal bar, e, is the average strain at the level ofthe point being
0,50 0.30 considered, c," is the minimr¡m distance ftom the surface of concrete to the
Domestic o¡ office areas
0.70 0.60 sr¡rface of the tension steel (the cover to fåe tension steel), 'å is the overall depth
Congregation and shopping areas
Storage areas 0.90 0.80
ofthe section and ¡is the depth to neutral a¡ris.

Traftic areas, vehicle weight S 30 kN 0.70 0.60 This formú:la is applicable only when the strain in the tension reinforcemeut is
Traflic areas, 30 kN 3 vehicle weight < 160 ll,l 0.50 0,30 limited to 0.E Íy f E, , which is generaily satisfied at serviceability limit state.
Roofs 0.00 0.00
The average strain a. may be evaluated from the cury¿fure calculated as per
Clause 7.3.6 of the code based on the following assumptions:
Table 7 .2 Values of ø, and 2 for imposed toads to be used in the new code r plane sections remain plane;
Building area over which imposed load is applied Vt Vz ¡ the reinforcement is elastic with an el¿stic modulus of 200 kN/mm2;

Domestic or offrce areas 0.50 0.25

. the concrete in cômpression is elastic;

0.60
r the tensile strength of concrete is not negligible and there is a long-term
Congregation and shopping areas 0.70
tensile^ stress varying linearly ftom 0 N/mm2 at the neutral axis tõ 0.55
Storage areas 0.90 0.75 N/mm' at the level of the tension reinforcement within the tension zone.
Traffic areas, vehicle weight 3 30 kN 0.70 0.60 Altematively, as an approximation, the average stain may be evaluated by first
0.50 0,30 calculating the average strain of a c¡acked section with the tensile strength of
Traffic areas, 30 kN < vehicle weight < 160 kN
0.00 concrete neglected and then reducing this average strain value to take into
Roofs 0.00
account the stiffening effect ofthe concrete in the tension zone.

wbere it is exp€cted that the concrete may be subjected to abnormally high

7.2.2 Control of cracking without direct calculation (deemed-to-satisfy) shrinkage (> 0.0006), the value of e, should be increased by adding 50"/, of lte
expected sh,rinkage stain (this requirement is based on BSgll0: part 2: 19g5).
Clause 7.2.2 of the code suggests that if the deemed-to-satisS rules in the code F¡om clause 3.1.3 of the code, it can be seen that the sh¡inkage of concrete in
with respect to minimum reinforceme
usually no direct calculation needs to
Hong Kong is in general much higher than that of concrete in ur.
rne author is
at the moment conducting a fairly large number of shrinkage tests on typical
crack width limits, As an altemativ normal-strength concrete, high-strength concrete and high-performance con-c¡ete
direct calculation may be carried out to check compliance with úe crack width i1 Hong Kong. The tests ars not completed yet butjudging from the test ¡esults
limíts. According to Clause 7.2,1 of the code, in most cases, direct calcul¿tion to obtained so fæ, it may be roughly estimated that at a relative humidity of 50%,
check compliance with the crack widtb limits would atlow wider bar spacings to
be used, especially for shallow members. llf-:h4nk1C9 strain ranges ftom 0.0007 to 0.0009, while ar a relarive humidity oi
75%,1l,.e shrinkage strain rangcs from 0.0006 to 0.0008. It is now almost cefain
that the sbrinkage of concrete in Hong Kong is generally higher than 0.0006.
Hence, the value of e. will have to be increased by 50% of the expected
7.2.3 Assessment of crack widths
shrinkage strain. Assuming the expected shrinkage stain to be 0.0006, the value
of e. will be increased by 0.0003 (0.0006 x 50%). This increæe may amouû ro
Only the crack width ds to be assessed, The loads
to be applieit should combination (for prestressed more than 30% ofthe original value of e, aad is definitely not negligible. The
membeii with bonde anent load combination (for omission ofthe effect ofsluinkage in serviceability limit state design in the past
¡einforced members and prestressed members with unbonded tendons), while the is one of the ¡easons fo¡ having, from time to time, concrete cracking problemì
in
elastic modulus of the concrete should be the long-term modulus taken as one- Hong Kong. Inclusíon of this requirement in the new code wouljsignificantly
half of the short-term modulus' lmprove the crack control ofconcrete sûuctures.
 It should be noted that with üre effect of shrinkage taken into accoun! the
statement in Clause 7.2.1 of the code, which says that in most cases, direct
calculation 1o check cornpliance with the crack width limits would allow wider
bar spacings to be used, especially for shallow members, is no longer valid' In
fact,'with the effect of shrinkage taken into accorm! the crack contol
Ì
:r1
1 During the course of de
requirements of BS8ll
Railways h¿s been made
width limits, which a¡e
maximum allowable crack width va¡ies from 0.10 mm for exteme exposwe
the complete code ofpractice philosophy, including the loading_intensities to be
adopted, the material it*r"s ø be used, the concrete cover to be specified and
the ìolerable crack widths eÎc. It will be dangerous to single out the crack width
fomruta out of context. Crack contml is uot an exact science because cracking is
7.2.4 Earþ thermal cracking
Early thermal cracks are caused by internal or extemal rcstaints against the
thermal movement of the concrete as the temperatue of the concrete changes due
to the heat generated from the chemical reactions of the cementitious materials.
Internal restraint is the major cause of cracking in massive concrete structu¡es'
while extemal resûaiDt is the major cause ofcracking in concrete structures cast
against movement resûaints.
Under adiabatic condition, the þdration of cement would produce a temperature
rise of about l2oC per 1 00 kg/mr of cement regardless of the tlAe of cemeìt used.
Partial replacement of cement by pulverized ñrel ash @FA) could reduce the
âmount ofheat generated. The temperature rise due to the pozzolanic reaction of
PFA is dependent on the temp€råture, but is generally lowe¡ than that due to the
hydration of cement. For ins f 600C, the pozzolanic
reaction ofPFA would produce 5.5oC per lOb kg/m3 of
PFA. Mo¡e information on the rise can be found in the
following publications :
FitzGibbon M.E., "Large pours - 2, heat generation and control", Conctete,
10(12),1976, pp33-35.
Building Resea¡ch Establishment, Technícal Guide on Contol of Early Thermal
Cracking in Concrete, PSA Specialist Services, UK,1992, l4pp.
The temperature distribution within the concrete mass is dependent on both üre
heat generation from the concrete causing temperature rise and the heat
dissipation through the boundary su¡faces causing temperature drop and
temperature gradient within the concrete. Taking into account the various facto¡s
deærmining the heat generation and dissipatíon, the temperature distibution a¡d
variation with time may be evaluated for members with simple geometry using an
analytical melhod and for membe¡s with complex geomety by the furite element
method, However, even with fhe temperature dist¡ibution known, the evaluation
of the tensile stresses developed within the concrete is not an easy task. This is
partly because apart fum the extemal restaint against thermal movement, the
concreæ itself can also act as an intemal ¡estaint against differential thermal
movement arising from uneven temperature distribution and partly because the
restraint-induced sl¡esses a¡e dependent also on the changing elastic modulus and
creep characteristics ofthe concrete at early age. In fact, rhe mechanisms ofhow
the intemal and external restraints cause tensile stresses and thermal cracking are
often misuderstood, leading to inappropriate practices during curing that often
further aggravate the thermal cracking problem, as explained in the following.
Inlernal restraint:
Due to heat dissipation, the ûemperature ofthe concrete nea! the surfaces is lowe¡
than that of the concrete at the cenüe. The resulting tempe¡åture difference may
induce tensile stains læge enough to cause thermal cracking within the surface
zones when the tempelah¡¡e is rising and in the inæ¡io¡ when the temp€rature is
d¡opping, as depicted in Figure 7.1. When the temperature is rising, ihe conc¡ete
at the cent¡e has a higher temperature and iæ thermal expansion will induce
bursting pressure onto the outside causing the concrete withi¡ the surface zones
to be subjected to tension and possibly cracking. The themral cracks formed
within the su¡face zones can be observed after removal offormwork and are thus
well known. When the temperatue is dropping, ihe concrete nea¡ the su¡faces
will cool down to ambient temperature first while the concrete at the cente will
continue to cool down. Cooling of the concrefe in the interio¡ within a bigger
mass of concrete will induce tensile strains that may cause cracking. Since the
thermal cracks formed in the interior cannot be seen unless coring is carried out,
many engineers are not awa¡e of this problem. As the movement resfaint that
induces these cracks is from the concl€te mass itseE ít is called internal restraint.
 --

P r ovision of in s ula tion :

There is a common misunderstanding that insulation ís a proper way to deal with
thermal cracking, regardless ofthe site conditions. In some contacts, it is even
stipulated in the specification that double-layer insulation is to be provided in all
cases when curing freshly cast concrete. Actually, it may help to avoid thermal
cracking in some oases but may also aggmvate the problem in other cases,
depending on the type of restr¿int. If there is no external resEaint, then the most
tncmlôt likely cause of thermal cracking is intemal resùaint. In this case, insulation
c!açk crack
should be applied. lt could ¡educe the temperafire differe4ce within the concrete
(a) when temperature is rising (b) when temperature is droPPing and thereby alleviate the thermal cracking problem. However, it should be noted
that the insulation and also the formwo¡k should not be removed until the
temperatwe of the coucrete has dropped to near ambient temperature. Moreover,
Figure 7.1 Thermal cracks i¡duced by intemal restÉint
after removal offormwork, no water spraying onto the consrete surfaces should
be applied, It is not uncommon in Hong Kong t}at the formwork is removed
while the concrete is still hot and cold water is then sprayed onto the concrete for
curing, leading to temperatwe shock a¡rd immediate cracking. If there is extemal
Freshly casl shed wall restaint, tåen it is the thermal contraction of the concrete during temperatue
cooling dowo to
drop that will be the major cause ofthermal cracking. In such case, insulation
Mbient tçmpcrature
would increase the temperatue rise of the concrete and of course also the
subsequent temperatue drop, thereby aggravatiug the themral cracking problem.

-c Therefore, insulation should not be applied when there is any external rest¡aint.

v¡-- I Reducing the heat generation by changing the mix design ofthe concrete would
<,,' help but if the situation remains critical passive o¡ active cooling might have to
Existing shcu be applied. If there a¡e both intemal and external restraints, as in the case of a
wall at ambient
lemperalufE thick cóncrete member to be cast against existing stuctures that will resfiain its
thermal movemen! then it is not a question of whethe¡ to insulate o¡ not to
insulate. The only acceptable method is to reduce both the peak temperature of
(a) edge rest¡aint þ) end restraint the concrete and the temperature differential within the cotrcrete. One feasible
solution is to provide active intemal cooling so as to d¡aw out heat directly from
Figure 7.2 Thermal cracks induced by exlernal resüaint theinterioroftheconcrete. Loweringtheplacingtemperafureoftheconcreteand
reducing the heat generation ofthe concrete should also be considered.

External restrainl: Mix design:

Subsequent temperature drop after the concrete has reached the peak temperature
causes tlermal contaction. If the thermal contraction is ¡estrained by existing
structwes bonded to the freshly cast concrete, as shown in Figure 7.2, tensile
stresses will be induced and thermal crack, which are generally through cracks,
may be formed. As the movement restraint is from outside, it is regarded as an
extemal restraint. There a¡e two types ofextemal restlaint: edge restraint and end
n when a shea¡ wall is constn¡cted storey-
at the lower storey restrains the thermal
at one edge, while end restraint usually
occurs when a wall or slab is consfucted bay-by-bay in which case the existing
walls or slabs resftain the thermal movemenl of the freshly cast concrete at two
opposite ends, The autho¡ had once encountered a thermal cracking problem,
which occurred when a shear wall was cast using high-strength concrete' Du¡ing
curing, the tempeJature of the concrete reached a peak of around 80oC' After the
moulà was removed, fairly wide therrral cracks were observed. The conl¡aclor
had tried to resolve the problem by providing better insulation bur the problem
persisted and in fact was worsened due to reasons explained below'
It is advantageous for a cotrcrete stucfure to have a small temperah¡re rise during
curing. A small temperature dse ensures minimal thermal movement and reduces
the risk of early thermal cracking. In many cases, it is possible to alter the mix
proportion of the concrete mix without affecting its rheological properties and
desired compressive stength to reduce the heat generation of the concrete so as
to keep the temperature ¡ise as small as possible. This is especially so for today's
concrete, which usually contains a fairly high cementitious materials content.
The heat generation from a concrete mix is affected mainly by the quality and
quantity of cement, mineral admixtu¡es and chemical admixtu¡es, If cement is
the only cementitious material in the concrete mix, its hydration cont¡ibutes to all
tàe heat generated and the lemperatu¡e rise would be proportiona.l to the cemeut
content. The fineness of the cement may also affect the temperatue rise; a finer
cement hydrates and generates heat at faster rates and, with relatively little lime
to dissipate the heat, the resulting temperatue rise would be higher. Replacing
part ofthe cement by PFA can significantly reduce the amount of heat generated.
This is because the pozzolanic reaction of PFA does not start until the alkalinity
of the pore water is high enough to start the ¡eaction, which requires part ofthe

86 87
-
cement to hydrate for releasing calcilm hyd¡oxide into the pore wâtet. On the
other hand, the efflect of replacing part of ttre cenent by condensed silica ñune
(CSF) on heat generation is more complex. CSF provides nucleation sites for
hydration and would thus speed up the heat generation due to hy&ation of
cement. Mo¡eover, since CSF is much fine¡ than PFA, the pozzolanic reaction of
CSF is faster than that of PFA, As the increase in heat seneration due to addition
of CSF could outweigh the decrease in heat generation resulting ftom reduction
in cement content, the replacement of part of the cement by an equal weight of
CSF might not help to reduce the temperature rise. Lastly, the addition of a
superplasticizer can improve the rheology of the paste and reduce the water
demand of the conc¡ete mix, but generally has little effect on the heat generation
of the cemeniitious materials. Nevertheless, the addition of a superplasticìzer
does have an indirect effect on heat generation. Whilst the superplasticizing
effect does not seem to affect heat generation much, it reduces the water demand
of the concrete mix and thus with a superplasticize¡ added the paste volume of
the concrete mix could be ¡educed while maintaining the same workability. After
reduction of the paste volume, there should be pro-rata reductions in the
cementitious mate¡ials content ând the amount of heat that would be sene¡ated
Thermal analysis and control:
In the past few years, the autho¡ has investigated a nr¡mber of thermal cracking
problems in Hong Kong and found that in virtually all cases, the thermal cracking
was associated with hieh cementitious materials content of the concrete mix and
onc¡ete durinq curing.
ds 450 kglm" or the
the contractor should
adiabatic curing test or a TRET. Then, if the temperatue rise of the concrete as
measured by a true adiabatic cudng lest is higher than 45"C or as measured by an
acceptable TRET is higher than 40"C (TRET tends to yield a lower temperature
rise because ofheat loss), the contractor should be required to caÍy out thermal
analysis and submit a proposal for thermal control ofthe curing concrete.
7.3 Deformations
7.3.1 General conside¡ations

Adiabatic curing test: The¡e are several reasons for controlling deformation ofa structure:
Although simple nrles for predicting temperatue rise are available (such as . avoid causing unacceptable appearance;
BRE's lecb¡ical guide), they are only rules ofthumb and are not expected to be . ensì.Ee proper functioning;
accu¡ate. In fact, lhe tempetatu¡e rise is dependent on the material cha¡acteristics ¡ avoid causing drainage problems;
and the local conditions, and thus the only reliablo way of estimating the
. avoid causing damage to adjacent structurEs;
temperature rise is to garÏy out adiabatic curing test of the concrete mix using
exactly the same constituent materials. The purpose of the adiabatic curing test is
. avoid causing damage to non-strucn[es such as finishes/partitions/fixings.
to simulate the adiabatic condition (i.e. no heat gain or loss condition) inside a
Iarge volume ofconcrete and measure tïe temperature rise ofthe concrete unde¡
The deforrnation control requirements may be satisfied by either following the
deemed-1o-satisfy rules with respect to maximum span/effective depth ratio given
such condition. From the test, üre va¡iation of the concrete temperaturc with time
in Clause 7.3.4 without carrying out any direct calculation or by direct calculation
can be obtained, based on which the rate of temoeratu¡e rise at different times
of the deflection of the structue according to Clause 7.3.5 and, checking of the
after casting ¿nd the total temperatue rise mayL evaluated. These data a¡e
deflectíon results against the permitted values.
needed for finite element analysis of the tempetature and thermal stress
dist¡ibutions within the concrete, the outcome of which would help to design a
proper thermal contol scheme for mitigating early thermal cracking. Although To avoid impairing the appearance and general functioning of the stucture, the
permanent sag (the tolal long-term vertical deflection under quasi-permanent
the principles of adiabatic curing test are well founded, the¡e is still no generally
loads) should not exceed span/250. Pre-cambering may be used to compensate
accepted or slanda¡dized test method. As a result, the details ofthe test va¡y from
one laboratory to another. Some resea¡cheÍs use a computer contoiled for some or all ofthe permanent sag, but any pre-cambering should not be larger
than spar:r/250.
environmental chamber to simulate the adiabatic condition by automatically
adjusting the temperatwe of the chamber to be constantly equal to the
temperature at the centre of the concrete mass as measured by thermocouples
To avoid causing damage to adjacent st¡uctural or non-structural parts of the

buried into the concrete, while others just cure fhe conc¡ete specimen ãnd
sfuctue, the long-term vertrcal deÍIeclion after consîruclion under quasi-
permanent loads should not exceed spar/500. It should be noted that it is the
measure the temperature at the cenfre of the specimen using thermocouples
deflection of the structure after the adjacent paÍs are con¡ected to the structu¡e
without providing any kind of insulation or temperature contol. Frankly
(i.e. after construction) that causes damage 1o the adjacent parts; the deflection of
speakiDg, ifthe adiabatic curing conditions are not exactly simulated, the curing
the structure beforethe adjacentpafs are connectedto the structure has no effect
test is not a tue adiabatic cu¡ing test and should be called a temperature ilse
Hence, pre-cambering would not help to avoid causing damage to the adjacent
evaluation test (TRET). In Hong Kong, adiabatic curing tests a¡e often carried
parts. I1 should also be noted that although the dead loads are usually applied
out without exercising any temperatu€ contol; these should have been more
before the adjacent parts are connected to the st¡uctr¡e, the creep ofthe st¡uctu¡e
appropriately ¡efer¡êd to as TRET. However, the adiabatic curing test is more
under dead loads could contribute to deflection ofthe slructure after the adjacent
suitable for use in a well-equipped laboratory and for field tests ihe less
parts are connected to the structuÌe. In reality, fte adjacent parts could be
sophisticated TRET methods are preferred.

88
 connected to the structure at diffe¡ent times during construction. Therefore, if
considered necessary, the actual time of connection should be taken ínto account
in the deflection calculation.
7.3 2 Excessive response to wind loads
The response of tlte sûucture under wind loads may be assessed using either
static or dynamic analysis.
When carrying ovtstatic analysis,lhe static cha¡acte¡istic wind loads should be
applied 1o the building and then the lateral deflections at top and every storey of
the building evaluated. The lateral deflection at top of the building should not
exceed height/50O, while the relative late¡al deflection in any one storey should
not cause damage to tie pafitions, claddings and finishes etc. Depending on the
wind environment, the height/width ratio and the structural efficiency (in terms of
lateral load ¡esistance) ofthe building, this top deflection limit ofheight/S00 may
govem the design of a tall building. Hence, for a relatively tall building (more
than, say, 40 storeys high), every effo¡î should be made to maximize the lateral
stiffness of the building, e.g. by converting the non-structural padition walls into
sfuctural walls, aligning the shea¡ walls so that they can be coupled together,
increasing the depth ofthe coupling beams so as to increase ÎÏe effectiveness of
the coupled shear walls, incorporating out-rigger trusses/beams to make good use
of the peripheral columns, putting in diagonal bracing members if practicabie,
and reducing the shear lag of any framed tubes etc, On the other hand, if the
relative lateral deflection in any one storey is likely to cause darrage to the
partitions, claddings a¡d finishes, the mountings ofthe partitions, claddings and
finishes should be detailed to allow for such detbrmation.
Where a $tnamic analysis is undefaken, wind speeds based on a lO-year retum
periodofl0minutesdu¡ationshouldbead-opted. Theresultingpeakaccele¡ation
ãfth"boildir,gshouldnotexceed0.l5m/s2ioraresidentialbuilding or0-25m/s2
for an office or a hotel building If dampers are incorporated to reduce the wind-
induced vibration, the design should be suppofed by dynamic analysis following
specialist advice,
It is interesting to compare the above peak acceleration Iimits of 0.15 m/s2 and
0.25 m/s2 in co¿e to the corresionding acceleration limits for diflerent
ihe
comfort factors given in the explaaatory notes to the Chinese Code JGJ 3-2002,
which are summarized below:
Comfo¡t factor Accele¡ation limit (in terms of e and m/s')
sensitive occupants will feel disturbing, while the peak acceleration limit of0.25
m/s2 falls well within the ra¡ge that average occuPants will feel disturbing.
7.13 Excessivevibration
Excessive vibration may cause discomfort or ala¡m to the occupants. 1o avoid
causing discomfort or alarm, long-span floor/footbridge structues should be
analysed for thei¡ dlnamic behaviour under fluctuating loads so as to ascertain
their accepøbility in tenns ofvib¡ation control.
If there is any machinery generating vib¡ation that could cause nuisa¡ce to the
occupants, the machinery should be isolated so that its vib¡ation would not be
transmitted 1o the floor.
lf there is, any sensitive equipment whose functions could be affected by
vibration, it should be easier to isolate the equipment ftom the floo¡ vibration
than to ¡educe the vibration ofthe floor.
7 3.4 Limiting deflection without direct calculation (deemed{o-satisS)
In normal circumstances, provided the deemed-to-satisfr limits given in the code
with respect to span/effective depth ratios are complied with, no di¡ect
calculation ofthe deflection is required. However, iffor some reason, deflection
limits more stingent than those stipulated in the code a¡e to be enforced, more
rigorous checks would be necessary.
Table 7.3 of the code gives the maximum limirs of span/effective depth ratios to
be applied to rectangular beams, flanged beams and solid slabs with different
support conditions for complying with the permanent sag and deflection after
constuction requirements in Clause 7.3. l. No allowanco has been made for any
pre-camber in the derivation of these limits.
The deemed-to-satisff maximum limits of span/effective depth ratios in Table 7.3
are fairly conservative and may be exceeded if direct calculation of deflection has
been ca¡ried out to verify compliance wilh the requirements in Clause 7.3. 1.
If it is not necessary to limit the deflection of the structue after construction (i.e.
if the deflection of the structure affer construction is unlíkely to cause any
damage), Table 7.3 may be applied to spans longer than 10 m without any
modification. But if it is necessary to limit the deflection of the stucture after
construction, the values in Table 7.3 should be multiplied by the ratio lO/span
(the span should be expressed in m) before applying to spans longer than 10 m,
except for cantilevers where the design should be justified by direct calculation.
To allow for the effects of the amount of tension ¡einforcement (which may be
estimated from úe design ultimate moment) and the service stress (which may be
estimated from the ratio of area of tension reinforcement required to a¡ea of
tension reinforcement provided), the maximum limits of span/effective depth
ratios in T¡ble 7.3 of the code should be multiplied by the modification factors
given in Table 7.4 of the code. The modification factor va¡ies from 0.76 for a
heavily loaded member to 2.0 for a lightly loaded member.
 -i'ií¡dú'!ffæ

To allow for the effect of the amount of compression reinforcement (expressed in

steel area ratio provided), the maximum limits of spar/
Table 7.3, after being multþlied by the modification
.4, should be fr¡rther multiplied by the modification
factors given in Table 7.5.

The code s The above methods are applicable only to relatively simple cases In more
high or if complicated cases, the finite element method may have to be used,
span/effect
mentioned
be delermined. The author would like to propose tha! in such case, the deflection 7 3.6 Calculation ofcurvature
control should be exercised by direct calculation ofthe deflection of the structure,
taking into account the creep/shrinkage of the concrete and any abnormally The curvature of a section may be calculated based on the following assumptions
adverie conditions, ald checking the deflection results against the perrrissible Cracked section;
defleclion limits. unfortunately, the shrinkage of conc¡ete in Hong Kong is quite . plane sections remain plane;
high and it seems tlnt we cannot avoid this problem. . tïe reinforcement is elastic with
an elastic modulus of 200 kN/mm2;
. the concrete in compression is elastic;
There is actually another problem that may be more serious. The elastic modulus ¡ the tensile stress develo¡ed in the conc¡ete varies from 0 N/mm2 at the
of the concrete in Hong Kong is relatively low and thus the maximum span/
neut¡al axis to I -0 N/mm' instantaneously or 0.55 N/mm2 in the long-term at
effective depúr ratios given in Table 7.3 ofthe code, which a¡e based on BSSl l0: the level ofthe tension reinforcement within the tension zone
paff | lggi, may have to be adjusted to reflect the difference in elastic modulus. Uncracked seclion
Further research on this topic is needed' o plane sections remain plane;
¡ the reinforcement is elastic with an elastic modulus of 200 kN/mm2l
. the concrete in compression and in tension is elastic;
7 .3.5 Calculation of deflection . the maximum tensile stress developed in the concrete is not larger than 1.0
N/mm2.
For deflection conÍol by direct calculation, only the long-term deflections (the
total deflection and the deflection after constnrction) under the quasi-permanent In assessing the long-term curvatufe, the creep effect has 10 be taken into account
load combination need to be evaluated and checked. The quasí-permanent load by using the long-term elastic modulus ofconc¡ete in the analysis The long-term
elastic modulus may be taken as ll(1+ø) times the shoft-term elastic modulus.
where / is the appropriate creep coeffrcient.

The shrinkage curyature ofthe concrete should be added to the curvature due to
dead and imposed load. It may be calculated using the following equation:
realistic structual modelling of the non-stnrctures, allowing for creep effects and
allowing for tension stiffening of cracked concrete etc. t _ po¿cs
1", d
The defleclion of a member subjected to flexure may be calculated f¡om the
crrrattlle using the following equrltion:
where /r"" is the slrinkage curvature, e". is the shrinkage strain, p" is a
coefficient depending on *re tension and compression steel ratios and d is the
ô2a-7 effective depth of the section. The values of po are given in Table 7.7 of the
^) rx
dx- code. It is noteworthy that the addition of mo¡e tension reinforcement misht not
where ¿ is the deflection at x and is the curvature at
/1 x.
The deflection a help to reduce the shrinkage curvatwe but increasing the effecdve depth wòuld.
sections along the
may be obtained by calculating the curvature fr* at successive
From Figure 7.2 of tLrc code, it can be seen that the total long-term curyature
rnembe¡ and solving the above equation by numerical integration. Altematively, consists of th¡ee components :
the following approximate formula may be used: (1) the long-term curvature due to the permanent load;
(2) the shrinkage curvature; and
a: K! (3) the instantaneous curvatue due to increase ofpermanent loadto total load.
tt
 v-

bend ¡adius is not less than the respective limiiing

REINFORCEMENT : GENERAL REQUIREMENTS value
code and one of the following conditions is satisfied,
the
de the bends needs not be checked:

' tre anchorage ofthe bar does not require a renglh mo¡e tha¡r 4/past
the end
8.1 General ofthe bend;
normal steel ' the ba¡ is assumed not to be stressed beyond apoint 4þpastthe end
The reinforcement detailing rules given herein are applicable only to of the
bend at ultimate limit søte; or
reinforcement,welded fabric andlrestressing tendons. They are not applicable . there is a cross ba¡ ofdiameter > /inside the bend.
to painted, epoxy or zinc coated bars.
otherwise, the design bearing stess ìnside the bends should be checked to avoid
applicabte only to normal building structrues
concrete failure. The design bearing stess should be checked using the
icloading. They are not applicable to structwes/ following
equation:
loading, impact loading, machine vibration and
fatigue.
bearingstress=4= r'f* -
rQ It*z!)
For reinforcing bars to be fitted within two concrete fa
ion
\ ¿ul
on the bendin! schedule should be determined as the
the

concrete minus the nominal cover on each face minus

ion where 4r is the tensile force Èomultimate loads atthe start ofthe bend, r is the
given in Table 8 1 ofthe code internal ¡adius ald ao is tüe cente-to-cente distance between the ba¡s or the
cover plus bæ diameter, whicheve¡ is the smalle¡.

Spacing of reinforcement
8.4 Anchorage of longitudinal reinforcemett
and arrangement ofthe bars should be such that lie concrete can
be
The spacing-compacted
placeå and properly a¡ld the bars are adequately bonded or ancho¡ed' All longitudinal reinforcing bars, wires or welded fab¡ics should be properly
anchored at each of their ends to transmit their forces to the concrete-wiihout
Theclearhorizontaldistancebetweenindividualpafallelbarsandtheclear causing longitudinal splitting or spalling. Longitudinal splitting or spalling may
less than
vertical distance between horizontal layers ofparallel bars should be not occur because the bond stess developed aÎ úe barlcóncrete inte¡face -coulä
the maximum of the following: onto the surrounding conc¡ete thereby subjecting the
. bar diameter; I tension. Ifnecessary, transverse reinforcement should
. maximum size of aggregate + 5 mm;
kind of failure.
o 20mm
At any sect at both sides ofthe section
Bars positioned in separate horizonøl layers should be located vertically
above
by an appr chorage. provided this is
undemeath
each other so as to allow access for vibrators to compact the concrete done, local
the bars.
The anchorage bond st¡ess fo may be assumed to be uniformly dist¡ibuted ove¡
the effective anchorage length and taken as the force in the bar divided
bv the
Pernissible internal radii for bent bars effective anchorage area, as given by the following equation:

The minimum internal radii for bent bars should be such that: f = 's
. the ba¡s would not fail by having bending cracks; rôlt
. the concrete inside the bends would not fail because of high bearing st¡ess where .Fl is the fo¡ce in the bar and /o is the effective anchorage length.
acting ftom the bars.
The anchorage bond stress
fr should not exceed the design uitimate bond shess
[" given by:
fo": P,[h
the manufacturer.
where / is a coefficient dependíng on bar type and
f" is limited to 60 N/mm2.
 .T

For bars in beams where minimum links have been provided in accordance with 8.5 Anchorage oflinks and shear reinforcement
Table 6.2 ofthe code, the values of p may be taken from Table 8.3 ofthe code.
For bars in beams where minimum links have not been provided in acco¡dance Each end of a li¡'k or shear reinforcement shourd be ancho¡ed
following means: by one of the
with Table 6.2 of the code, the values of / should be taken as those appropriate
to plain bars irrespective of the bar type. For bæs in slabs, the values of p may
. passing bar of at least the same bar size to fo¡m
or hook a bend
be taken from Table 8.3 ofthe code. It should be noted that although the code certain mi¡imum length beyond the
. bend, o¡
says that Table 8.3 is for bars in tension, the table is actually applicable to both
rveldins two perpendicular bars (the p".p""ai.uf..
nray be ì ,e¡se).
U".,
bars in tension a¡d bars in compression.

The delailed requirements for the bends/trooks

The required anchorage length may be calculated using the following equation: or werded bar anchorages are
shown in Figure g.2 ofthe code.
t=fsó
.o
4fr,
is the ultimate stress developed in the bar taken as 0.87 fr. Substituting 8.6 Anchorage by welded bars
where fi
the values of fr and /0" into the equation, the required anchorage lengths, Additional anchorage be provided by welding üansverse
cross bars) bearing on ffy
expressed as multiples of bar diameter, may be obtained as tabulated in Table 8 4 bars (also called
the concrile, as shoin in Figle S.3
of the code. Note that although the equations used a¡e the same as those in of the code.
BS81 10: Pal. l: 1997,since the material safety factor of steel is 1. 15 in the new If there is one such welded transverse bar, the anchorage
code but is 1.05 in BS81l0, the required anchorage lengths in the new code are calculated as:
capacity fi,o may be
not the same as those in BS81 10. As a rule of thumb, an anchorage lengrh of 40
times ba¡ diameter should be sufficient in most cases. 4,¿ : /,¿ /, o,r, but not g¡eater than _q*d
in which /,u is the design rength oft¡ansverse
bar, d is the diameter oftransverse
Bends and hooks a¡e effective anchorage fo¡ bars in tension but a¡e not effective bar, o,o is the conc¡ete stress and
for bars in compression. Bearing failure of conc¡ete inside the bends should be d,o is the design shear strength of weld.
prevented by ensuring that the minimum intemal bend radius complies with the
If there are two such welded tra¡sverse bars, one on
requirement in Table 8.2 and that either the anchorage does not require a length anchored, the capacity given by úe above
each side of the bar to be
more than 4{past the end ofthe bend or the bearing stress inside the bend equation ,lro,ri¿ ¡. doubled.
satisfies the requirement ofEqualion 8.1. If the¡e a¡e two such werded tr¿nsverse bars, on the
same side of the bar to be
anchored with a minimum spacing of
Regarding the effective anchorage length ofa bend or a hook, both BS81l0 and 3 fimes the ú. ,;;;, the capacity should
multiplied by a factor of 1.4 be
Eurocode 2 have provided some guidelines. Herein, it is suggested to follow
BS81l0. According to Clause 3.12.8.23 of BSSl l0: Part It 1997, the effective
anchorage length of a bend or hook is the len$h of a straight bar which would be 87 Laps and rnechanical couplers
equivalent in terms of anchorage 10 that portion of the bar between the start ofthe
bend and a point 4 times lhe ba¡ diameter beyond the end of the bend. This Axial force may be transmitted from one
effective anchorage length may be ta.ken as: bar to the next in the same directíon by:
r lapping ofbars;
For a 9f bend: either 4 times the intemal radius of the bend with a maximum of . welding (not forjoints subjected to cyclic loading
l2 times the bar diameter, or the actual length ofthe bar, whichever is greater. r to avoid fatigue failure);
using a mechanical coupler.
Fot a l80o hook: eithe¡ 8 times the intemal radius of the hook with a maximum of
24 times the bar diameær, or the actual length ofthe bar, whichever is greater. lle requirements for mechanicar coupre;rs
have arready been stipulated in section
If the intemal radius is at least 3 times the bar diameþr, then the effective ;1^r:.,I:':.:
only,the requiremenrs f";
ñ;;d".;ïä,, secrion. Mosr of rhe
anchorage length should be the maximum of (12 times the bar diameær or actual requrremenls for raps in new cocre ar: aaoptla from Èuroco¿e
Part 1: 199i. '.e 2 and BSg110:
lengfh of bar) for a 900 bend or the maximum of (24 times the ba¡ diameter or
actual lenglh ofbar) for a 1 80o hook.
The detailing of laps between
bars should be such that:
For bottom bars in simply supported beams, no bend or hook should begin before I t¡ansmission offo¡ces from one bar to
the next is assured;
the centre of the support or before half of the effective depth ùom the face of the t spalling ofconcrete cover near the lap
does not occur; and
support (see Clatse9.2.l.7 of the code). ' longitudinal cracks along the ba¡s do not
occur.

I
 -
sriixlt&åij;e"'""=*-'"-

Main bars in compression and secondary (distribution) ba¡s in tension or

compression may be lapped in one section, However, main baIS in tension musl
not ùe lapped in one section unless the requirements stipulated in Figure 8.4 of
the code are complied with

The requirements stiPulated in Figure 8.4 ofthe code are:

. clear transverse distance between two lapping bars should not be greater
thøn 4þ or 50 mm; otherwise, the lap length should be increased;
. longitudinal dista¡ce between adjacent laps should not be less than 0.3 times
the lap length; and
. clear distance between adiacent bars in adjacent laps should not be less than
Compression lap length;
the main bars in tension and in one layer may be
despite satisfaction of these requirements, main _be_at
least equal to 1.25 times the design
ers should have at most 50% ofthem lapping at he lapped ba¡s a¡e of unequal size, the lãp
any one section.
Tr ansv e r s e r e info r c e m e nt
Minimum lap length:
The minimum lap length should be:
for ba¡ reinforcemenf: not less than 15/ or 300 mm;
for fab¡ic reinfo¡cement: not less than 250 mm

Tension
The ten d be at least equal to the design tension anchorage
lenglh. ofunequal size, the lap length may be based on the
smaller wing conditions, the tension lap length needs to be

(c) adjacent laps is less than 6þ ot 75 mm, the lap

1.4 times the anchorage length; E.8 Addition¡l rules for large diameter bars
(d) apply, the lap fength should be increased to 2'0
times the anchorage length;
(e) ifboth conditionsla) and 1o¡ uppty, the lap length should be increasedto 2'0
 T-
-PT--

'When
large diameter bars are used, crack contol should be achieved either by a staggered manner as shown in Figure g.g
of the code. Basically, the second
using swface reinforcement or by direct calculation (as per Clause 7.2.3 of the should be curtailed at a distance ofìt reast
I .3 lo from the end of the first
ba¡
code). The code refers to Clause 9.2.4 fot tlrre design of the su¡face reinforcement ba¡ and
the thi¡d ba¡ shourd be cufai.red at a dislance
but there is no Clause 9.2.4 in the code. The code is actually referring to Clause
åf at reast /b f¡om the end of the
second bar (in this case, the value of /o
9.2.4 of Eurocode 2. Acco¡ding to Clause 9.2.4 a¡d' Annex J of Eurocode 2' should be based on the diamete¡ of
an
individual bar).
surface ¡einforcement should be provided to resist spalling wherever the main
¡einforcement is made up of large diameter ba¡s. If provided, the surface
reinforcement should consist of wire mesh or small diameter bars and be placed
Compression anch lsars;
outside the links. The area of surface reinforcement should not be less than
For compressiorr ¡dle of bars, regardless oi its
diameter, may be equivalent
0.01 ,4"r,.n in the direction perpendícular to the large diameter bars a¡d 0.02 A.,."*l ståggering. For a bundrc or. bars rvith an
equivalent diamete¡ greate^r than or t:qua-l-to i) --, ut least io.¡ ri*. nrrrg
parallel to those ba¡s diamere¡ ofnot ress than r2 mm shouJå l:eprovii"J"tl""o
encr ofthc bundle and"
should be^provided jusr U"yona ,n" of rhe bundle. The fir,e
Large diameter bars should be anchored with mechanical devices. As an i-L3_.^r.l"k
¡ttu<s are to serve as confining ",ìisplitting.
reinforcement to prevent
altemative, they may be anchored in the fonn ofstraight ba¡s, but links should be
provided as confining reinforcement 1o prevent spliuing' Lapping ofbundled bars
For a bundle consisting of not more tha¡
In the anchorage zones of the large dìameter bars, if lransverse compression is diameter less tha¡ 32 mm. the bundl: of ba¡s
not present, in the form of links, additional to that (in this case, the equivalent diameter shoLrld
b
provided for provided. For straight anchorage lengths, Otherwise, the lapping should be staggcred i
the additiona should not be less than the following: least i.3i" (in this case, the value of l" .s
. inthe directionparalleltothetensionface, .{rn :0.25 A,n, individual bar). A fourth bar of tåe same s
rhe diametc:: .ri an
rh¡,- rL-
¡ inthe directionperpendiculartothetensionface, A', :0-25 A"n, lap without cranking rhe bars or shifting ';i',ililåÈ:'::
The additional tansverse reinforcement bars should be wriformly disûibuted in shown in Figure 8.9 of the code At ar,/t hould nor be nrore
the anchorage zone and their spacing should not exceed 5 times the diamete¡ of than fou¡ ba¡s For this reason, a buncilc
ofmore than three bars srrol¡rcr nor he
the large diameter bars. lapped.

Generally, large diameær bars should not be lapped, except in sections with a
minimum dimension of 1.0 m or \\¡here the st¡esses in the bars are not greater 8.10 Prestressing tendons
than 80% of the design ultimate stength. For large diarneter bars, the use of
mechanical couplers should be prefened. Layout o/ pre +e n s io ne d tendons
The minimum crear horizontar spacing of pre{ensioned
tendons should be not
less than the maximum of the foilówins:
8.9 Bundled bgrs . tendon diameter;
¡ maximum size of aggregate + 5 mm;
In a bundle, all the bars should be of the same type and grade. The bars may be . 20 mm.
ofdifferent sizes provided the ratio ofdiameters does not exceed 1 7. The minimum clear vertical.spacing ofpre_tensioned
' tendons should be not less
than the maximum of the following:-
A brmdle ofbars rnay be treated as equivalent to a notional bar having the same . tendon diameter;
sectional area. The number of bars in the bundte is limited to a maximum of 4
for vertical bars in compression and bars in a lapping joint or to a maximum of 3
t test results have shown satisfactorv
for all other cases. Moreover, the equivalent diameter is limited to not greater
;r aachorage of the tendons, concrete
than 55 mm.
of conc¡eÎe and placing of concrete
rns should not be allowed within the
rwn that there are no problems with

nored thar rhe above minimum .,".,

,0".;;1:'å1.:l*:.T:"t:::
those stipulaled in Eurocode 2 and ^,t:,:ll"r:'*::
d'ring field applìcations whethe¡ the above
.;" ,h;;i; be exercised
therefbie to check
."quir"*.or, sufficient to ensure
satisfactory perfonnance. ".e

100
 f.I: i r;
i- iiï

Layou 1 oÍ po s |-tens i one d ten dons : equation may be used !o estimate

the ha¡smission lengtå /, for
tle minimum clea¡ horizontal spacing of post-lensioned tendons should be not prestressing forces ttp to
75%o of the ch¿¡acteristic -initial
less tha¡ the maximum of the following:
stength of the tenàons.
¡
.
r
duct diameter;
maximum size of aggregate
50 mm.
f 5 mm;
in which
#
/. is the conc¡ete strergthâttransfer,
The minimun clear vertical spacing of post-tensioned tendons should be not less te¡don and K, is a coeffìcient depending
/ is fhe nominat diameter of the
than the maximr¡m offte following:
presenæd in the code, it can
* til;; of tendon. F¡om the varues
o duct diameter; be seen t'at (, vad;; ûom 240
or super stand to 600 for plain or for 7_wire standa¡d
o maximum size of aggregate; indented wi¡e.
r 40 mm.
An
ducts
used,
on of lh
a
'!":"#:*

r bursting ofthe side cover perpcndicular 1o the plane ofcurvafiue;

r spalling ofconcrete around the anchorage;
and
¡ bursting oflhe cover in the plane ofourvature; tock.

Table E.7 ofthe codc.

Anchorages of pre-tensíoned lendons :

The desip ofthe anchorages should be such tlrat:
{ . sufficient transmission lengths are developed to avoid bond-slip failu¡e; and
¡ longitudinal splitting of is avoided'
The lransmission length of a
. type, size and su¡face condition ofthe tendon;
. degrce ofoompaction and stength ofthe concrete; and grcqulrements:
r location oftendon.
'h11 those for
stand both longifudinal and ha¡sve¡se
Transnission lengths for tendons near the Ûop are generally greater rd t¡ansmit ttresã forcesìo thì;*;#;
identical tendons ncar the bottom. This is due to sedimentation and bleeding of
o the deviato¡ shourd have a suitable
radius of curvatu¡e to ensu¡e that
tendon within the deviaro¡ would nor;;;;y;;^tussing the
or damage ro
-lr,

DETAILING OF MEMBERS AND PARTICI'LAR RULES

9.1 Ge¡er¡l

The detailing requirements in this chapter supplement those in chapter 8 and

have to be complied with at the same time.

9.2 Be¡ms

9.2.1 Longitudinal reinforcement f continuous beams, the

in the
åî
"f-tt"
r;;i;;;;Ë;
m"y ¡e .prr"¿ ouã,
jìro;

the flanges are dispersed into the flanges. ' beyond the point at whioh it
is no longer needed to ¿ssist
in resisting bending
momenr for a distance equal ro
tlr.
nor the a¡ea of sûengrtr (0.871);or "f.h";;rì;ä appropriare to irs design
ofthe bar sizes
ection. ' to the point where theìesign shear
capacity of
-' the
twice the design shear forcrãt
ü,"t r".tiär,; ;
- section is greater than
For the purpose of craok control, longítudinal bars should be added to the side
¡ ro rhe point where
."lr,nuing pasr that point provide
",_n:l !1: double the
faces of beams exceeding 750 mm overall depth. The minimum size of the area required to resist the design bendiniÅ"_*t that section.
longitudinal bars in the side faces should be lF¡øwhere s, is ttc ba¡ "t
uld lead to
spacing and ö is the breadth of the section or 500 mm if D > 500 mm' The ba¡
shor¡Id be
spacing ofsuch longitudinal bars should not exceed 250 mm' in heavily

Also for the purpose of crack control, the clesr distance between bars or groups
of bars near the tension face should be limiæd by: support, ztleast 50Vo oftl¡e calculated
be extendcd to the end supf;;;
clear spacing .7oog h < 3oo mm
¿y
mes the bar diameter beyond
the centre
/o is the moment redistribution ratio. Alternatively, the clea¡ spacing may hould begin before the
where ..rrt , oiã,
be assessed by:
diameter beyond a distance
clear spacingS 12999 g3g9*rn
;1i,"ã"S:,ï3ffil:;l
in which Lis the estirrated sewice sfiess in the rei¡forcement. The distance
supporr of a conrinuous beam, at
betweon tbe face ofthe beam and the nea¡est longitudinal ba¡ in tension should 1l,T_t-"t:T:*are
mrd-span bottom ¡einforcement should be ,ontioìå*
leasr 30% of the calculated
not be greater than half ofthe clear distance determined abovs' tfräugh the support.

r04
T

.lVhereacontinuousbeamorslabisextendedbeyondtheendsupporttoforma
should be
cantilever extension, the top reinforcement ofthe cantilever extension
e*ended back to beyond the point ofcontmflexure in the adjacent span'

Allcompressionreinforcementinbeamsshou]dbeproperlycontainedbylinks
passing ,ouna each comer ba¡ and each altemate ba¡ in an outer layer
having an
for the case ofcompression ¡einforcement in
in"u¿ã¿ angle of 135o or less, as
columns (see Clause 9.5'2.2 of the code)'

9.2,2 Shea¡reinforcement

The shear r€inforcement may consist of a combination of:

olinksenclosinglhelongitudinaltensionbarsandthecompressionzone;
r bent-up bars; and
Figure 9.1 Some to¡sion link arrangements
. cages or ladders not enclosing the longiJudinal bars but are properly
anãho¡ed at each end in the compression and tension zones'

Atleast50%ofthenecessaryshea¡reinforcemenlshouldbeintheformoflinks. 9.3 Solid sl¡bs

TheminimumtinkstobeprovidedshouldbeinaccordanoewithTable6.2oflhe
code, i'e.: 9.3.1 Flexu¡lreinforcement
A* > v, b, s, /(0'E7 f")
= The minimum tension reinfo¡cement to be provided
in which v, is the nominal shear stretrgth to be provided' F9Jf"
49.N/ml" v' in each direction shan be the
Í taKen as
s¡me as that for rectangular beams.
0.4 N/mmt while for >f 40 N/mmz,v,= 0'4.(1"J40)"'N/mn'with'/cü
g0 N/mrn2 when it is trigber than 80 N/mm2 (note that the formula given in
erse ¡einforcement ofnot less than 20To of ¡he
40 N/mn1'
Clause 9.2.2 of the code is applicable only whenfo 3 provided. In a¡eas nea¡ supports, Eansverse
bars needs not be provided whe¡e the¡e is ¡o
maximum spacing of the links should not exceed tansverse bending moment.
In the di¡ection perpendicular to the span' every
be within 150 mm from a vertical leg of a link' a The maximum spacing of.reinforcing bars in slabs
should comply with the
cage or a ladder. fo-llgwing requirements, which are aaolted from Eurocode
2 (r, i, tË"io;ñ.;tl,
of slab):
In general areas wif¡out concentated loads:
9.2.3 Torsion ¡einforcement ¡ fo¡ the principal reinforcement, mrü. spacing S 3 h s 4Ì0mm;
and
fo¡ the,secondary reinforcement, max. spacing < 3.5 <
The torsion reinforcement shor¡ld consist ofrectangular
closed li¡ks togethcr with -¡
In areas with conce¡hated loads or areas of åaximlum
å 450 mm.

longitudinalreinforcementandisadditionaltoanyrequirementsÏofsnearol ¡ for the principal reinforcemen! max. spacing <2 hmoment:

bending. o for the secondary reinforcement, rn*. ipu.ing < 3 å<250 mm; and
< 400 mm
The torsion links should be closed It is also suggested in ùe code that no fi¡rthe¡ check
¡ecommended shaPes are shown in I eirher (this is derived from BSg l l0: part l: 1997):
is required on bar spacing if
perpendicular to the axis of the me . max. spacing S å < 250 mm (gade 250 steel);
minimum of (the smaller dimensior . mÐ(. spacìng < å < 200 mm (grade 460 steel); or
the link or 200 mm). On the other h peroentage of requircd tension reinfo¡cement
_. the
r_
distributed evenly rormd lhe inside is less than 0.3%.
lvhcre none ofthese conditions appries, the bar spacings
betwecn them ofnot exceeding 300 shourd be limited to the
valæs calculated in Clause 9,2.1.4 of the code.
should be Provided.

106
-ff

Curtailment and anchorage of tension ¡einforcement in slabs may be carried out

risk of cracking at the root
inaccordance with Clause 9.2.1.6 of the code as for tension reinforcement in ofrhe cantireve¡ can be reduced
å::iïi:ilrîd by
beans. r adding a sp.lay undemeath
the root of the cantilever;
or
At an end suppoÍ, which acts as a simple support, zt leasl 50/o of the calculated ' ;å1iT#;plav
wilh diagonal tie back il hrå;;; top orthe ¡oor orrhe
mid-span bottom reinforosment should be extended to the end support and
anchored therein in accordance with Secfion 8.4 and Clause 9.2.1.7 as for beams. eme¡t and the tensile
Ifthe desiga ultimate shear stress at the face ofthe support is less than tralfofvo, o be provided to avoid
a straight length of bar beyond the cent¡elùrc of the support equal to 1/3 of the canrilever, rvhich may
width of support or 30 mm, whichever is the greater, rnay be considered as
effective anchorage.

Hogging moments due to partial fixity of a support could lead to cracking. To

control this, an amor¡nt of top reinforcement capable of resistng 50% of the
maximum mid-span sagging moment should be provided at the support and the
top reinforcement so provided should have an effective anchorage inlo the
support and extend not less than 0.15 times the span length or 45 limes the bar
diamete¡ into the span. As before, the author does not ¡ecommend such practice
ofneglecting the partial fxity in the design calculations because this could lead
to cracking of the joint between the support and the slab. The partial fixity
should be t¿ken into account in the desip calculations and the hogging moment
(a) Possible cracking
so induced properly desigrred for. at root ofcantilever

At an intermediate support of a continuous slab, at least 40%o

of. thc calculated
mid-span bottom reinforcement should be continuous tkough the suppof.

If the deøiling aÍangements a¡e such that lifting of a comer is restrained, suitable
reinforcement should be provided, as required by Clause 6.1.3.3 ofthe code.
supporting
Along a free edge, the slab should be provided with both longitudinal and
fiansverse reinforcement, as shown in Figure 9.4 ofthe code.

(b) Adding a splay undemeath

9.3.2 Shearreinforcement rhe root

Shea¡ reinforcemen! ifrequired, should be provided in accordance with Section

6. I of the code. Howevû, it should be noted that it is in general difñcult to bend
and fix shea¡ reinforcement a¡d assu¡e its effectiveness in a thin slab, especially
if the slab is less than 200 mm thick.

9,4 Crntilevereilprojectingstructures

The codc specifies a minimum top tension ¡einforcement of 0.25%, a minimum

bar size of 10 ¡nm and a maximum bar spacing of 150 mm. AII the top tension
reinforcement should be a¡chored with a full anchorage length beyond the centre
Iine ofthe supporting membe¡ and no reduction in anchorage length due fo aotual
(c) Adding a splay on top
bar sûess should be permitted. Apart from complying with these requirements, ofroot
the author suggests that particular attention should be paid to the risk ofcracking Figure 9.2 Adding splays to avoid
at the top surface of the root of the cantilever due to stess concentation. Stess c¡acking ofcantilever sEuctures

108
 9.6 Vlrrlls

9.5 Columns
9.6.1 General

9.5.1 Longitudinal reinforcemenl in-plane forces, the provisions in tbis section apply.
are:
for longitudinal reinforcement in-plane and out-of-plane forces, both thc provisions
The minimum requirements
rninimum total area: 0'8o/o; ions for slabs in Section 9.3 apply.
'- colr'¡mns;

:î"#J""î:äi*î'i#:cula¡ 9.6.2 Velical reiuforcement

Toavoidsteelconçstion'lhetotalareaoflonginrtlinalreinforcementshor¡ldnot
od
zes of the longitudinal
the column should not
ations.
The minimum total a¡ea of vefical reinforccment is 0.4%. Where the minimr,,,r
a¡ea of reinforcement contols the desigr¡ half of the a¡ea should be located at
each face so that tlrere would be atleast 0.2o/o vertical reinforcement at each face
to ¡esist expected or unexpected out-of-plane bending and to contol cracking.
To contol cracking, the maximr¡m spacing between vertical ba¡s should be 3
times the wall thickness but not larger than 400 m-, To avoid steel conçstion,
the total a¡ea ofvertical reinforcement should not exceed 4yo.

9.6.3 Ho¡izontal reinforcement

9.52 Transverse reinforcement Where the vertical compression reinforcemett 3 2Vo, horizontal reinforcement
confinementto tht::1"*" should be provided such that:
is to provide nooinal It may be in the
longitudinal bar;-4"'n-ú;King'
¡ total areaz>0.30Yofot fr=250N/mm2or) 0.25yofot
d =460N/mm2;
sPiral reinforcement' ¡ minimum bar size: 1/4 of the largest vertical bar size but not less than 6 mm;
. ma:<imum spacíng: 400 mm.

The minimum requirements not less than ó mm;

. 1/4
minimum bar size: longihrdinal bar'
9.6.4 T¡a¡sverse reinforcement
r maximr¡m sPacing: 12
Where the vertical compression reinfo¡cemenl> 2%o, tansverse reinforcement in
the formofli¡ks should be provided such that:
r minimum ba¡ size: U4 of the largest vertical ba¡ size but not less than 6 mrr;
r maxinum vertical spacing: 2 times the wall thickness or 16 times the largest
ve¡tical ba¡ size, whichever is the smaller;
¡ maximum horizontal spacing: 2 times the wall thickness;
r every vertical compression bar is enclosed by a link;
Ito the concrete core' o no vertical compression bar is fr¡¡tl¡e¡ than 200 mm from a ¡estrained
vertical bar;
. at a restained vertical bar, the link should pass round the ba¡ with an
included angle of not more tåan 90o.

9.6.5 Plain walls

To provide nominal resistance and to control cracking, minimum reinfo¡cement

in both the ve¡tical and horizontal directions should be provided.

llt
110
 9.8 Corbel¡
9.7 Foundations

9.7.1 Pile caPs

concentrated in the stress zones

The main tension reinforcement should be to ensure that the total area
between the tops tit piilt' ðËck should be made
of tension
"f
,"ioro,"t-iT'i; b*ìilthe minimum required as stipulated in
Table 9.1 ofthe code'
""

The code sussests that the sides and top

surfaces t"I!: l*. ,î1f,ï:ifi:i,Tj: 9.9 Detailing for ductility
åå";i:t'i"ïïå;ä"i;Ë;"r"'h';'''î'"":I"-:Yn1.1:l j'^ï^d:i":'"ti:1"*fi :
äJ,il:,i":iìÏilä:"#i:"'ilä;i"pt"c
rnsr' rs 6¡] r¡5À - 'tÏiî¿l- 3:l'. 1?;.li*
¿ti tti"ltlr ""tti"or
"'t movement
whelner of the piles The ductility of a rcinforced concrete member is dependent on many factors,
cap is normallY Curte. thi¡k a
including the following:
::t"t:iäffi iiä'iÍ'ä''ä'i''ã's-sü;'r':e*i'9,,T::::1T::,::i'*':il
:H:",.#:"iJ;ä::i;í,.ilfttñ'"'i'^"t:l'rP,;1""1::*:"1"T:lT::',:t,T:
suucrur'' Devouùe "' ""-"
åi tn"- ¡tt of tension developing in any
piles, we are actually- n""';':;t;
. . .,. .- -
-:r^ ^^- ¡. ,,c'ailv, maqsive concrete
;äìil:pä"r,nii,.ãi'üi*'t1,X.f :f-::'"1,:'y,ll',îlillinü;;,iål
:ffi:i:'åi'd;;ì: Ìii';.'J,h; ;ì;t "i'äìv'r'""i"r :'":lÌ,q:. :ï:;*:å*h:i
1

iïiff#:#;;iliüö;;ñ"*ärapite"31l?{'ÌS:îlï;:,1,ff
wouro rur rçeuru¡r¡euri t'""å#t;i"¡;i""''ttti'tipulÀ"d
minil in Table g'l shoilld be
any surface, at least the
Provided'

A Pile caP is actuallY a trans

Howeve¡, the Provision of
difficult and uneconomical' "i;:
sufficient thickness for the pile cap
necessary. and

g.7.2
\.It¿.,twiow
Column andwall footings

Tension reinforcement should be provid{

where necessary' Check should'be lr^1^
tension-reinforcement is not less than the
made to ensure that
*i"i*t '#;;i;;tif
required as stipulated in Table 9'l of the code'
{'ow'Ptt^n-
travip { :ttf*"
r¿r*.>
- recommend
As before, the author^ would not ^:lå'1"-l:"9
least the minimum reinforcement
totally unreinfor."¿' ïtãv twqo¡'. at
stþulated in Table 9'l should be provided'

9.7.3 Tic beams

lt3
 9.9.1 Beams

L ongitudi n al r e infor c e me n t;

comPression bars would

comf rcssion ba¡s should
reinforcement to Prevent

confined by links or ties satis$ing Equation 9.6.

Tra nsvers e re inJorcem en t :

forces would onlY act inwartls'

9.9.2 Columns

Lo ngindinal r e inþrce m ent :

wind loads.

higily commended.
 l0 GENERAL SPECIF'ICATIO\ CONSTRUCTION
WORKMANSHIP A¡ID

10.1 Objectives
bars enter the beam or the foundation member. No column bar should be
temrinated in a ioint a¡ea without a horizontal 90o standard hook or an equivalenl
ancnorage
anchorage -^-- ^' --'^"^^rt.' -^-^:r'là r^ +ilo fo' side
":" of *'":î"'
the beam and
:' Unless illlc_ r. r¡ -c" standards fi¡r
close¡
not closer of cntry'
ol cntry' unless thc |ill l['
|l red slanda¡ds
is
column is leg of _J j lL--
f the horizontal ttg s or noted on
.L- "f-J lL-
strenoth welded--U-
i
ä. L--J -L^..rr L^
U*JrfroofA be +^-,--¡-
towards +L- r-^a of
Ç-- face
the far the ¡nlrrmn
^f +hp column. FullFrrll strength welded--l¿-
Ince. ng should be

10.2 Constructio¡toler¡nces

each lapped bar is confined by lbks or ties satisfying Equation 9'6'

igh-stength reinforced concrete

moment over a finite lengtÏ as
depicted below ( il is the axial load and l, is the gross area ofsection): ofin-situ c

r For O < N/(A.fJ < 0.1, the extent of critical region is taken as l'0 times whether o
, workmar
the greater dimension of the section or where the moment exceeds 0'85 of working condítions. For superstructt
the maximum moment, whichever is larger. mm while for substuctures, tlre devir
r For 0.1 < N /(AEIJ < 0'3, the extent ofcritical region is taken as 1'5 times mm. For piles, the deviation can amo
the greator dimension of the section or where the moment exceeds 0.75 of
the maximum moment, whichever is larger.
¡ For 0.3 < N /(4Iò < 0.6, the extent of critical region is taken as 2'0 times
10.3 Concrete

the grcater dimension of the section or where tlle momenl exceeds 0.65 of
the maximum moment, whichever is larger. 10.3.1 Constituents
The transverse reinforcement should be arranged so tbat each longitudinal bar or
bundle ofba¡s would be restrain€d by a leg The constituent materials should comply
diameter of the üansverse reinforcement 0
with the acceptabre st¿ndards, as listed
in Annex A ofthe code.
less than 6 mm o¡ l/4 of the diameler of the
the greater. The spacing of the transverse reinforcement along the column should
notixceed fre smallerìf l/4 ofthe teast lateral dirrension oftbe column or 6 10.3.2 Mix specification
times the diameter of the longitudinal bar to be restrained. Each end of the
transverse reinforcement should be anchored by l35o or l80o hooks. Anchorage The recommended methods given in
tl
by means of90o hooks or welded c¡oss bars should not bo permitted' ofthe code should be followãd. Each
number. For fit
Transverse reiltlorcement outs ide crit ical regions : variations in
in accordance wilh Clause s!
Outside the criiical regions, the normal requirements retarder and s f(
9,5.2 of the code would applY' tanspo¡tation time, sbould be ailowed.
proportions is unavoidable, the varied
number and separate approval should
be sought before use.
I
production contol and Eansport
10.3.3 Methods of specification"
given in the list of acceptable
The recommended methods
ofth" cod" should be followed'
assessing conformity
10.3.4 Sampling, testing and
The speciñcations glen in Hong fong.go*F*ion
*"tpi.ií"-r"t¿ttÀt iino"t* À
Additlonal cubes:
For sPecial PurPoses'
coûcrete at Eansfer a¡
i"J",tr#tJï Eåi
and curing shor¡rd be
sarrpling should be^
cured u¡der the san
additional cubes shou
Conctete cubesfor cc
The comPressive stre
150 mm cubes at the
be taken from the ûe
ftom a single batch'
10'1 of the code u
identifred bY a distinr
of fresh concrete tak
number sequentiallY
cubes shall be adeqt
avelage compressrve
be taken as the test æsult'
crìleriafor
AccePtance
itt tl¡" standard deviation
th¡n"L5 MPa or is n
Where tlrere is sufl
the same Plant und
of 40 test results
Previous consccuti'
-CZ
stt¡t be adoPtc'
individual test rest
MPa and the aver
sPecified grade str
rãsult shall not be
average of 4 cons
str€4th Plus 3 M
result shall not bc
average of 4 corx
strcngth plus 2 MPa. when ihere are less than 4 test results, the test resufts shalì
be teated as ifthey were 4 consecutive test results. when there are 4 o¡ more
standa¡ds in Annex A test resufts, the average ofeach set of4 consecutive test results sh¡ll be calculated
to checkfo¡
test
the
¡esult and a¡e
not satisfied fish
whether the conc¡ete represented by tåe failed test ¡esult is acceptable or not.
Standard CSI and the list
of Acceptance cilteña Íor I 00 mm cubes :
should be adopted'
nd the conditìons orstorage
t;äå:å"îîi:ft,Ïi:itj;
-
in the -"mb.r.. These
:sting'
specifiedgrade sûengthpl 0D, each
individual test result shall minus 2
MPa and the average of rhm the
specified grÊde stength plus 3 MPa. when the¡e are less than 4 test results, ûre
test results shall be Èeated as ifthey were 4 consecutive test resuhs. when the¡e
are 4 or more test results, the average of each se1 of 4 consecutive test results
shall be calculated to check for compliance each time a new t€st result is
produc the
above be
canied t is
acceptable or not.
1 50 mm cubes versus 100 mm cttbes:
Most of the testing laboratories in Hong Kong are equipped with compression
i'50 mn cubes: testing machines with a ¡¡ryi¡¡rrm loading capacity of around 2000 llrl. These
of 40 previous cons machines are capable of testing 150 mm cubes up to a coDcrete stength of 90
MPa. With the development of high-strength corcrete, the capability of these
testing machines may not be sufficient. The wear and tea¡ of the machine Dlatens
under high loading will also be very severe. The use of 100 mm cubes fo¡
compliance testing could alleviate such problems. The 100 mm cubes would be
easier to lundle and could result in saving of materials and labour. The siorage
and cwing space, as well as the testing .ims, can also be reduced. As a matter of
fact, the Mass Transit Railway Corporation and tbe local concrete producers have
specified the use of 100 mm cubes for internal quality contol. In 1993, the
Standing Committee on Concrcte Technology (SCCT) requesûed the Public
Works Central Laboratories (PWCL) to carry out an investigation to examine the
suitability of using 100 mm cubes for concrete qua.lity contol purposes. The
general conclusions a¡e: (l) the 28-day strength of the 100 mm cubes is on
average higher úran that of the 150 mm cubes made ûorn the same batch of
ll9
118
 Inducted ftom 1998 to 1999'

10.3.5 Placing and comPacting

To oastop
no formed'
wo placing
should be used'

e following afüibutes:
r consistence;
¡ deforrrabilitY/flowabilitY;
r Passing abilit$ and
. segregationresistance(oohesiveness)'

T\e consisle i
is likely to be passing
dryness/wetn compaction to ensure
with a small of the concrete mix at
formed. These lumPs have rather iss
smaller lurnPs when dishubed' BY ID
formcd would grow in size and becor ids
is adde{ there is a certai¡r stage that i
the
a Paste. At such sþge, the .Paste is if
sr¡face tension forces that pull the w mar<imum size of aggregate is 20 mm, then apart from taking the necessary care
during compactio& the content of the l0 - 20 mm aggÌegate ihould be rimitcd to
a maximum of 25% by volurne or to a maximum of 660 kg/m3 by weight, Most
codes allow the gap width between bars and the thickness ofconcrete cover to be
ize ofaggregaæ plus 5
the ooncrete mix, this
of
'bulk volums Powdors)' ratio of smaller than L3
egate to Ãot more than 2
is added' then
If, after form
less
become
itY arrd flowab SegregaÍíon is the separation of some of the ingredients, especially the coarse
deforma imparts the aggregate particles and the paste, ûom the bulk ofthe concrete mix due to ûree
the
the deformabi
sonvenlional falling or sliding through suifaces during placing ofthe concrete, Ieading to non-

t21
 should be carr.ied out under
io likely to take a ìong time,
bc checked by carying out

10.3.6 Curing

The code states that curing is the prevention of loss of moisture ûom the new

Hence, curing sbould comprise of at least fIrc following:

r moistu¡e conhol;
o thermal control;
. vib¡ation conhol; and
r movement/deform¿tionc'ontol.

123
-1F

by a tn¡e adiabatic 10.3.7 Concreting in hot weather

þst. Then, if the temperature ¡ise of the concrÊtc as rncasured
cudng test is higher than 45oC or as meas
evaluation test is higher than 40oC (a ter
yield
l' ä"ii:,ï"ï:'i"ïnllJJ,"lt;ng rime
requÍ
of the s of mix water arising ûom accelerated
intem
Detailed design of the thermal control measurres is beyond
the scope of the code e early fhermal cracking due to high ûemperature
rise.
anct specialist literature should be consr¡lted
when necessary'

a¡e rock blasting' pile

vibratlon control is not normally reguired unless there

following GEO rePort:

Studv of the Effects o'f
[*rt À.f.fl. unå L". P.K.K., GEO Report No'102' 4Engineering Offrce' Civil
Blasting Vibration on Crr.n-öoott"tt, ðeotechnical
Brgi*ã¡rg Department, The 59PP'
tefer to this report where
roî."0o-tã"n¿ätions on vibr
*i vibration control limits for t¡pi different ages has been
" "f
provided.
nless the adjacent
Nonetheless. this
concretc in some
segmeil idge' the stitching
cases. For instance, in precast

curing ofthe stitchitrg segmeût' no cons

deck. Ifnecessary, a temPorary connect
completed decks at the two ends of I
minimize the deforr¡ation of the stitch
example, when an existing bridge is wid
to the old deck and connecting the twc
stitching slab, lhe concrete
remain open to trafhc, whi 10.3.8 Formwork and falsework
deformation of the stitchin
widening, it is generally impracticable Formwork and falsework shoul<r be designed
and consûucted to safely withst¿nd
of time. Hence, trafftc vibration ûom all the loads and vibrations that may or.,ü*inltnr
constuction process.
stitching slab is unavoidable, In sucl
connections hotding the old and new <

owing PaPer;
to coÂcrete stitches in bridge
decks", Proceedings, Institution ofCivi s, Bridge Engineering' Vol'159'
No,2, June, 2006, P953-62.

L¿1
 10.3. l0 Consùuction joints

Construction joints mum number necessary. Their

locations should be tfr, ."""J.-i.'oì"ää¿ *,
should normally nrn the membe¡.
by the following factors:
The time for removal of formwork is determined Tbe concrete at å vcrtical constructionjoint should be
o uihether the conc¡ete structue has developed suffrcient strength to suppori fonned against a stoD end
and bonded with the concrete subsequently pru."¿
all the loatls acting on it; fo¡ movement,
ii
JiÃîïr-'"."íi.ì"',
"loiort
o whether the concrete h¡s been suffrciently cured;
r avoidance ofearly thermal cracking; and The top surface ofconcrete to become ahoñzontar constructionjoint
o the need for subsequent surfacc teatment' level and, reasonably flat. If a kicker is provided, it
should be
should be
and carefirlly constructed. "i-r"árr-io,r-'riet,

Where it until the conc¡ete hæ

turdened
immediatelY.
brasrturg;
dislo
may
dp"ï.ffii,ffJ.,,f,*
used.

d ûee from loose particles before the Aesh

10.3.9 Surface finish
waY that there is no
ete srufaces. Blow
should otherwise be
10.3.1 1
ay be needed prior to the new concretc beins
New concrete placed close to the ioint mul
Movement joints

The movement joints must have sufficient gap widths to

allow future movements
at the joints.

Joint fÌlle¡ forming the g If

more than one strip is us ed
to prevent grout leakage,

Flexible water stops should.be fixed so that they cannot become

displaced from
their intended positions and so that the concretã surrounding tt"m
compacted.
ð- b; frrliy

contractionjoints may altematively be inûoduced by the use ofcrack inducers.

127
-.ilÍ.-

10.4 Reinforcement 10.4.5 Laps andjoints

Laps and joints should be made in

accordance,with the specifications
10.4.1 General details shown in the drawings, and the
rhe engineer.
", "r;gr"J;;
Reinforcement should conform to Hong Kong Construction Standard CS2:
Carbon Steel Bars for fle Reinfo¡cement of Concrete and the acceptable 10.4.6 Wetding
st¿r¡da¡ds listed in An¡ex A ofthe code.
Generally, welcling should be car¡i,
or workshop; welding on site
shou
10.4.2 Cutting and bending only on reinforcing steel ¡hat has
tl
generally be applied at or
near b,
Reinforcement should be cut in accordance with the acceptable standalds. There accordance with the acceptab.le
oould be many different ways of cutting the required lengths of ba¡s Aom compctence should bc de¡nonstrate,
standa¡d lengths (normally 12 or 18 m) but the amormt of leftover (short lengths operations.
remaining after cutting that cannot be used as reinforcement) varies wilh the
rrethod of cutting. Hence, it may be advisable to optimise the cuttitrg schedule to Welding may be used for the following
purposes:
minimize the leftover; commercial software is now available for such purpose. ¡ non_sEr¡ctr¡¡al spot welds for fixing$eel
bars in position;
Reinforcement should be bent in accordance with the acceptable standa¡ds. ba¡s or between steel
Bending should be carried out using mandrels such that each bend has a constant The
curvatu¡e. Grade 250 reinforcement bars projecting from concrete may be bent, elec flash butt welding and
re-bent or straightened provided the internal radius of bend is not less than that subj g maY also be used
specified in the acceptable standards. Grade 460 reinforcement bars projecting ma¡ufacturer,
from concrete should not be bent, re-bent or sfiaightened without the engineer's Welded joints between parallel ba
approval. After bending, re-bending, staightening or any kind ofreshaping, each staggered in the longitudinal d
ba¡ should be inspected for signs offracture. engineer. Welded joints nray be I
digance bettveen them is not ljss
ùa:
10.4.3 Fixing Wel
ld have fo¡ each single pass a length
not ofrun
The reinfotcement should be firmly tied together and securely fixed in position
relative to the formwork to maintain the bars in their prescribed positions during
concreting. Nominal conorete cover should be specifredto all steel reinforcement
including liriks. The specificd nominal cove¡ should be maintained by the use of
fi# ;:*ï',i""':,L':r:n:n;* "*lls
approved spacers and chairs and the achral cover provided should not be less than f0,5 Prestrerslngsúeel
the uominal cover minus 5 mm. The position of reinforceme¡t should be
checked before and during concrcting to ensurs that the nominal covcr is
maintained within the prescribed timits. Where conside¡ed appropriate, a cover 10.5.1 General
meter may be used to check the cover to reinforcement in hardened concrete'
Prestessing tendons
ptable standards listed in Annex
A ofthe code. TheÌt ¿rs and sheaths should be tlose
10.4.4 Swface condition specified in the design as such.
The surfaces of reinforcing bars should be clean and free tom deleterious
substances such as oil, grease, mud, loose rust o¡ mill scale that may adversely I 0.5.2 Transport and storage
affect the bond between the bars and the corcfete. Ifnecessary, the steel su¡faces
should be cleaned by brushing and,/or washing before the reinforcing bars are
fixed. The su¡face condition ofthe reinforcing bars should be examined again uld be protected ûom harmful
prior to concreting.

128
t29
 To ensure protection ofthe tendons, lhc following should be avoided:
¡ mechanical deformation or heating during handling;
. cxposr¡¡e to rain o¡ contsct with the ground;
¡ welding or cutting of steel in the vicinity oftendons without protecting the
tendons from splashes;
¡ after manufacture, any welding, on-site heat teatment or metallic coating
such as galvanizi¡g; and
. contamination likely to afect the durabilþ or bond ofthe tendons.

10.5.3 Fabrication

At the timc of incorporation in the sûuctural member, all prestressing tendons,

sheaths or ducts shoutd be ûee from h¡mrfirl matters sr¡ch as loose rust, oil' paint
o¡ other lubricants. Oiled or greased tendons may be used under certain
circumstances, ifagreed among the parties involved. Tendons may be cleaned by The tensioning appa¡atus shor¡Id
have been calibrated witl¡in six
wire brushing. months.

Low and nonnal relaxation wi¡es and prestessing sûands should be tnnsported
in coils of sufficiently large diameter to ensure that they run of staight from the
coils. Prestessing bors should be sÈaight; small adjustuent for staightness may
be ca¡ried out on site by hand in cold under the supervision of the engineer but
ba¡s bent in the tb¡eaded portion should be rejected.

Cuning to length and timning of ends should be by either high-speed abrasive

cutting or oxy-acetylene flame cutting. Neithe¡ the flame nor splashes should
come into contast \Ìith the anchorage or ofåe¡ tendons. Post-tensioned tendons
should not be cut lcss than one diameter Aom the anchor and the temperaturc of
the tendon adacent to the anchor should not cxceed 200oC.

10.5.4 Placing

sheaths should be aocurately placed in the

pemritted deviation in the locations of the
d bc +5 mm unless otherwise statgd. Ths
tendons, sheaths or duct fomrers should be securely fixed such that they will not
be displaced by distu¡bance from nearby activities or during concreting.

sheaths and extactable cores should be strong enough to retain tbeir correct
section and profì adequatelY
sealed lo prevent leæd. Ends
ofducts shouldb Extractable
cores should not be coated with release agent excePt with rhe approval of the

passage for the bndon.

I

determined by test and allowed for. Before tensioning, the lendons should be Duct design and construction;
shown to be free to move in the ducts. Dwing stressing, allowance should be Sudden changes in the diameter or al
made for the friction in the jack, unless load cells are used. Measruement of the should be provided at crests in thc
tendon extension should not commence until any slack in fhe tendon has been section of the duct, and slsewhere if re
points ifthe difference in level betwee
m' Vents shourd also be provided at aricho*ges. AI vents should be closable.
vents to be used as enty points shourd be ttuei'a.a to pr.mit the use of a screwed
ts and injection connections to the ducts
ld be able to withstand distu¡bance before
corrective measures should be taken with the approval ofthe engineer' Records during grouting. Lined ducts should be
should be kept of all tensioning operations, including the measured extensions, rrrosion of the tendons or having excess
pressr¡re gauge or load cetl readings and draw-in at each anchorage' I to blow dry oil_free aìr through a lined
¡n ifit is left ungrouted for a considerable
sorption
t0.5.6 Protection and bond ofprestressing tendons to flush
e sealed
General:
Prestressing tendons must be protected against mechanical damage and corrosion.
Protection against ñre damage may also be required. stressed tendons ale Properties oîgrout:
trormally required to be bonãed to the struct,'e after all the tensioning and Tl. grolt should hsve high fluidity (same as flowabiliry) and high cohesiveness
anchoring operations are completed. when plastic and tow shrinkage ànd adequate strength
when Éar¿en"¿. the
fluidity may be assessed using ihe immersiån ot t¡" test methods set out in
Protection ond bond of internal lendons: "ä" of the resistance to
the acceptable $andards. The cohesiveness is a measure
cement grout or cement ssnd gfout may be employed to protecl and bond intemal segregati ùon. Cohesiveness may be inc¡eased bv
tendons to the stucture. reducing io, adding putverizeoä"ì-"rh';ä;r";
viscosity 100 mm ãube strength ofthe grout shoUJ
not be less than 27 N/mm2 at 2g davs.
damage and
of adequate Composifion of grout:
nts that maY Tbe be c portland cement and
water but may
prolection ¡s also ,f¡ll Normally, sand is included
concFete or mona¡ and the¡e is the possibility ofundesirable cracking, B
pfimary grou d in "trì;
r of more than 150 mm. pulve¡ized
conosion by the differential movements fuel be a ¡nsidered suitable. Afthough putv*zeã
should be e required to be bonded to the ñ¡el ash is often regardedjust as fi[e¡ in thc gout, it actuallyurro ñrnctions as u
strucil¡e, tably reinforced and tied to the he cohesiveness ofthe grout.
structure. ay be added to improve the
, the use of gas generating
generation of an excessive
10,5.7 Grouting Ltially weaken ihe grout and therefore if
rûolled, the grout could be seriously
General: Ê years seen quite a number of failure
The main objectives of grouting ducts in post-tensioned concrete members a¡e: ng admixtures, Exteme care should be
I to prevent corrosion ofthe tendons; and mixtu¡e and trial mixing with different
r to allow effective transfer ofbond strcsses. rut should be ca¡ried out to evaluate the
To meet the first obj
cover the tendons. It
the second objective, Batching and mixíng o/ grout:
when hardened shoul The water/cement ratio should not exceed 0.44. For a neat cement grou!
the
effective bond between the tendons and the sides ofthe ducts' optimum water/cement ratio is probably around 0.40, witl a suitable pústicízing
admixture added, the water/cement ratio may be reduced to 0.¡S.
Theìuantitv oi

133
 l1 QUALmy ASSTTRANCE AND QUALTTY CoNTROL

ll.l Scope

meâsures for the design and construction

mprise of essential actions and decisions
, as well as checks to be made, to ensure

ll,2 Quality lssunnce

Grouting procedure:
standârds'
The groiting procedw€ should bs in acoordancc wíth the acceptable

Some other ge
The Concrete No'47: Durable Bonded Post-tensioned
ConcreteBrid recommendations:
¡ Ducts and vents to be pressure lested.
r points in
Vents to be positioned at lolil poi¡ts, high points and beyond high

¡ to be provided covering fluidity' bleeding'

e, strength and sieve tests'
r Use of a plasticizer in the grout'

ll.3 Cl¡ssific¡tion ofthe control mersures

There are th¡ee basic control systems: (l) internal control; (2) external contror;
and (3) conformity contol.

Internal conirol:
any
Itis
nal"

External control:
Exæmal control is carried out by an indepen with this
task by the client or by the relevant authority. ist ofthe
ve¡ification of internal control measures extemal
r

requirËments or additional checking procedures independent of the intemal

contol system. -lg."tfU
g renote areas_ or where the volume of
50 m'. Even fo¡ these ..exceptional;pr"¡"".1
obrained from a supplier oplrating i.
Conformity control: ; ;pp;;;;
Conformity contol is exe¡cised 1o veriff that a particular seryice or production
function has been carried out in conformity with the specification previously
established. Conformity control is generally part of extemal control. tnslruclion:
constuction conhol are summarized
rnd constructio¡ control should "
in
i""lìÀ,
LL.4 Veriñc¡tion systems t,
rsttrction; and
The frequency and intensity ofcont¡ol depend on the chances and consequences
of non-compliance with the specification requirements in the va¡ious stages of propriate for:
design and construction, Different contol measures are oombined in a is to achieve a unifo¡m quality
of the
verification system to ensure the effectiveness ofthe control system.
. â single product: the aim is
For a single producf it mav
measures, in particular on initial
11.5 Confrol of e¡ch ctage of design and co¡struction process
Inilial tests:
According to the purpose and lime of the cont¡ol, the following stages may be
distinguished:
in order to
r control ofthe design; d standards
o cont¡ol ofthe production and construction; and detrilnental
o control of the completed strucfi[e.
C hecks during cons lruc lion :
11.6 Control of design If.ffi"ï:,,ï ild": ,tj."^0,.1:: "|h...r"ú,¡.ls and componenrs built into fhe
::,"d:':::i,i,.,l;
sqr¡¡¡Ë Lvuruueuon-
iiï:H".1î.jlr":::"r:::_g"l::-",rn;.",i*;;;sï;",;_ffi
measurJment
the verification Ihe results ol
Control ofdesign shall conform to the statutory and adminishation procedures. It rot *'"..,ouotîf
i""ïff ii:iÏ
paries concemed. For ,""ÍTt:-:1":lo--Tt'"1*tl {]"a uìi
should be noted that despite the approval grven by the relevant authorit¡ it is
ullimately the responsibility of the designer to ensure that the design assrunptions
#ff
:i:.:.li:: *i,h ';"Xî,n1vï¡;.;;;.öi#'å.'iri;fi
.:::"#f gsps.c: r"i,.i"¡or.ioÈ ri.;ì;'r#äj;
rçr¡¡¡lru.rg sreer, rne delrvery ticket sbourd i'crude
are oorrect, the design calculations contai¡ no mistakes, the computer programs :i.Tlii:,:,îl^$"_"jigiî.*d
should have labets and ¡-olling
oi,r,",""i
the ideniity ¿.livered and ùe
the sreel
steer
used are suitable for the type of structure being designed and have been marks for idån¡fl;;ìt;
thoroughly verified, the drawings are accurate presented, the specifications are Conformil controls:
proper and the completed structure will perform reasonably u/e[ 10 meet with the
requiremenls of Section 2.1 of the sode. It is up to the desigaer to aÍange
intemal contol to ensure that the design is ofan acceptable standa¡d.

lt.7 Control of production ¡nd construction

Objectives:
The cont¡ol of production and constuction comprises of all measures (e.g.
inspections and tests) necessary to maintain and regulate the qualily ofmaterials td straclure:
and standard of the workmanship in conformþ with the specified reguirements.

P roduction of concrete :
Structural concrete for all works should be obtained from concrete suppliers who
a¡e ceÍified under the Quality Scheme for the Production and Supply of Concrete
L2 PRESTRESSEDCONCRETE

l2.L Baris of design

In the design ofprestessed.concrete sfiuctu¡es, it is

difücult to prodict whether
the ultimatc limit state.-or the serviceability limit
søte will be ihe criticJ one
because the serviceability limit state requirements
for prestre"."¿ .ã*ãt"
structures are generaily more striugent than those
for ordina¡y reínforced
structures. Hence, the usual practice in the design "o""i"t"
or'r.¡oior..l.ãnìiJt"
sfructr¡es of firsr designing fbe stucture in accordance with the
state requirements and then c "lti-"rc-lirit
the serviceability limit state
ultimate limit state wil¡ be th
design ol
limit star
at tlìe han

Altemative design re tha¡

those made in the c riate
especially forthe de
fi"ää'Ëi:.:f,lö.orconc** -'- -
es

In the assessment of the rikery behaviour of a prestessed

member, tåe amount of flexr¡¡al tensile stresses allowed
dependent on the serviceability classification
concrete stucture or
r.*i"" -a..
ofthe stucture or member.
i*ã i.
ä:"ri1iîïil"*:.*]:
3;t""IîÉ *' *;d;:
lll"-ottttg tendons aæ given in clarr." ì ;.":;":i^_-
låå'"**nrtenaona,eli*i"cr"*.i.ãä"äil'":ïj,iË:ïi"iï,ffi
pre-tensioning

steel
¡einforcement anì
a¡e c35
.h","dË;ïr"ï;;
,îi,i:*,:"ïtr"î:ff
There
a¡e three sewiceability classes:
o Class l: noflcxuraltensilestesses. 12,2 Structures ¡¡d ¡f¡uctur¡l fr¡me¡
r Class2: flexuraltensilestcssesallowedbutno visible cncking.
¡ class 3: flexr:¡ar tensile st¡esses ariowed but subjected to a maxirnr¡m
surface crack width of 0.1 mm for members i" opor*ì
conditions 3 or 4 and a maxim,r¡ su¡face c¡ack width analysed using the
of 0:t; .3 or followinã the
for all otler membe¡s.
should be givãn ro
The definition for class 2 given in ct¡ruse l2.r of the may be analysed æ
code is rather vague and
ambiguous because it is not known what is meant by ..no
visiåIe ,räfng;i
whether a crack is visible or not is dependent not only on the c¡ack
*¿tri, Ë"i
ilrg 9n F: Iighting condition a¡d the èyesight of the ãbserve¡. Io ta.t, cilu"" ess should generally
12'3'4 of the code redefines class 2 as ;flr*l*d tensile sgesses be based on tåe concrete
arowed but the 5.1.¿orrnecode. section as
gtr
of
of eding C70, ¡edistribution
in of momenß mav
Iitions are satisned:
:mal a¡rd extemal
forces is maintained.
comparing the above crack width limitations for class 3 prestressed
concrete Le moo.mum
dt:srgn moment derived
:T:q"i with those stþulated for prestessed members with bonded tendons in
Table 7'1, it can be see¡ that the above limitations are more
>xceed2üyo. Êom
stringent tan nos"
stipulaûed in Tabre z' l,
Based on the spirit that if the¡e ¡. *y
¡"'"o*¡.t"nüü neutal axis
te¡ than the
Ç

12,3 Be¡ms ¡ Class ll No flexu¡al tensile stress.

o Class 2:

12.3.1 General

The geometric properties ofprestressed conc¡ete beams are defined and limited in
the same ways as for reinforced concr€te beams in Clause 6.1.2.1 of the code, t*"o
except that the overall depth of the member should be used instead of the . class 3: "t
0'45 JE )'
Although cracking ,.
effective depth, section is uncracked a¡d "rr:t
exist at the limiting crack
stesses are given in Table
12.3.2 Slender beams , modified by the coefficier
When additional reinforcr
Clause 6.1,2.1 should be refened to regarding the slendemess limits for avoiding zone, and is posifioned clc
lateral instability. Particular attention should be paid to possible instabilify of the these modified design I
beams during conslruction as well as when the beams are in their ñnal positions. increased by an amount tlu
;sed as a percenf^ge of
12.3.3 Continuous beams in the tension zone).
in (see Table 12.2 f;;
Continuous bearns may be analysed using an elastic analysis method with the reinforcement, the st¡esses
following anangements of loads considered: (1) two adjacent spans loaded with members in groups a) and
maximum design loads and all other spans loaded wilh minimum design loads; group c). For other percer
(2) altemate spans loaded wilh maximum desip loads and all other spans loaded stresses may be increased i
with minimum design loads; and (3) all spans loaded with maximum design Whel a significant propo
loads. Redíst¡ibution of the moments obl¿ined by elastic analysis may then be transitory so that the whãI¡
ca¡¡ied out as per Clause 12.2.3.
ûequent load combination
imposed load; see Section ?
load combination ofdead lc
12.3.4 Sewiceabilþ limit state for bearns load is not a permanent ¡oa
stresses may be exceeded r
Section analysis: characte¡islic loads).
For analysis of sections at sewiceability limit state, it may be assumed tlat plane
sections remain plane and that the materials are elastic and linear. In general,
only the load arrangements at fhe following two stages: (l) immediately afte¡ the 12.3.5 Sþess limitations at tansfer
for beams
transfer ofpreshess, and (2) after all losses ofprestess have occurred, need to be
considered. For such analysis, the effects of dead and imposed loads on the Compressive sÞesses n concrete;
forces in the tendons may be ignored. The compressive stresses should
not exceed 0.5
for nea¡ ru¡iform distributions of prest
f
at the exteme fibre nor 0.4/,
Compressive stresses ln concrele: *h"J I
is the concrete slrength at
tra¡sfer. "rs,
In flexue, the compressive stess should not exceed 0,33 /- excep within the
range of support Eoments in continuous bea¡rs and other statically indetenninatc
Flexural tensile stre¡.re., in concrete:
sûuctur€s where the compressive stess may be increased to 0.4f,". In direct The tensile stresses should not cxceed
the following limits:
compression, the compressive st¡ess should not cxceed 0.25 f^, r Class l:
l.0N/mm2.
¡ Class 2; for p¡e_tensioned members
Flexural tensile in concrete:
ctrs,Jses O,asrff, and fo¡ post-tensioned
At mo¡ta¡ or conc¡ete joints of precast units, no tension should be allowed. members ß6,[8.
Elsewhere, the tensile stresses should not exceed the following limits for different r class 3: same as for c.lass 2; if the tensile
classes of prestressed concrete structures/members:
st¡ess- limit is exceeded, the
section should be considered as
cra"k"J.
12.3.ó Deflection of beams 12.3.8 Design shear resista¡ce of
beams

There a¡e no deemed-to-satis$ rules. When it is considered necessary to Mmimum design shèar stress;
calculate the deflection, the following methods may be used:
The maximum design shear stress
. should be lirnited
-- to
--
Class l: Elastic analysis based on the concrete section. ur ?.0 l\/rnm
"'"V''cu or N/mm2
Whichever is the smaller. '.gfi ''u '
¡ Class 2: Elastic analysis based on the concrete section.
r Class 3: Ifthe frequent load combination results in stresses not greater than ,:
those in Table 12.1, an elastic analysis based on the concrete I
section may be used. Otherwise, rigorous analysis based on the
moment-cun'atre relationshio fo¡ cracked section should be
.rl
'"li
ned
carried out.
necessary, shear ¡einforcement
should If
Secti
12.3.7 Ultimate limit state for beams in flexu¡e
The shnc
Section analysls:
corre a ma flexure Z*
centoidal axis of üre section of stress ¿t the
The following assumptions should be made: I = 0.2 of %o' tl¡e
¡ Plane sections remain plane.
design value ofthe presrress at the centoidai axis
s
¡ The desip. stresses in the concrete in compression a¡e either derived f¡om
which f is rhe compressive sress ar rhe
cen'oidallÏntilïîJr:,?;
the stess-stain curve given in Figure 3.8 of the code with r,
= 1.5 or taken as positive. Based on ihese desiga
values, the value of %o is determined
as:
as 0.45 f-over a depúr (from the compression face) equal to 0.9 times the
depth ofthe compression zone. V", = 0.67 b"h +0.Efof,
o The tensile strength ofconcrete is negligible.
Typical values of Z- so determined
. The desigl stresses in any additional reinforcement and in bonded tendons, for conc¡ete ofgrade C30 to C60
and
whether initially tensioned or untensioned, are derived from the appropriate
ranging ûom 2 to 14 N/mm2 are
labulated in Table 12.5 of the code.
f
stess-strain cuwes as given in Figure 3.9 and Figure 3.10, respectively.
Seclions
. The desip stress in unbonded tendons is limited to the values given by The desi istance ofa section cracked
Equation 12.2 unless more rigorous analysis justi-ñes a higher value. calculate equation:
in flexure Vu may be

Designformulae: ( ¡\
The resistance moment M, of a beam containing bonded or unbouded tendons, v. = *u.v or
\lt-o.ss!:zþ.4a 'M 0.1 ,l f_bd , whichever is the greater
all of which in the tension zone, may be oblained as: .¡el
M, = fùAp,@ -d")
For a rectangula¡ beam or a flanged beam with flange thickness not less tbatr
0.9x, dn may be taken as 0.45x, For bonded tendons,St and x may be obtained
from Table 12.4 of the code. For rmbonded tendors, $¡ and .x may be obtained
from Equation 12.2 and, Equation 12.3 respectively ofthe code. lable 12.4,
Equation 12,2 ø¡{ Equation 12.3 of the code a¡e the same as those given in Shear reinforcement;
BS81 l0: Paft L: 1997, except that the design stength of fhe tendons is taken as The amount of shear reinfo¡c
be provided is gíven as follows:
0.t7l* insæad of O.95lw as in BSBI10. where r< % throughout the"o|ä1"
no require
of minor importance;
Allowance for additlonal reìnforcemenl in the lension zone: minimum
The additional reinforcement in the tension zone may be t¿ken as equivaletrt to Where 0.5%
,rd(see Clause 6.1.2 ofthe code):
prestressing tendons wilb an area of Alfylfp, minimum
F ---:=--r;;:-r'r¡ll

rühere Z" + n< Vi t2,7 Prestrecsing

links to be provided such that A"" > su(Y -Y")/O.t7îwdt in which dr is the
The jacking force should not normally exce
depth from the extreme compression fibre eithe¡ to the longitudinal bars or
to the centroid ofthe tendons, whichever is the greater,
of the tendon bur may be i¡lcreased 1o g0%
given to safety and to the load_exterlsion
At both corners in the tensile zone, a link should pass round a longitudinal bar, a
transfer, the initial plestress shourd not normaüy exceed 70%o
tendon, or a group of tendons having a diaÍ¡eter not less than the link diameter. ofthe characte¡istic
strength ofthe tendon and in no case should it ixceed,TS%o.
A link should extend as close to the tension and corrpression faces as possible,
with due regard to cover. The spacing of links should nol exceed 0.75 d, or 4
times the web thickness. If Zexceeds 1.82", the maximurn spacing should be 12.8 Loss ofprestress, other tban friction losses
reduced to 0.5 d, .

12.8.1 General
12.3.9 Torsion
In the calcuration of the design forces in the tendons at va¡ious stages of
Where torsional resistance is necessary for equilibrium or significant lorsional construction, ¿rlowance shourd be made to trre possibre ro.r..
stesses may occur, it is necessary to calculate tJle torsional moment and stresses forces. Other than friction, losses of prestress may aris" Aom:
Jp...ti.ring
that could be developed and provide torsional reinforcement accordingly. The o relaxation ofthe tendon stgell
method adopted for reinforced concrete beams may generally be used. ¡ elastic deformation and subsequent shrinkage and creep ofthe
concrete;
r draw-in oftendons ¿t anchorage during anchoring; and
¡ other causes in special circumstances.
12.4 Sl¡bs

The recommendations given in Clause 12,3 for beams apply also to slabs, except
12.8.2 Relaxation of steel
that shea¡ reinforcement needs not be provided if / is less lhz¡ V..

The long-terrr loss ofprestress due to reraxation ofsþer may be obtained aom
The design ofprestressed flat slabs is outside the scope ofthe code and should be
the 1000 h reraxation val'c of the steel by the extrapolatioi m.-aoJ-or
canied out in acco¡dance with appropriate specialist literatr¡¡e.
multiplying trre 1000 h relaxation value with an appropriáte relaxation
facto¡.
Table 12.6 gives the relaxation factors to be used in the'c"r.r orlr.-trr,rioJog
and p_ost-tensioning. The 1000 h relaxation value should be
L2.5 Columns taken tárn tL
manufactu¡e¡'s appropriate certifi cate.
For columns in framed stuctur€s, where the mean design stress in lhe concrete
section imposed by the tendons is less than 2.0 N/mm', these may be analysed as
ordinary reinforced conc¡ete columns.

12,6 Tension member¡

12.8.3 Elastic deformation of concrete
The tension capacity should be evaluated based on the desiga strength of the
The immediate loss of prestress due to elastic deformation of the concrete at
presû€ssing tendons (0.E/*) and thc actual tensile stresses that could be
developed in the additional reinfo¡cemeil (subjected to a maximurn value equal
transfe¡ should elastic modulus of tåe concrete at the
time of transfer ioning, the loss of prestress should be
to the design strength of the reinforcement, i.e. 0.8/r) when the tendons reach
calculated on a the stress in the a_djacent
their design strength. Stain compatibility should be used to evaluate the lensile in
the case of post-tensioning if "oorr.t". a
the tendons a¡e not stressed simultaneousl¡
sEesses in the additional reinforcement. The reason for determining the tensile
progressive loss occurs. In such cæe, the average loss
st¡esses in the additional reinforcement in this way is that whilst the additional may be calculated oíthe
basis of half the product of the modular ratio and the stress in the
reinforcement should have suffrcient ductility to maintain its design stength after concrete
adjacent to the tendons averaged along their length or alternatively the
yiclding until the fendons reach their design stength, the tendons may not have loss of
prestress of each tsndon may be exactly computed on the
sufficient ductility to maintain their design strength, if they reach their design basis ofthe sequence of
tensioning.
strength firs! rurtil the additional reinfoÍcement also reaches its design slrength.

t44
12.8.4 Sbrinkage of concrete the tendon and the sides of the duct and so produce friction. Because of such
frictior¡ the presfessing force p* decreases as the dista¡ce ûom the jacking point
¡ increases. The rate of change ofthe presÈessing force with the distance from
å"J'i::i*ii'.i'.ii,'"""ili the jacking point due to such füction is given by:

âP
after rhe curins stops and at the t¡-. oræorior,¡iåï",iåiå*ffi;äÏt?':i: = -'KP'
total shrinkage might have already taken place. =øx
dependent on tåe type of
cient of its inside surface,
12.8.5 Creep of concrote employed in placing the
not less than 33xl0a m'r
The loss ofprestress due to creep ofconcrete mry be obtained as the product of
dso
the creep per unit lenglh ofthe concrete adjacent to the tendons and the elætic
ybe
modulus ofthe lendons.
alue

12.8.6 Draw-in during anchorage

The above equation is given in a different fonn ftom that given in the code
because the friction due to r¡ninûentional variation from the specified profile
cannot be conside¡ed independently from the friction due to curvatuæ oftendons.
The friction due to unint€ntional variation from the specified profile and the
friction due to curvafure of tendons have to be considered at the same time
because one will affect the other.

L2.9 Loss ofprestress due to frlction

1 2.9.4 Friction due to curvature oftendons

The cuwature of a tensioned tendon produces radial force acting against the
12.9.1 Gene¡al inside surface of the duct o¡ sheath. Such radial force üren produces friction
along the di¡ection ofthe tendon and thereby causes the prestessing force P, to
decrease as the distance fion the jacking point x increases. The rate of changc of
the prestressing fo¡cc with the distance ûom the jacking point due to such ftiction
is given byr
AD
?: -p1/,^
øx

should atso be calculated øking into account the variation in force along the where p is the coefficient of friction and r- is the ¡adius ofcurvature. The value
length ofthe tendon for later checking on site during the ænsioning operation. of p is dependent on the type and the surface condilion of the tendon and the
duct. Typical values of ¡r are as follows;
12.9.2 Friction in jack and anchorage ¡ liÊlrtly rusted stand running on unlined concrete ducl: 0,55
¡ lightly rusted stand running on lightly rusted steel duct: 0.30
The ûiction in jack and anchorage should be detennined by calibration of the r lightly rusted strand running on galvanized duct: 0.25
actual jack and the type of anchorage to be used. ¡ bright strand running on galvaaized duct: 0.20
r greased strand nrnning on plastic sleeve: 0.12

12.9.3 Friction in the duct due to uninlentional variation from the specified profile Considering the ûiction due to unintentional variation from the specifÌed profile
and the friction due to curvature of tendons at the same time, i'e' combining the
Whether the specified tendon profile is straight or cuwed, slight unintentional above equations together, the following equation is obtained:
variations from the specified profi.le cause additional points of contact between

146
 T

t2.8.4:
,n? rì
+-l rpt
.)

-.ùe rp, varies along the length ofthe tendon, this equation

-.r by numerical integration, stErting at thejacking point where r =

, : P,. One simple way of evaluating the values of P, along the length
rne tendon is to input the profile of the tendon (in the fomr of the three-
dimensional coordinates ofthe tendon along its length) into a computer program, attention should also be paid to end blocks.having a cross_section
which automatically evaluates the cuwature at point along the lengtl¡ in shape ûom that of the general cross-section of úre beam. In srsh
calculates the total frictional loss within each short of the tendon and tb¡ee-ct¡mensionaì finite element analysis of the end blook may be needed.
then determines the values of P, along the length of by step-by-step
inægration starting at the jacking point.
I 1.3 l-Itimate limit sfate

12.9.5 Lubricants In the case of members with bonded tendons, no checking of the design bursting
tensile force at the ultimate limit state is necessary.
formulation, lubric¿nts may be wed to ease the movement
If of satisfactory
lendons in the ducts. Lowe¡ values for p may t}ren be used subject to their being In the case ofmembers with unbonded tendons, the design bursting tensile forcc
at the ultimate limit state should be assessed from Table 12.7 of tlre code on the
determined by tial.
The criteria of Clause 10.5.3.1 regarding application of
basis of the characteristic tendon force; the ¡einforcement provided to resist this
lubricants and subsequent cleansing should then be satisfied if the tendons are
force may be assr¡med to act at a stess of 0.871,
subsequently to be bonded into tho structure. .

12.10 Tr¡¡smi¡sion lengths in pÌe-tersiored member¡ 12.12 Consider¡tions ¡ffecti¡g design defailr

Clause L10.2 should be ¡efened to.

12.11 End blocks in post-fensioned m€mberg a¡e to supplement those given in Section 8.10 ofthe code.

l2.l1.l General 12.12.2 Limitations on ofprestessing tendons

In the design ofend blocks, considerations should be given (a) bursting forces The size and number tendons provided should be such thst
a¡ound individual anchorsg€s; (b) overall equilibrium of end block; and (c) cracking of the concrete ultimate failure of the beam. This is to
spalling of the concrete from the loaded face Only the avoid sudden fait¡re ofthe bea¡r any prior signs offailure,
brusting forces a¡e dealt with herein. Special be ¡efcrred to for I
Ì
consideration of overall equilibrium and spalling of at the loaded face. The above requirement may be cousidered to- if the ultimate moment
of¡esistance exceeds the moment necessa¡v to a flexr¡ral tensile süess in
the conc¡ete al the extreme tension ûbres equal to 0.6.[], .

12. 1 1.2 Sewiceability limil statE

At üre sewiceability limit state, the design force in an individual square
end block loaded by a symmetrically bearing plate may be derived
from Table 12.7 ofthe code, which relatcs the bu¡sting tensile force in each ofthe
two principal di¡ections to the half width of loaded a¡ea to half width of end
block ratio. This bursting tensile force should be laken as distributed in a region
extending ûom 0.2 times the half width of the end bock to 2 limes the half width
12.12.3 Cove¡ to presÉessing tendons
Bonded tendons:
The nominal cover to be provided should be sufficient to protect f¡e tendons from
conosion and fire, as per the requirements in Sections 4.2 a¡d 4.3 of the code.
For pre-tensioned tendons, the ends ofindividual tendons do not normally require
 (

ofthe end block from the loaded face, and resisted by ¡einforcement in the form
9?--r-(*.Y) J

ox ¡P'./
\
In actual practice, since r- varies along the length oflhe tendor¡ this equation
f

jacking point where x =

has to be solved by numerical integration, starting at the
0 and P, :
1". One simple way of evaluating the values of P, along the length
of the tendon is to input the profile of the tendon (in the fonn of the úuee-
dimensional coordinates ofthe tendon along its length) into a computsr prog¡am, Special attention should also be paid to end blocks.having a c¡oss-section
which automatically evaluates the cu¡vature aÍ every point alotrg tle lengtlu diferent in shape f¡om that ofthe general cross-section of the beam. In such
calculatcs tho total frictional loss wilhin each short scgment of the tendon and case, tlree-dimensional furite element analysis ofthe end block may be needed.
then determines the values of P, along the length ofthe tendon by step-by'step
integration starting at tlejacking point.
l2.l1.3 Ultimate limit stat€

12.9.5 Lubricants In the case of memben with bonded tendons, no checking of lhe design bursting
lensile force at the ultimate limit state is necessary.
If of satisfactory forrrulation, lubricants may be used to ease the movement of
In the case of members with unbonded tendons, the design bunting tensile force
tendons in the ducts. Lowc¡ values ¡
for may then be used subject to their being
at the ultimate limit state should be assessed from Table 12.7 of the code on the
determined by trial. The crite¡ia of Clause 10.5.3.1 regarding application of
basis ofthe characteristic tendon force; the reinforcement provided to resist this
lubricants and subsequent sleánsing shor¡ld then be satisfied if the tendons are
force may be assumed to act at a st¡ess of 0.87 f" .
subsequently to be bonded into the structur€.

12.10 Tr¡nsmission lengths in pre-tensioned members 12,12 Considerations alfectlng design details

Clause 8. I 0.2 should be referred to.

12.12.1 General

12.11 End block¡ in post-tencioned nember¡ These considerations are to supplement those given in Section 8. I 0 ofthe code.

12.l1.l General l2.L2.2Limitations on area of prestessing tendons

In the design of end blocks, conside¡ations should be given ûo: (a) bursting forces
a¡ound individual anchorages; þ) overall equilibrium of ùe end block; and (c)
spalling of the concrete from the loaded face around anchorages. Only the
bursting forces are dealt with herein. Special literature should be referred to for
consideration ofoverall equilibrium and spalling ofconcrete at the loaded face.
12.l1.2 Serviceability limit state
The size and number of prestessing tendons provided should be such that
cracking of the concrete would precede ultimate failu¡e of the beam. This is to
avoid sudden failure ofthe beam without any prior signs offaih¡re.
The above requirement may be considered to be salisfied if the ultimate moment
ofresistance exceeds the moment necessary 1o produce a flexural tensile sfress in
the concrete at the exheme tension fibrcs equal to 0.6,[h .

At the sewiceability limit statc, the desþ bursring force in an individual square
end block loaded by a synmetrically placed square bearing plate may be derived
ûom Table 12.7 ofthe code, which ¡elates the bursting tensile force in each ofthe
two principal di¡ections to the half width of loaded a¡ea to half width of end
block ratio, This bursting tensile fo¡ce should be taken as distributed in a region
extending from 0.2 times the half width of the end bock to 2 times the h¡lf width
12.12.3 Cover to prestessing tendons
Bonded lendons:
The nominal cover to be provided shoirld be suffrcient to protect the tendons from
corrosion and fire, as per the requirements in Sections 4.2 and 4.3 of the code.
For pre-tensioned tendons, the ends ofindividual tendons do not normally require

149
 concrete cover and should preferably be cut off flush with the end ofthe concrete l3 LOAD TESTS OF'STRUCTURES ORPARTS OF
member. Forpost-tensioned tendons, the equivalent bar size for ducts containing
a number ofstrands shouldbe calculated from the total a¡ea ofthe strands inside
STRUCTURES
the duct and the minimurn cove¡ to the outside oftle duct should be nol less tha¡r
the minimum dimension of üre duct c¡oss-section nor less than half the lareest
dimension of the duct cross-section.
13.l Gener¡l

Unbonded lendons: of a structùe or structural

The cover to the ducts of unbonded tendons should be sufficient to protoct the Model or prototype testing
tendons from corrosion and fire, as per the requiremenls in Sections 4.2 and4.3 ice are not included.
ofthe code. In any case, the nominal cover to the ducts should not be less than
The need for Ioad tests may arise during construction unde¡ the following
25 mm.
circumstarices:

Exlemal lendons:
o where the compliance procedures indicate that the materiars used may
be
sub-standa¡d o¡ defective
Where extemal tendons are to be protected by a concrete cover added
.
:

subsequently, the concrcte should be of grade at least C40 and the thickness of workmanshio on
the cover should be not less than that required for tendons inside the stn¡ctural sign;
coucrete urder similar conditions, The concrete cover should be anchored by
. ns or in sensitive
reinforcement to the prestrrssed member and should be chccked for craok cont¡ol. structu¡al members;
. where a check is required on the quality ofthe construction.

12.12.4 Spacing ofprestressing tendons and ducts

There are no additional requirements apart from those given in Clause 8.10.1 of
the code.
level appropriate to thc servic the tests, sufñcient
measurements should be taken gn in predicting the
12.12.5 Longitudinal reinforcement in prestessed concrete beams ultimate stength and long term

Longitudinal reinfo¡cement may be added in presEessed concrete members to

increase the strength ofsections, to tie up the shea¡ reinforcement or links, or to
conhol cracking.
12.12.6 Links in prestressed concrete beams
Links may be provided in prestressed concrete membe¡s to act as shear and
torsional reinforcement, to resist the bursting tensile forces in the end blocks of
post-tensioned members, or to resisting the bursting forces arising from the
anchorage bond stresses in the ûansmission length ofpre-tensioned tendons,
13.2 Test loads
The total load should be not less th¿n 1.0 times the cha¡acteristic dead load plus
1.0 times the characteristic imposed load and should nonnally be the greater of
(a) the sum of the characteristic dead load and 1.25 times the cha¡ìcteristic
imposed load or (b) 1.125 times the sum of the characteristic dead and imposed
loads.

12,12.7 Impact loading

Where a prestressed concrete beam is rcquired to resist impaot loading, it should

be reinforced with longitudinal reinforcement and closed links, preferably ofmild 13.3
steel. Other methods of desigr and detailing may be used provided it can be
shown that the beam can develop the required ductility,
Àssessment of re¡ulh
assessing the results is to compare tbe measured
pected on the basis of the design calculations. This
should be made for any differences in material strengÛr

t50 l5l
 or stt€ss or other characteristics ir
assumed in the design. Steps shr
parameters as accurately as possit
tensioning records etc. Due alloll
environmental conditions that have <¡

13.4 Tesf criteria

In assessing test data, the following criteria should

be considered:
o Thq initial deflection and cracking should
bein accordance with the
reguirements. design
. occurred under the test loads, the
loading cycle should be at least equal
should be at least 75%o for reinforced
concrete structu¡es and gS% for Class
1 and
>tures.
o The sfiuctu¡e should be examined for unexpected
defects, which should theu
be evaluated by recalculation. Cr
results are impofant. Where thr:
should be to check that the sr.
different from that assumed in thr
or to the influence of non-stru¡:
need ûo be checked as well.

13.5 Special tests

Special tests that may be required should be
agreed in advance by all the parties
concemed.

r52

!-
t2.8.4,' \ould preferably be cut off flush with the end ofthe concrete 13 LOAD TESTS OFSTRUCTURES ORPARTS OF
^9
/a
.ioned tendons, the equivalent bar size for ducts containing
STRUCTTIRES
rld be calculated from tle total a¡ea of the st¡a¡rds iruide
cover to the oulside ofthe duct should be not less than
-, of the duct cross-sestion nor less than half the largest 13.1 General
it cfoss-section.
only the testing of whole stuctures, finished parts of a structr¡re or structu'al
,endons: components during the construction phase is covìred. Modcl or prototype testing
.ver to the ducts of unbonded tendons be sufficient to protect thÊ
and appraisal ofsl¡¡ctu¡e after a certain period of service are not includeà
.dons from conosion and fire, as per the in Sections 4.2 and 4.3
ofthe code. In any case, the nominal cover to lhe should not be less than The need for load tests may arise during construction r¡nder the following
25 mm. circumstances:
o where ttry cgmn]igcg procedures indicate that the materials used may be
Exlernal îendons: sub-standard or defective;
Where extemal tendons a¡e to be protected by a cover added
¡ where supervision and inspection procedures indicate poor workmanship on
subsequently, the concrete should be of grade at least C40 and thickness of
-tendons site, producing consüuction outside tJre specification and design;
the cover should be not less tha¡ that required for inside stn¡ctural
concrete under simila¡ conditions. The concrete cover should be
¡ where there are visible defecß, particularly at critical sections o¡ in sensitive
shrctural members;
reinfo¡cement to lhe prestessed member and should be checked for c¡ack
. whe¡e a check is requited on the quality ofthe consfuction.

72.12.4 Spacng of prestessing tendons and ducts

There are no additional requirements apart ûom those given
the code.
12.12.5 I'or.glttrdinal reinforcement in prestressed concrete
It should be recogrized that loading a structu¡e to its desip ultimate loads may
impair its subsequent performance in sewice, without necessa¡ily giving a kuc
8.10.1 of measr¡re of its load carrying capacity. While such overload tests may sometimes
be justified, it is generally ¡ecommended that the structure should be loaded to a
level appropriate to thc sewiceability limit slates. During the tests, sufficient
mcasurements should be taken to calibrate the original design in predicting the
ultimate strength and long tern performance of the structure.

Longitudinal reínforcement may be added in concrete members to

increase the strength ofsections, to tie up the shea¡ or links, or to
13.2 Test lo¡ds
control cracking,
The total load should be not less ttnn 1.0 tirnes the char¿cteristic dead load plus
L0 times the cha¡acteristic imposed load, and should normally be the grcater of
(a) the sum of the cha¡acteristic dead load and 1.25 times the cha¡acteristic
imposed load or þ) 1.125 times the sum of the cha¡acteristic dead and imposed
loads.

post-tensioned members, or to resisting the bursting forces arising from the Test loads should be applied and removed increrrentally with a 5 min interval
allowed between load increments for recording defomration measurements. At
anchorage bond st¡esses in the tansmission length ofpre-tensioned tendons.
least tuto cycles of loading and unloading should be applied, with a minimum of
t h between the two cycles. Conside¡ation may also be given 1o a third
application ofload, which is left in position for 24 h.
12.12.7 lmpactloùng

Where a prestessed concrete beam is rcquired to resist impact loading, it should

be reinforced with longitudinal reinforcement and closed lints, preferably ofmild
13.3 Assessment of re¡ults
steel. Other methods of design and detailing may be used provided it can be
shown tbât the beam can develop the required ductility.
The main the resufts is to compare the mcasured
performanc the b¡sis of the design calculations' This
means that made for anv differences in material strengfh