SHEAR STRENGTH

OF REINFORCED CONCRETE BEAMS

by

ALEXANDRE PLACAS B. A. Sc., Civil Engr., m. A. Sc.

A thesis submitted for the Degree of Doctor of Philosophy

in the Faculty of Engineering of the University of London

Imperial College of Science

and Technology, London November 1969

1
BEST COPY

AVAILABLE

Variable print quality

TEXT BOUND INTO

THE SPINE
 ABSTRACT

The invesigation reported in this thesis is concerned with

the general subject of the shear resistance of reinforced

concrete beams and the findings are applicable to rectangular,

T- and I-sections.

To assist the theoretical analysis, tests of 75 normal sized

beams, in which the parameters affecting shear varied system-

atically, were carried out. Concentrating on the types of

members where lack of sufficient experimental data existed, all

but 10 of the beams contained shear reinforcement and 45 had

compression flanges.

Virtually all the beams, having high strength longitudinal

reinforcement with good bond quality, failed in shear in one

of the following modes

1) Diagonal Tension

2) Shearing (Shearing of concrete above a shear crack)

3) Shear Compression

4) Web Crushing

The theoretical analysis is to a large extent based on

observations of the behaviour of these beams during the tests

and at failure. As a result equations are presented representing

each mode of failure. If the strengths corresponding to modes

2,31 and 4 are greater than the diagonal tension load, the

final solution is given by the mode of failure equation pre-

dicting the least strength; in other cases failure is by dia-

gonal tension. In this manner not only is an adequate numerical

prediction of the failure load obtained, but a comparison of

2
the results pertaining to the various modes provides a good

indication of the actual typo of failure that is likely to

take place.

Intrinsic to this approach is a rational solution of the

problem of compression reinforcement and its influence or lack

of influence on the ultimate shear strength of rectangular

beams. The increased strength of T-beams over that of rectangles

composed of their webs is also adequately reflected in the

proposed theory.

To establish the equation for the shearing mode of failure

a new adaption of a Coulomb type "plastic shoar" failure cri-

tenon for concreto is presented.

Sinco collapse loads are not the only design criteria, in

addition to the equations for the four modes of failure mentioned,

formulae are also developed, substantiated by test results,

to cover two other possible criteria

1) Shearing off of flanges in T-beams

2) Shcar crack widths

The various equations put forward in this thesis show

adequate correlation with test results, giving on average a

10% safety margin at a coefficient of variation of loss than

0.109

3
 ACKNOWLEDGEMENTS

The author is grateful to Professor A. L. L. Baker, D. Sc.,

F. I. C. E., F. I. Struct. E., under whose supervision the in-

vestigation reported herein was conducted.

My sincere thanks are due to Dr. P. E. Regan, Ph. D.,

D. I. C., SUR, for his advice and guidance throughout the

duration of the investigation.

My thanks are also due to the laboratory staff, in parti-

cular Messrs 3. Turner and F. Turner, for their assistance

in the preparation of the, test beams.

The author is indebted to the following organizations,

without whose cooperation the research reported in this

thesis would not have been possible: The Construction Industry

Research and Information Association (C. I. R. I. A. ) under whose

sponsorship the research was carried out; the National

Research Council of Canada for the award of a grant that made

the author's stay in England possible; The Ecole Polytethnique

of the University of Montreal for the granting of the necessary

leave of absence.

4
 CONTENTS

PAGE

ABSTRACT 2

ACKNOWLEDGEMENTS 4

CONTENTS 5

NOMENCLATURE 11

CHAPTER1 INTRODUCTION 15

CHAPTER2 CRITICAL REVIEW OF EXISTING SHEAR THEORIES "25

2.1. Introduction and general comments - 25

2.2 The morsch theories -s - 30

2,3 Laupa's theory and other semi-empirical shear com- -33

pression theories- Zwoyer, moody

2.4 Waltherls theoky and other theoretical shear com- 41

prossion theories - Ohja, Bjuggren

Plane @f minimum strength methods - Borishanski 50

-2.5
Hillerborg

296- Regan's theory 57

2.7 Other shear theories - Taylor, Petersson, Jones, Kani 64

.
2,8 Review of some research on T-beams 69

2.9 Dowel and aggreagte interlock actions 72

2.9.1 Dowel action in reinforced concrete beams 73

Beams without shear reinforcement

Beams with shear reinforcement

2.9.2 Aggregate interlock action in reinforced concrete 82

beams

CHAPTER3 DETAILS OF THE EXPERIMENTALINVESTIGATION 88

.
3.1 Concrete 89

3.2 98
Reinforcement

5
3.3 Geometry of beams and arrangement of reinforcement 108

3.4 Casting and curing 115

3.5 Test procedure 117

3.6 Parameters investigated 122

4
CHAPTER INVESTIGATION OF SHEAR FAILURE IN 125

RECTANGULAR
BEAMS

4.1 Behaviour of rectangular beams failing in shear 125

4.2 Rectangular beams without shear reinforcement 131

4.2.1 Strain measurements on concrete and reinforcement 137

4.3 Influence of the degree of shear reinforcement 145

4.4 Influence of the ratio of main reinforcement 156

4.5 Influence of the shear span to effective depth ratio 1G4

4.6 Influence of the concrete strength 167

4.7 Influence of the compression reinforcement 173

4.8 Influence of the inclination of shear reinforcement 185

4.9 Strain measurements 189

4.9.1 on the shear reinforcement 189

4.9.2 on the main reinforcement 192

4.9.3 on the concrete 194

CHAPTER5 INVESTIGATION OF SHEAR FAILURE IN T-BEAMS 207

5.1 Behaviour of T-beams failing in shear 207

5.2 Influence of tha degree of shear reinforcement 216

5.3 Influence of the ratio of main reinforcement 223

5.4 Influence of the shear span to effective depth ratio 229

5.5 Influence of concrete strength 235

5.6 Influence of the inclination of shear reinforcement 240

5.7 Influence of transverse flange reinforcement 246

5.8 Influence of flange breadth 255

6
5,9 Influence of flange depth 259

5.10 Strain measurements 263

5.10.1 on the shear reinforcement 263

5.10.2 on the main reinforcement 266

5.10.3 on the concrete 269

a) top surface

b) sides

CHAPTER 6 INVESTIGATION OF WEB CRUSHING FAILURES 298

6.1 Behaviour of the thin-webbed I-beams 300

6.2 Influence'of concrete strength 307

6.3 Influence of shear reinforcement 309

6.4 Influence of spacing of shear reinforcement 311

6.5 Influence of inclination of shear reinforcement 311

6.6 Strain measurements 313

6.6.1 on the shear reinforcement 33.3

6.6.2 on the main reinforcement 32.7

6.6,3 on the concrete 321

CHAPTER 7 SHEAR STRENGTH OF REINFORCED CONCRETE BEAMS 331

7.1 Behaviour of beams failing in shear 332

7.1.1 Shear cracking 333

7.1.2 Shear failure 339

7.2 Equilibrium conditions at an inclined plane 340

7.2.1 Behaviour of beams following shear cracking 340

7.2.2 Equilibrium conditions at an inclined plane 344

7.3 Shearing modo of failure 347

7.3.1 Singly reinforced rectangular beams 352

7.3.2 T-beams 354

7.3.3 Beams with compression reinforcement 356

7
7.4 Shear compression mode of failure 358

7.4.1 Singly reinforced rectangular beams 358

a) Closely spaced vertical stirrups 362

b) Closely spaced inclined stirrups 363

c) Bent-up bars 364

d) Widely spaced shear reinforcement 365

7.4.2 Beams with compression reinforcement 365

7.4.3 T-beams 367

7.5 Web crushing 368

7.5.1 Equilibrium conditions 370

7.5.2 Considerations on the stress-strain relationship 372

in the web

7.5.3 Formulation of equations for the wob crushing 376

strength of reinforced concrete beams

7.5.4 Validity of web crushing equation 382

7'6 Horizontal shear and transverse flange reinforcement 385

in T-beams

7.6.1 Beams without transverse flange reinforcement 385

7.6.2 Transverse reinforcement in flanges 390

7º7 Detailing requirements 395

7.7.1 Effect of shear on forces in main reinforcement 395

a) Beams without shear reinforcement

b) Beams with shear reinforcement

7.7.2 minimum requirements for shear reinforcement 402

7.8 Short and deep beams; Moment-shear interaction 405

7.8.1 Beams without shear reinforcement 406

7.8.2 Beams with shear reinforcement 419

7.9 Physical interpretation of shearing equation 425

8
 7.9.1 Failure criterion for concrete 425

7.9.2 Rectangular beams 428

7.9.3 T-beams 433

7.10 Discussion 437

CHAPTER 6 COMPARISON OF TEST RESULTS WITH THEORIES 441

8.1 Comparisons with the proposed theory 443

8.2 Comparisons with Codes of Practice 449

8.3 Comparisons with other shear theories 458

8.4 Discussion of the comparisons presented in 8.1 to 8.3 462

8.5 Influence of compression flanges - comparison of 471

rectangular and T-section

8.6 Application of the proposed theory to T-beams tested 474

by other investigators

8.7 Comparison of the proposed web crushing equation with 477

the provisions of the Draft Unified Code of Practice

for Structural Concrete

CHAPTER 9 SERVICEABILITY CRITERIA 480

9.1 Shear crack widths 480

9.2 Deflection of beams failing in shear. 494

End rotations

CHAPTER10 CONCLUSIONSAND SUGGESTIONSFOR 510

FURTHERRESEARCH

10.1 Conclusions 510

10.2 Suggestions for further. research 513

10.2.1 Aggregate interlock 513

10.2.2 Effect of axial loading on shear strength 515

10.2.3 Effect of manner of load. cpp. ication 518

10.2.4 Prestressed concrete beams 520

9
 10.2.5 Continuous beams and influence of shear on moment521

- rotation characteristics

REFERENCES 526

APPENDIX A MODIFICATIONS TO EXISTING SHEAR COMPRESSION 540

THEORY

A. 1 Failure criterion 541

A. 2 Influenco of stirrups on the compatibility conditions 549

A. 2.1 Comparison of top fibre strains 550

A. 2.2 Investigation of the effect of taking account of 561

the stirrups in the compatibility conditions as

expressed by equation (A. 2.2)

A. 3 Influence on the ultimate load of the inclination 569

of stirrups

A. 3.1 Ultimate load for stirrups inclined at 450 571

A. 3.2 Development o f an approximate equation 573

APPENDIX B TABLE OF STRAIN MEASUREMENTSON BEAU D-2 580

10
 NOMENCLATURE

The following schedule gives symbols used repeatedly through-

out this thesis. Symbols used on only one occasion are defined

in the appropriate section and are not included here.

A cross-sectional area of longitudinal compression rein-

so
forcement

A crossLsectional area of one member of shear reinforcement

q
A cross-sectional area of longitudinal tension reinforcement
st
of
A cross-sectional area of one member of transverse flange

reinforcement

a shear span; length over which the shear force diagram

retains the same sign

b breadth of rectangular beam; overall breadth of com-

pression flange in flanged boams

br web breadth of beam (; b in rectangular beams)

c projection of shear crack length measured along axis

of boom

0 overall depth of beam

dl effective depth of beam from extreme compressed fibre to

centre of tension reinforcement

d2 effective depth from extreme compressed fibre to centro

of compression reinforcement

d* depth from compressed surface to anchorage of shear rein-

forcement (i. e. in the case of stirrups, depth to lowest

layer of tensile reinforcement; and in the case of bent-

up bars depth to layer of reinforcement from which bars

11
 are bent up)

dd lateral translation, due to shear, of flexural moment

diagram

Ec modulus of elasticity of concrete

Est modulus of elasticity of steel in tension

a strain ratio = E/c

Fb bond force

Fcc component, parallel to tension reinforcement, of com-

pressive force in concrete

Fd dowel force in tension reinforcement

Fi interlock force at a crack

Fso compressive force in compression reinforcement

Esq force in shear reinforcement crossing a shear crack

Fst tensile force in tension reinforcement

fcw compressive stress in web concrete

fse compressive stress in compression reinforcement

fst tensile stress in tension reinforcement

fsq tensile stress in shear reinforcement

ft tensile stress in concrete

fy yield or proof (0.2%) stress of reinforcement in general

fyq yield or proof (0.2%) stress of shear reinforcement

k1k3Ucyi average stress in compressive zone with fully deve-

loped stress block

k2ndI depth from extreme compressed fibro to lino of action of

comprossivo force in concrete

La internal lcvor arm (botwoen tension in main stool and

compression in the concrete)

m oxtornal bending momont

12
fý1W bending moment resisted by shear reinforcement crossing

a shear crack, including moment about tensile reinforce-

ment due to longitudinal forces in concrete balancing an

ovontual horizontal component of the force in tho shear

reinforcement

Mult external bonding moment at limit state of collapse

me bending moment resisted by the couple due to the force

in the tension reinforcement crossing a shear crack and

tho balancing compressive force in the concrete

m modular ratio

ndI neutral axis depth

nedi roforence neutral axis depth corresponding to Ecc= 0.35%

with the main stool elastic

ndl ultimate neutral axis depth at failure by shear com-

pression tf
A
p ratio of transverse flange reinforcement = ts*

Q external shear force

QC shear force supported by concrete

qw shear force supported by shear reinforcement

Qwc shear force carried by web at web crushing stage

qcr shear force causing shear cracking in web

Q*r shear force causing longitudinal shear cracking in flange

Quf shear force corresponding to limit of T-beam action in

a beam with transverse flange reinforcement

Quit external shear force at limit state of collapse

q nominal shear stress; subscripts as per shear forces

A
rq ratio of shear reinforcement =
=Qbrs
sin rC A
rsc ratio of compression reinforcement = bsso
1

13
 Ast
r ratio of tension reinforcement = brdl
st
s spacing of shear reinforcement measured along axis of beam

s* spacing of transverse flange reinforcement

t thickness of flange

Ucyl cylinder strength of concrete (6" x 12" cylinders)

Uw cube strength of concrete (6" cubes)

Ut tensile strength of concrete (uniaxial)

Usplit cylinder splitting strength of concrete (Brazilian test)

w- width of widest shear crack at any given load stage

rf
factor (-q Yq)
- shear reinforcement
cyl
z 'internal lever arm in general

CX angle of inclination of shear reinforcement (angle bet-

ween shear reinforcement and axis of member)

angle between interlock force F1 and axis of beam. Also

angle between thrust line in concrete above shear crack

and compressed surface of the beam

angle of shearing failure of concrete in compressive zone

e angle of inclination of compression in web between shear

cracks

compressive strain in the concrete

c
E. t
limiting compressive strain in the concrete
cc
8o compressive strain in concrete corresponding to maximum

compressive stress.

14
 CHAPTER 1

INTRODUCTION

Ever since Ritter's and Mörsch's tests at the turn of the

century, it has been recognized that the influence of shear

forces must be taken into account in the design of reinforced

concrete. The methods of calculation used to do this have

however been far from satisfactory. A lack of fundamental

knowledge concerning the way in which shear is resisted has

lead to the formulation of theories which are frequently un-

duly conservative, and yet manage in some circumstances to

produce unsafe designs.

Up to the mid 1960's experimental shear research was

characterized by a prodigious number of tests of heavily

reinforced, point loaded, simply supported beams without

shear reinforcement, and albeit smaller number of other type

of tests: The conclusions drawn were mostly in the form of

empirical equations, and basically related to the type of

member described. Everything else was dealt with by extens-

ions therefrom. Not all the approaches to the problem were

empirical however as some investigators attempted to formul-

ate theoretical solutions. As a result there are quite a

number of shear theories presently in existence. Tho more

fully developed of these can be grouped into the following

categories.

15
(a) G1odified, ß1örsch Truss Analogy - ACI, CE8, CP114, Leon-

hardts proposals for the German Code. Even the recent Shear

Study Group Report, although based on considerations associ-

ated with (e) below, is in its proposed form very similar to

the ACI's modification of the truss analogy.

Sbi Plane of Minimum Strength Methods Borischanski (Russ-

-

ian Code) Hillerborg

(c) Semi-Empirical Shear Compression theories - Laupa,

Zwoyer, Moody

(d) Theoretical Shear Compression theories - Walther,

Bjuggren

(e) Combination of (b) and(d) - Regan

A thorough critical appraisal of these methods is pre-

sented in CHAPTER 2; this can be summarized in the form of-

the following general comments on the existing methods.

(1) The basic overall objection to these theories is that

either they assume a crack pattern and behaviour of the beam

not substantiated by the tests or that the ultimate shear

strength equations are based on the assumption of a unique

mode of failure. Provided that the crack pattern and behav-

iour are well reflected a theory may thus be acceptable for

the case in which the actual mode of failure agrees with the

An illustration of this situation is afforded

one assumed.

. by Regan's shear-compression theory which has been found

through tests to describe adequately the phenomenon of shear

failure. With certain modifications presented in

compression

16
VPENDIX At this theory forms the basis of calculations when

the mode of failure is shear-compression. It remains however

that not all beams failing in shear at loads higher than

those at which the first shear cracks appear (diagonal tens-

ion) do so in the same manner and this must of course be re-

flected in a shear theory of more than limited applicability.

(2) Important differences between the existing theories

arise not only from the choice of the parameters considered

to have an influence on the shear strength of a reinforced

concrete beam, but also from the different degrees of influ-

ence accorded to a given parameter by the various theories.

A detailed study of the particular influence of every major

parameter apt to affect the shear strength was therefore

warranted.

(3) Another criticism of the above theories is the

general

fact that both dowel and interlock action are neglected.

This may be acceptable if the theory is purely empirical, but

in a theoretical approach - attempted even by the American

shear compression theories notwithstanding their empirical

nature - these actions should either be considered, or, if

they are neglected, a more definite demonstration of the

reasonableness of the assumptions on which this is done

be presented. (Actually Regan has done the latter but

should

only in connection with dowel forces. )

(4) Most of the theories consider the ultimate shear

17
strength of T-beams as being that of the rectangle forming

their web although tests have shown that in many instances

increased shear strengths are obtained in T-beams. The only

theories that consider the effect of the flanges are Regan's

and Laupa's shear compression theories. Unfortunately the

limited test data on T-beams tend to indicate that shear

compression is not the normal mode of failure of T-beams

failing in shear.

It can thus be seen that the existing theories do not

provide an adequately general description of the behaviour

of beams failing in shear. It was therefore considered

necessary to develop a new approach to this problem whose

requisites would be to satisfy rationally the following

conditions in a given beam failing in shear.

(a) Adequate prediction of the shear cracking and ultimate

loads. An adequate prediction of the ultimate shear

strength is also a sine qua non condition if the

theory is to be used as a basis for design purposes,

since the present tendency in Codes (ACI, S. S. G.

Report is towards a limit state concept. (Russia

adopted an ultimate load basis of design as early

as 1938. )

(b) The actual physical bchaviour of the beam - crack

pattorr mode of failure - should be reflected.

(c) The influence of the various parameters apt to affect

the shear strength should be reasonably reflected.

1B
 (d) The assumption dowel
concerning and interlock action

should be compatible with the beam's behaviour and,

when possible' based on an interpretation of the

available research on the subject.

Since the number of available test results for beams with-

out shear reinforcement is large and the renges of variables

covered is comprehensive only a very limited number of such

tests were carried out; not so much to provide further evi-

dence for establishing a semi-empirical equation but mainly

to emphasize the importance of dowel and aggregate interlock

actions in that case and also, by comparison, to assist in

the understanding of the behaviour of the more realistic

beams containing shear reinforcement.

In the case of members with shear reinforcement the amount

of data available is much smaller and the ranges of influent-

ial parameters covered by it vary limited, as there appears

to have been little systematic testing involving varying one

parameter at a time. The present tests are intended to fill

this gap. They are not sufficiently numerous to cover all

possible combinations of variables, but they do include at

least one systematic investigation of each variable, with

all other parameters held constant. Particular attention is

paid to T-beams where the available test evidence on realist-

ic beams is scarce.

On the basis of the above considerations on existing

theories, on what a rational shear theory should reflect,

/with
and on the requirements of systematic testing of member

19
shear reinforcement a test programme was decided upon, in-

volving 75 beams. Furthermore it was realized at the outset

of the investigation that to produce definitive conclusions,

not only was a large number of tests necessary but realistic

sizes of beams would have to be used. The beams were thus

12" deep and their spans varied from 4' to 14', the weight

of the member in the latter case being nearly 1 ton. Another

feature common to the tests was the use of very high strength

main reinforcement in order to ensure shear failures with

realistic percentages of longitudinal reinforcement. The

beams' dimensions, the materials used, and other testing

details are described in CHAPTER 3. Because of the relatively

large scale testing carried out only simple span conditions

were used in this investigation and the loading applied was

in the form of concentrated point loads, since in the opinion

of the writer - and in the opinion of the CEB's relevant com-

mission as expressed by its spokesman Dr Soretz(108) - dis-

tributed loads are less severe, from the point of view of

shear, than concentrated loads. Soretz further considers that

it is sufficient to study shear resistance under the effect

of concentrated loads since they define a phenomenon more

stringent than what actually occurs in practice. Although

this may be so the study of shear resistance under uniform

loading should be undertaken if a rational design method is

to be produced; however this should be done only after the

phenomenon of shear failure under concentrated loads is more

clearly understood than. seems to be the case at present.

The tests themselves and the influence of the various

20
parameters on the behaviour and shear strength of reinforced

concrete beams are presented in Chapters 4,5, and 6.

CHAPTER 4 is concerned with rectangular beams and the para-

meters investigated in this case were

1) percentage of main reinforcement

2) amount of shear reinforcement

3) inclination of shear reinforcement

4) percentage of compression reinforcement

5) shear span/effective depth ratio

6) concrLte quality.

The effect of a parameter was assessed not only on the

strength of these tests but also on applicable test results

of other investigators. Particular attention was paid to the

physical behaviour of the beams.

CHAPTER 5 is concerned with T-beams and since available test

data for T-boams with stirrups is scarce this lead to a

relatively extensivo series of tests. Not only were all the

parameters listed in connection with rectangular beams in-

vestigated, but this investigation had to be carried a step

further since a T-beam can be designed for a greater flexural

capacity and can thus fail in shear in instances where such

behaviour would not be possible for a rectangle (e. g. long

spans). Furthermore certain parameters intrinsic to Ti-beams

were also investigated. These were the geometry of the flange

- breadth, thickness - and the amount of flange reinforcement.

This last parameter was studied in an attempt to investigate

the phenoirenon of shearing off of flanges in T-beams and thus

pr. oduce an experimentally based theory for the need for, and

21
amount of flange reinforcement.

As in the case of rectangular beams particular attention

was paid to the physical behaviour of the beams and use is

made of any applicable tests of other investigators.

Although CHAPTER 6 is also concerned with flanged beams

it was considered to be an entity in itself because it exam-

ines a particular mode of failure likely to occur in thin

webbed T or I beams. When stirrups act as the tensile members

of a lattice or truss system, the web concrete forms the

complimentary compression part of the lattice. If the web is

thin and the amount of shear reinforcement is large failure

in shear can occur by crushing of the web. The tests in this

series were designed to investigate the effects of the

following parameters on the behaviour of very thin-webbed

beams:

1) strength of concrete

2) ratio of shear reinforcement

3) spacing of shear reinforcement

4) inclination of shear reinforcement.

Following the presentation and discussion of test results

in chapters 40 5, and 61 equations are formulated in CHAPTER

7 for the shear strength of reinforced concrete beams. Thoso

equations are based on the actual behaviour of the beams as

observed during the tests, the failure loads obtained and, a

theoretical interpretation of of ttts problomvof shoor resist-

ance c mpatible witb, the obovo m.entionod bohaviaur's Of-beams fail-

ing in shear. Thcoc equations are in ogroemont with the viows pro-

22
scntbd prcviously, rogurding tho roquircmonts of a shoar theory.

Basically they represent different modes of shear failure

but their field of application is not prescribed in an

arbitrary manner but is dependent on minimum strength con-

siderations.

In CHAPTER8 the ultimate shear strength predictions of

the proposed theory are compared to the predictions of six

of the existing theories, using as a criterion the failure

loads obtained in the tests described in chapter 4,5, and 6.

Additional comparisons to illustrate the influence of flanges

in T'-beams are also presented and the increase in strength

of T-beams intrinsic to the proposed theory is further sub-

stantiated by comparisons with the test results of other in-

vestigators.

Although limit state design concepts are based on ultimate

loads it is possible that considerations other than strength

may determine the maximum design load in a given case. In

the present context these other considerations are those of

serviceability involving control of crack widths, deflections,

and local damage. Because of its structural nature, local

damage - e. g. shearing off of flanges - aspects have been

included in chapter 7 but the question of deflections and

crack widths is examined in CHAPTER 9. Considerable attention

is paid to the hitherto neglected subject of shear crack

widths and crack measurements with a graduated telescope

were taken in every beam tested; shear cracks tend to become

wider than flexural cracks shortly after shear cracking and

a knowledge of their development is necessary since the 'limit

23
state, concept in design may sometimes produce design loads

higher than those causing shear cracking. In this chapter

therefore an equation is presented for the prediction of

shear crack widths.

In CHAPTER 10 a summary of the conclusions of the present

theoretical and experimental investigation is presented to-

gether with some suggestions for further research, the

necessity for which became apparent in the course of the

work reported in this thesis.

24
 CHAPTER 2

CRITICAL REVIEW OF EXISTING SHEAR THEORIES

2.1 Introduction and General Comments

a
A critical appraisal of the existing theories and design

methods is instructive in that it gives an indication of the

types of approach that are possible; furthermore from their

study an evaluation can be made of the implications of the

various assumptions employed in their derivation.

It is generally accepted that the appearance of inclined

shear cracks extending to the level of the main steel is a

prerequisite for the shear failure of a reinforced concrete

beam. With the formation of shear cracks a redistribution of

internal forces occurs and the structural behaviour of a beam

changes in that the distribution of forces in the tension

reinforcement in the beam no longer follows the distribution

of external moments.

In most beams with shear reinforcement and in the case

of some short beams without ultimate loads considerably in

excess of those causing shear cracking can be withstood. On

the other hand in the case of most beams without shear rein-

forcement a sudden failure is the immediate consequence of

the formation of shear cracks.

It can be seen thus that the study of both shear cracking

load load is important in any given

and ultimate case as the

former determines the need for shear reinforcement and the

latter the amount needed to avoid shear failure.

25
 A considerable amount of work has been done in both these

fields in the last two decades and a great number of tests

on rectangular beams without stirrups and to a much lesser

degree on rectangular beams with stirrups performed. The new

shear cracking equations are mainly empirical but in compar-

ison with older codes of practice they make considerable re-

ductions in the critical shear force to be withstood by the

concrete. They generally take account of the influence of

the main reinforcement and in some cases make allowance for

the influence of the ratio It is the writer's opinion

Qä .

that all equations for the shear cracking load of a beam

will have to remain at least semi empirical till more is

known on the question of dowel and, more importantly, of

interlock action. latter can be quite important in the

"The

vertical cracks of small width existing before shear crack-

ing takes place; it can be expected to be much less important

in the case of ultimate failure of beams with stirrups where

large inclined cracks exist prior to failure and are capable

of absorbing the vertical displacements that are instrumental

in the creation of interlock action.

Because of the modern tendency to put at least a minimum

amount of stirrups in a beam it was decided in this research

to dwell mainly on the shear failure of beams with shear re-

inforcement which collapse at loads higher than those pro-

ducing shear cracking; as a result the various theories

examined are viewed mainly in terms of the shear failure

loads they predict rather than shear cracking. However shear

cracking equations are included in some of these theories,

26
which are then reviewed from both the cracking and ultimate

failure aspects. Some theories developed uniquely for shear

cracking are reviewed in section 2.7. The questions of dowel

and interlock actions are discussed in section 2.9.

Returning now to the case of beams in which the shear

force exceeds the shear cracking load, a simple improvement

on the inaccurate 1örsch truss analogy can be obtained by

increasing the calculated shear resistance with the addition

of a term representing the contribution of the concrete.

This approach is used by the ACI-ASCE Joint Committee 3264

as well as Leonhardt's latest proposals accepted on an in-

terim basis by the CEB26 in both of which the concrete's

contribution is taken to be approximately equal to the shear

cracking load. Although these design methods are an improve-

ment on Morsch's as far as predicting ultimate loads is con-

cerned, they nonetheless contain a märsch type term for the

effect of the web reinforcement which is tantamount to the

assumption of 450 cracks; test evidence does not substantiate

the adoption of this unique crack inclination. Another

criticism of this approach is the fact that failure of beams

with stirrups generally takes the form of a longitudinal

compression or as indicated in chapter 7 shear failure of a

compressive zone that has been reduced by the rise of shear

cracks, while the above equations are phrased in'terms of

transverse forces only.. Although this can be done in some

cases it cannot have the universal application implied in

these approaches. Because of their importance as National

Building Codes the methods proposed by the ACI-ASCE Committee,

27
the CEB and the modification to CP11421 proposed by the

Shear Study Group107 are reviewed extensively in Chapter 8

where the results of the present series of tests are com-

pared to these design methods and other theories including

the one developed in this thesis. - mörsch's original theory

is reviewed in section 2.2 of this chapter.

The majority of the theories dealing with the actual

cause of shear failure can be divided into two groups. Those

predicting a shear-compression type of failure and those

based on the concept of a plane of minimum strength.

Shear-compression theories treat the failure of the beam

as being caused by the failure in compression of the concrete

in the zone at the head of a shear crack. The American

approaches, reviewed in section 2.3, were semi empirical and

in deriving them the authors were primarily concerned with

point loaded beams without shear reinforcement. Subsequent

work by European authors was much more theoretical and far

more general in application. These. methods are reviewed in

section 2.4. They are very complicated and difficult to use

in practice; furthermore they fail to give any real means of

predicting the crack pattern which determines the structural

action of a beam and must therefore influence its strength.

Although the adoption of a typical crack pattern may be sub-

stantiated in some cases - the treatment of T-beams in this

thesis is an example - the generality implied in these theo-

ries tends to cast some doubts on their analysis of a given

cracked structure.

The theories based on the concept of a plane of minimum

28
strength attempt to provide a solution to this problem of

crack geometry. These methods, reviewed in section 2.5 give

general equatiorrIrelating the shear strength at any possible

crack, that is at any inclined plane, to the geometry of the

crack. The overall strength of a beam is determined by the

plane which first reaches a critical condition. In other

words when all cracks are regarded as possible,, failure will

take place on the crack (inclined plane) on which the mini-

mum strength can be mustered. Viewed frone a dosign point of

view these methods require that the shear reinforcement placed

to cross each possible crack should be sufficient to ensure

that the total shear resistance of concrete plus shear rein-

forcement be equal to the desired ultimate shear strength.

The main criticism that can be levied against this method is

the fact that in the case of beams failing in shear com-

pression, equations of transverse forces are used to predict

ultimate loads while failure is due to longitudinal actions.

A recent development of both the shear compression ap-

proach and the minimum strength plane concept has been

carried out by Regan94 who has combined the two and produced

a shear compression theory in which the depth of the final

compression zone at the head of a shear crack is related to

the geometry of the crack. The overall strength of a beam is

determined by the crack at the head of which critical com-

pressive conditions are reached first. This theory, reviewed

in section 2.6, seems to be the most rational approach to

the shear compression type of failure in beams. Apart from

some modifications carried out in Appendix A of this thesis

29
the only major criticism of this method lies in the fact

that not all beams failing in shear do so in a shear-com-

pression manner.

In the remainder of this chapter various other shear

theories and research are discussed in section 2.7 while

section 2.8 reviews the rather limited research that has

been carried out on the specific subject of ultimate shear

strength of T-beams.

2.2 The Morsch Theories

45
The traditional Ritter-florsch truss analogy has been

the basis of shear reinforcement calculations in the codes

of the West European countries. It is still considered the

basis for the CEB code and CP114, and the ACI/ASCE code is

basically ofthat Accord-

also an empirical adapt3t3on method.

ing to this analogy the structure of a beam containing shear

cracks is likened to a pin-pointed truss. The compression,

chord of that truss in, constituted by the compressive zone

of the concrete and is parallel to'the compressed surface;

the main tensile reinforcement forms the tension chord. The

lattice between the two chords is formed by the shear rein-

forcement in tension and the concrete in compression, the

latter in the form of struts inclined at 45° to the main re-

inforcement.

The conditions of vertical equilibrium at a point in the

tension chord and at a section through the truss give the

following equation in a beam with vertical stirrups, when

30
the concrete in the web is assumed to be evenly stressed.
fA
f sin 45 cos 450 = Sý
WC s
fAL
q sq q- a
and s

where fwc is the compressive stress in the concrete struts

fsq is the tensile stress in the stirrups

This theory neglects the contribution of the concrete in

resisting shear through an arch action by the inclination of

the main compressive thrust as well as any dowel action.

Thus it underestimates the shear resistance of a beam in

varying degrees and comes close to reality only in beams with

heavy shear reinforcement when the inclination of the main

thrust is less pronounced (see Appendix A/2) and the shear

force carried by the concrete relatively small. Furthermore

as can be seen from the photograph of the beams tested the

450 inclination of the shear cracks implied in this analogy

is not substantiated"

Some of the code methods derived from the Morsch truss

analogy - ACI/ASCE, CEB - add a term to take account of the

contribution of concrete. An additional inherent wecknoss of

this procedure is the fact that the contribution of the con-

crete is thus independent of the amount of shear reinforce-

ment. Tests however show that in most rectangular beams

shear reinforcement hinders the propagation of shear cracks

leaving thus a largor resisting area of concrete intact.

In the case of shear cracking Mörsch proposed the well

known equation for the shear stress at any cross section in

a reinforced concrete member.

31
 qQ`bL
ra

Since longitudinal tension in the concrete is ignored and

vertical and hcrizcntal stresses are equal this value re-

presents the shear or diagonal tensile stress in all the

tension zone below the neutral axis. Above the neutral axis

the magnitude of the shear stress reduces to zero at the top

surface. The fallacy of this method can be visualized by the

fact that if a suction through a flexural crack is considered,

the method implies shear stresses in the cracks. Even though

forces may be transferred in these cracks through aggregate

interlock, this was certainly not considered by morsch, and

the definite value given to these stresses by the Morsch

equation is not reflected in the limited aggregate interlock

tests reported so far.

In spite of its shortcomings this equation is still used

in all European codes which differ only in the value of the

shear stress q permitted in design. The discrepancies bet-

ween the various codes are quite considerable but their per-

missible values are generally in excess of those allowed by

the ACI/ASCE equation. The British Code as it stands at pre-

sent CP114 - 1957 - revised 1965 allows permissible shear

stresses almost twice those of the ACI code. Although it is

recommended that the design of members without shear rein-

/footingsInd
forcement should be limited to slab; members of

secondary importance, the difference between this code and

the ACI is far too important. This situation however is re-

medied by the S. S. G. report (see comparisons of chapter a).

32
 2.3 Laupa's Theory and Other Semi-Empirical Shear-Com-

Pression Theories

LAUPA61

Laupa's. theory is the only one of the early shear-com-

pression methods, which is sufficiently simple and general

to be of practical significance. It was presented in 1955

and its accuracy was reasonably supported by a considerable

amount of test data. The authors of'this theory, Laupa,

Siosss and Newmark reviewed the results of previous invest-

igators in the field of shear and diagonal tension and esta-

blished a general expression for the shear strength of rein-

forced concrete beams subjected to shear and flexure. A

basically empirical equation was derived for the shear

strength of simple span rectangular beams without web rein-

forcement. The equation was extended to include beams with

web reinforcement. The same equation was modified to apply

to simple span T-beams and restrained beams.

It was assumed that the total shear force was resisted

by the compression the

solely area of concrete and that the

final failure of a reinforced concrete beam without web re-

inforcement due to combined flexure and shear was essentially

the same as that due to flexure' the only difference being

that the zone of concrete in compression was reduced by

diagonal cracks which extended higher than the flexural

cracks at failure; thus the maximum moment the beam could

carry in the presence of a diagonal crack could be much

smaller than in the case for pure flexure. Using this con-

33
 cept of shear-compression failure, an equation expressing

the shear moment Ml for beams without web reinforcement was

s
derived as follows:

1,11
C1 k2n) k1k3
2u^n -
bdS
1 cyl

In this equation the depth of compression of concrete re-

mained the only unknown. The solution to this had been

attempted with the aid of strain relations for beams which

failed in flexure and the results obtained were found to be

incompatible with test observations. Consequently no theo-

retical relationship was possible for the depth of compress-

ion zone at shear failure. It was realized empirically that

there oxists a similarity between the functions representing

and Pi, where n is the neutral axis factor calculated by

normall elastic theory. On this assumption the basic equation

can be rewritten as

S
(Ucyl) (2.3.1)
.............
bd2 u=nF
1 cyl

where F(Ucyl) is an empirically determined function of the

cylinder strength. It was evaluated for simply supported

point loaded beams whose Ucyl varied from 1000 to 6000

lbf/in2 and is as given

445U
F (Ucyl )' = 0.57 w --
1 05

EquatiocL 2.3.1" is appropriate for beams without compression

reinforcement; when such reinforcement is present, the equi-

valent equation is

34
 m1
`lr, + mrsc) F(Ucyl)
bd2 us
1 cyl
10000
where m is the modular ratio takes as 5
cyl

The final equation thus for beams without shear rein-

forcement is

4.5U
S 1)
(n + mrSý) (0.57 (2.3.2)
- -- 5 ......
bd1 Ucyl 10

The effect of the shear reinforcement was also investi-

gated in about 80 beams. Its contribution was given by the

following empirical equation

ms 2r f
-I+ 10
ml
s
where ms is the ultimate shear compression moment of a beam

with shear reinforcement and ml that of an otherwise similar

s
beam without shear reinforcement.

The above relations were found to give satisfactory agree-

ment with test results of 186 beams - in terms of average

Qcalc from the mean

that iss because important deviations

, Xis varied from 1.17 to 4. B0 and Uc from

- whose d values Y1
1000 to 6000 lb/int.

Since the moment-rdtation relationship of a T-beam differs

from that of a rectangular beam, Laupa extended his shear

compression equation to T-beams by the introduction of a

shape factor:
IT + ICR
Ft
IR+ ICR

where IT refers to the uncracked T-beam

35
 IR refers to an uncracked rectangular beam of breadth

equal to the flange breadth of the T-beam

ICR refers to the "straight line" transformed section

of either the rectangular or T-beam.

Ft is thus a semi-rational factor representing a comprom-

ise between the ratios of the momentsöf inertia in the

cracked and uncracked states; it is defined in this way,

because the extent of flexural cracking at the moment of

shear failure is unknown.

Equation (2.3.1) is thus modified to:

f111 4.5Uc
1
0.57 - -------
AccdlOcylEt
105

where Acc is the area of the compressive zone of the T-beam

and replaces the bnd1 term of rectangular beams. When shear

reinforcement is present the same method as in rectangular

beams is used to determine their contribution with the rq

term computed with respect to the web breadth.

Laupa's equation 2.3.2 was slightly transformed by R.

Jones52 to take better account of the concrete strength. By

writing equation 2.3.2 in the form

ms 4.5U
cL' 1
0.57 - --- =B
A 105

where A= bd2 Uýyl(n+mrsc)

1
ml
Ä
he observed that the ratio is not a straight line as

represented by the quantity 8 but a parabola represented

approximately by a reduction of the form '

2
8=C1-C2Ucyl+C3u
cyl

36
where C1 C2t C3 are constants.

Subbiah and Smith109 consider that Laupa's equation gives

better results for Ucyl < 6000 lbf/in2 while Jones' trans-

formation applies better to Ucyl > 6000 lbf/int.

In the case of long beams without reinforcement Laupa set

a
an upper d limit of 3.4 to 3.7 for the applicability of his
1
equation, as failures at a/d1 values larger than these occur

immediately upon shear cracking.

The various criticisms of this method are as follows:

1) The mode of failure in beams is not always one of shear

compression even within the limits of applicability of

Laupa's method. This is quite manifest in T-beams and

this criticism was also levied by Ramakrishnan91.

2) The affect of the compression reinforcement is, taken into

account at all times, and the universal influence of

this parameter as implied by Laupats method does not

agree with the results of other investigators, including

the writer. even in the tests considered by

-Actually
Laupa, five of the six results in which his theory gives

unduly large values were those of beams with com-

Qcalc
test
pression reinforcement.

3) There is an oddity involved in the use of the web breadth

for calculating the ratio of shear reinforcement in T-

beams, when in all other respects such beams are treated

as variants of a rectangular beam of breadth equal to the

flange breadth of the T-beam.

4) The shear moment equation depends only on the properties

of beam.-cross-section and suggests that this. moment is

37
 independent of the length of the shear span. This is

contrary to experimental evidence and leads to higher

Qcalc
values in the short beams than in the long ones.
Qtest
5) The linear effect of rqfyq is not substantiated at higher

degrees of shear reinforcement.

ZWOYER126

The first shear-compression theory was presented by

Zwoyer + Sieas in 1954. Based on this concept an empirical

formula was developed to predict the ultimate shear strength

of prestressed concrete beams and which was then extended

to ordinary reinforced concrete beams. This method considers

that failure in the reduced area concrete was due to crush-

ing at some limiting strain. As the assumption of linear

strain distribution is not valid when the diagonal tension

cracks rise above the flexural cracks, the crack pattern was

used to suggest an empirical method for determining n.

Zwoyer considered that since inclined cracks are directed

to one point of the compression zone most of the rotation

and compression strain would be at the zone above the crack

heads and this region would act as a hinge. An average value

of the concrete strain at the top surface of the concrete

and at the level of steel was determined empirically for

prestressed beams, reinforced beams without shear reinforce-

ment and reinforced beams with shear reinforcement' and used

to define the depth of the neutral axis. Using also a con-

stant value of k2=0.42 - where k2n d1 is the depth from the

compr©ssod 'surface to the line of action of the compressive

38
force in the concrete - an expression for the ultimate shear

moment was given.

Zwoyer's conclusion that the shear failure was caused in

effect by crushing in the compression zone of the beam at

the maximum bending moment in the shear zone supports Laupa's

approach to the problem although the consideration behind

his empirical approach to the determination of n are differ-

ent.

The general criticism of shear compression methods in

which, furthermore, the critical crack length is not taken

into consideration have been described in the introduction

and in connection with Laupa's theory. They are also applic-

able in this case with the addition of the following points

intrinsic to Zwoyer's method.

1) The values of the ultimate average compressive stress in

the concrete at the reduced zone are quite high. This

probably stems from not considering any force as being

carried by dowel action in beams whoeo details suggest

the opposite.

2) The expressions for the average strain in various types

of beams seem inconsistent with one another.

MMOODY78

This is an entirely empirical approach derived from the

tests of 42 simply supported and 36 restrained rectangular

beams with a/d1 values varying from 0.57 to-3.33. Equations

for initipl diagonal cracking load and ultimate shear

strength of beams are presented.

39
 The shear cracking load is considered to be dependent

upon the cross section dimensions and the quality of con-

crete. An empirical expression for its calculation is given

Qd
in terms of Ucyl, bdl, and e maximum limit corre-
, with

sponding to Ucyl=5000. This relationship is quite unsatis-

factory even from an empirical point of view as variations

±30% in the the tests
of are observed comparison with used

in its derivation.

To determine the ultimate shear strength a complicated

empirical expression for the maximum moment at shear failure

is given. It is basically composed of two terms: the first

term represents the moment that can be supported concrete

and the tensile and compression reinforcement in a manner

analogous to simple bending; the second term is the product

Fq x at the moment of the force in the stirrups crossed by

the shear crack.

Both terms are expressed empirically to fit the test

results. In practice the determination of the exact position

and value of Fq is not rationally possible on the basis of

''given
the consideration by this theory and can only be based

on the empirical expressions given which make the method too

complicated and restricted for any practical use. Its general

approach however is similar to that of Laupa which is much

clearer.

The same general criticisms made for Laupa's shear com-

pression theory are also applicable to this method.

40
2.4 Walther's Theory and Other Theoretical Shear-ComEression

Theories

WALTHER119

The more recent form of R. Walther's theory is contained

in a 1964 CEB article entitled "Le calcul de la resistance

ä l'effort tranchant des poutres en baton arme ou an baton

precontraint" and reference(119): both these are translations

of the same article that appeared in Beton und Stahlbetonbau

in November 1962.

The basic assumptions and considerations involved in this

method are as follows:

1) The Navier-Bernouli assumption that sectionremains plane

is abandoned. Conditions of compatability of deformations

are formulated, valid after the formation of cracks, be they

flexural or shear.

2) When a beam is subject to bending and shear, the com-

pressed zone of concrete is subject to normal stresses ';'

due to bending and shear stresses 4 (depending on the ab-

sence or presence of shear reinforcement the concrete takes

up the whole or a part of the shear force) and the resist-

to shear-compression Upt (Schubdruck festigkeit) ig

ance

always smaller than the prism strength Upr, by a margin

that is proportional to the shear force associated with the

bending action.

3) In order for equilibrium condition to be formulated, the

S 1""stresses
distribution of both and at the compressed

zone is considered uniform. This assumption is rather out

41
of place in such a particularly theoretical approach as

Walther's where some secondary points have been analysed

in more depth. Since the type of failure envisaged here is

not of a plastic nature the distribution has no theoretical

reason to be uniform and it is possible that the critical

e ?
combination of and giving the critical - for the

strength of the beam - value of apt may occur at an earlier

load stoge than that to the

corresponding critical combin-

ation of the average stresses.

4) Based on equilibrium and deformation conditions the shear

strength is expressed in terms of an ultimate moment at

shear failure - Mult'

5) Shear reinforcement has the following effects

(a) It reduces the deformation due to shear

(b) It increases the resistance to shear-compression by

taking up part of the shear force; this reduces the

7'
shear stresses acting on the compressed zone and

consequently delays the development of the limiting

stress Upt.

(c) Contributes directly in the increase of mult'

6) The method is based on conditions in a part of a beam

containing a crack inclined at 450 to the axis of the beam.

This assumption is not substantiated experimentally and as

mentioned in the introduction constitutes one of the theoret-

ical drawbacks of this method,

The application of the above to the formulation of equat-

ions will now be described briefly.

Using a Mohr's envelope based on a ratio of tensile to

42
compressive strength for concrete equal to 1/8 and the

stress conditions in a beam, Walther arrives at the follow-

ing

1+3.2 (2.4.1)
upr Q0dl .......
2

where Q0 is the portion of the shear force carried by the

concrete. It is taken as Q when no shear reinforcement is

present and as Q(1-rj) in the presence of shear reinforce-

ment; in this last expression q, is a term strongly reminis-

cent of the Mörsch truss analogy called "degree of shear

safeguard" (Schubsicherheitsgrad) and given by the following

expression
AfL
_a yq a
S Qult sin=`-

rl thus indicates the proportion of the shear force which

can, at failure, be absorbed by the shear reinforcement and

which therefore no longer acts in the compression zone.

It can be noticed that when rV--1 (full safeguard, indi-

cating that the shear reinforcement is able to take up the

U
whole of the critical shear force) the ratio = 1, i. e..
pr
the shear compressive strength attains the prism strength.

Equation 2.4.1 takes no account of the influence of any

vertical stresses on the strength of the compressive zone.

This has no significant influence for values >3 but

Qrd
1
may be quite important at lower values of this ratio. It is

approximately taken into account with the aid of a chart

given by Walther. This chart however is based on rather em-

pirical relationships and does not describe very well the

43
actual conditions since neither the position of the concen-

trated load is mentioned nor the value of the actual verti.

cal stresses produced is given.

In a region of relatively large bending moments and

moderate shear, the deformations of a beam are considered

to be primarily reducible to a rotation about the end of

the shear crack (Shubrotation). The following relationship

is thus obtained between the deformation of the extreme

compressed fibre ( and that of the main reinforcement

cc)

st).

cc n since (2.4.2)
41 1-n...........
st

where io the angle between the crack and the main rein-

forcement. With more severe conditions of shear stress there

will, in addition to shear rotation, be deformation of the

This is takan into account by a coefficient Ký. 1.0.

web.

Equation 2.4.2 then becomes

KZn sin
cc (2.463)
n ...... º..
st
Qccr)(1 Kln)
where K(Rult and Ki = 1.0 for
cr
inclined and 0.75 for vertical shear reinforcement. The

shear cracking load qcr =2x bdl. 0.654Uw when kgs and

cm units are used. It is relevant to the conditions at which

the first shear cracks penetrate to approximately the middle

of the web.

The deformation of the concrete LýN is taken as

cc
Fc d1 I. The dependence of this deformation on the
c,

extreme fibre strain in the concrete is of course reasonable.

44
The factor dlj/-n is justified by Walther on the basis that

(
it indic as the neutral axis moves closer to the com-

pressive face of the beam, the region which cooperates in

the shear rotation will be smaller. The strain depends

c
upon the extreme fibre stress = Upt attained at failure

and Walther assumes the following relationship for its

evaluation:
ýt
1 2
ýcc
u
pr
cc
where is the ultimate strain in the concrete in the

absence of shear and taken as 0.003.

The deformation of the tensile zone (! ý depends upon

st)
the steel strain e the extent of the region disturbed by
st, d1(1-n)
the shear crack, taken as the quality of bond -
sing-'
the influence of bond is taken into account by an empirical

coefficient Kb, related to the concrete strength, the ratio

of main reinforcement and the surface condition and diameter

the '1V which in

of the bars - and on a function of term

this case must be constrained to be used to take account of

the magnitude of the bond stress and of the influence of

the shear reinforcement on the deformation of the tensile

ýý
Zone.

is then expressed as:

st
. stdl(l-n)Kb
st - 2sin 9.
lU

Upon substitution of the expressions for the deformations

into equation 2.4,3, together with the equilibrium condi-

tion Astfst = Upt bndl the following expression for the

45
neutral axis factor n is obtained when the angle c. is

taken as 450
3 £1
ý
n3ý2 ccr st st
K` KbU
pr

the moment at failure is then

n1ult (1 2)
= Uptbd12 n + ý1q

where mq takes account of the direct contribution of the

forces in the shear reinforcement to the shear failure

moment: For evenly spaced vertical stirrups

n. Qultd1(1-n)
111q=2

This failure theory is admittedly-even by its own author

- not very suitable for practical applications, particularly

as the second modulus of elasticity for steel should be

used for Est, and is generally unknown until the ultimate

moment and consequently the ultimate steel stress are deter-

the terms V1
mined. Furthermore-the equation for n involves

and Kr , containing the degree of shear safeguard n, which

in turn involves the unknown ultimate shear force.

Because of these difficulties Walther has rephrased his

equation in more practical terms involving additional co-

officients and has given a number of charts and tables to

assist in calculations.

Apart from its complexity, the iteration involved in what

has just been mentioned, and the various criticisms con-

cerning the crack angle, the distribution of stresses in

the compressed zone and the coefficient Kr already made,

there are a number of other more or less questionable points

in this method.

46
 1) Although a 450 crack is assumed with a horizontal

projection of d1(1-n), the value of `is calculated on the

basis of stirrups acting along a length equal to the lever

arm.
(1-n)dl
2) The term representing a measure of the extent
sin ý

of the disturbed region in the expression for the deform-

"quite
ation of the tensile zone seem arbitrary. It does avoid

cc > is
an infinite value for when 300 but so could a
st
number of other forms of that expression.

3) According to U'alther's method shear strength increases

if bond is imprnvad It may be correct if taken literally

but in terms of ultimate strength of a beam, Walther's own

tests - in connection with Leonhardt64 - have indicated that

due to the fact that no shear cracks can form in an un-

bonded beam, such beams may have a greater ultimate strength

than otherwise similar bonded beams failing in shear. In-

creases6f strength in unbonded beams were also observed by

Evans36.

Although this method represents an attempt at the basic

reason for shear failure no evidence has come forth from

the great number of tests conducted by Walther and Leon-

hardt that it gives a satisfactory correlatiän with test

results.

OHJA83

This theory is similar to Walther's in that it uses his

shear rotation principle. The failure criterion for the

concrete of the compression zone is based on the shear dis-

47
tortion energy theory and seems to give adequate agreement

with the shear and compression tests of Bresler and Pister19

Ohja introduces some possibility of variation of the

critical shear crack and gives the following expression for

its value
2Qd1
beams without tan P-
reinforced shear reinforcement

450
1.5Qd1
reinforced beams with tang
-M
1.8Qd1
prestressed beams tanz, (-
m

9 X45°

In the tests these values are not substantiated since in

the shear compression failures the crack length seems to be

also quite dependent an the degree of shear reinforcement

and not only on its presence. On the other hand when the

failure was not of the shear compression mode a more or

less constant value of the crack inclination was obtained.

The method is quite complicated for any practical use,

is restricted to rectangular beams, and the author did not

perform any tests to verify it but used other investigators'

results. Of 215 such results used 22 had shear

only rein-

forcement. This disproportionate use of available test

results inevitably casts some doubts on the applicability

of this method in beams with shear reinforcement. The reason

of course is that rectangular beams with shear reinforcement

fail generally in shear compression in which case the crack

angle is quite important and as mentioned previously depends

to a large extent on the degree of shear reinforcement, a

condition which is neglected in this theory.

48
BJUGGRENIS

This theory involves the use of equilibrium equation and

compatibility conditions on a section of the boam bounded by

an inclined crack: The horizontal projection c of the crack

varies with the beam considered.

The theory considers the failure of the compression zone

of the beam under the action of the thrust produced by arch

action; the. failure is assumed to be caused by crushing of

the concrete at the crown of the arch and not to be influ-

enged by any shear stresses present in this region. Thus,

once the neutral axis depth at the crown has been determined,

the bending moment at the failure section can be calculated

C Fst FJct is
directly as x Zst where the force in any bar

crossing the shear crack and zst its lever arm about the

centre of pressure of the force in the concrete at the crown;

the stress in the shear reinforcement crossing the crack is

to be equal to that in the main reinforcement (but

assumed

fyqý.

To determine the neutral axis depth at the crown of the

arch Bjuggren uses a Walther type compatibility condition

however depends on the integration of strain in the

which

main steel and in the concrete, above the crack. Because of

the fact that the exact shape of the shear crack is unknown

and of the complexity intrinsic in the shape of the thrust

line in a general case, several approximations - simpler for

a straight thrust line, more complex in a general case - are

given for the values of the concrete and steel deformation-

in terms of the stress-strain relationships of the materials,

49
the horizontal projection of the critical shear crack and

the thrust inclination.

The most serious disadvantage of this method is the fact

that the determination of critical sections and cracks is

a matter of trial and error; there is also a theoretical

objection to the compatibility condition formulae as pre-

snted by Bjuggren - though not to the concept of integration

of strains - because when applied to a vertical crack they

give an infinite value to the neutral axis depth factor.

(Such discrepancy is avoided by Walther and Ohja. )

It remains however that the principle of a freely vari-

able crack length entering into the compatibility condition

is a logical approach to the study of shear compression

f ailur©s.

2.5 Plane of minimum strength methods: Borishanskiý Hi11er-

borq

This approach to the problem of shear is the basis of

the Russian Code of Practice and stems from the work of M. S.

Borishanski. The latest version of Borishanskits theory is

contained in an article16 entitled "Calculation for rein-

forced concrete members subjected to transverse forces"

published in Russian in 1964 and translated by the writer

and V. Anangelovic. A synopsis of this translation is ob-

tainable from the Concrete Structure and Technology Section

of Imperial College. Another source for this methol a 1964

CEB article17 by Borishanski entitled "Calcul ä 1'effort

50
 tranchant das elements f1echis en beton arme d'apres les

normos sovietiques". Hillerborg's43 method is a variant of

the Russian approach and its translation from the Swedish

is due to Or P. Regan.

The main interest of these methods lies in the fact that

they try to get away from the concept of a fixed crack

angle, implying a fixed direct contribution of the shear

reinforcement. This fault is common to the methods derived

from the morsch truss analogy, the empirical shear-com-

pression equation and Walther's theory. Ohja's less rigid

implications were not a success and Bjuggren's attempt al-

though logical is not systematic.

BORISHANSKI

The shear strength of a beam is composed of the contri-

butions of the shear resistance of the concrete and the

shear reinforcement. If a shear crack propagates it travels

both along the beam and into the compressive zone and on

one hand increases the number of stirrups crossing it but

on the other hand decreases the contribution of the concrete.

Since the length of the critical crack is not khowh a priori,

any crack can be assumed to be possibly critical, but, natut-

ally, failure will occur along the line of the crack at

which the sum of the two components of shear strength is a

minimum. This determination of the critical crack forms the

basic novelty of this method and since cracks are idealized

to straight lines the problem becomes one of finding the

'plane of minimum strength".

To determine the part of the shear force taken by the

51
concrete, Qc, 75 beams were tested, and they showed that Qc

depends on the following:

- the beam dimensions

- the concrete quality

- the inclination of the shear crack

Because of the complex state of stresses in the concrete

near an inclined crack no theoretical justification of the

proposed formula has been attempted. The reasoning given

by ßorishanski in reference 17 for the inclusion of the in-

clination of the shear crack as a parameter is that QC is

equal to the sum of the vertical projection of the principal

stresses, whose direction is determined by the crack angle.

QC will thus be greater when the slope is steeper. Although

one can agree with this conclusion, on the basis that the

larger the crack, the less concrete is left to resist shear,

Borishanski"reasoning is rather doubtful.

The empirical expression for Qc is thus

0.15 bd12 Rh
Qc (2.5.1)
C ...........

where Rh is the design compressive strength of concrete -

actually Borishanski took this value to be the cylin-

der strength when comparing the prediction of his

formula to the results of American investigators.

c is the length of the horizontal projection of the

crack - or plane - considered

Borishanski also notes that if in a beam without shear

reinforcement the inclined crack extends from the support

to the of load, formula (2.5.1) is

point of application

then in fact the constant moment relationship of'other

52
shear theories. To obtain a minimum contribution from the

only component present (concrete) the longest c must be

used, i. e. the above described crack in case of point loaded

beams, and the expression becomes

bd 2
mcin
Qult = 0.15 Rh (a)

The contribution of the shear reinforcement is taken as

QWý LAt ) (2.52)

t+ since. .......

where A is the shear reinforcement cross section

it it it It stress

t refers to vertical shear reinforcement

« if 11 shear reinforcement inclined at an

angle t

when vertical stirrups only are used

Qw c qX

where qx =tst A 'ý

and is defined as the shear force

carried by the stirrups per unit length of the beam.

The value of the stirrup stress used in the expression qx

at failure is taken to be 0.8fyq. This seems justified by

the fact that all stirrups crossing an inclined plane are

taken into account, which means that even those crossing the

uncracked concrete zone are considered (the plane is con-

sidered to extend from the level of the steel to the top

surface of the beam).

Thus if uniformly spaced vertical stirrups are used, the

shear strength of the boom can be expressed as

Q= QC + QW

0.15 bd12 Rh
{ (2.5.3)
+c qX ...........

53
 The most unfavourable inclination of the plane will then

be that for which the second member of this inequality is

minimum. This can be expressed by the condition

c _
Differentiating, the value of c is obtained as
/0.15bd12 Rh
c _-l (2.5.4)
4 .........
Yx

Substitution of this value of c in (2.5.3) gives the

minimum and therefore critical value of Q for a point

loaded beam.

Qult 0.6 bd12 Rýqx (2.5.5)

.......

In establishing equations (2.5.4) and (2.5.5) the stresses

in the stirrups were treated as"uniform load qx; this may

lead to appreciable errors when the stirrup spacing is

large. To avoid this a corrective term is added to expression

(2.5.5) in the form of the force carried by one stirrup for

the cases where cý 2s

q=0.6 bd12 % qx (2.5.6)

- qx. s .......

These equations enable the shear resistance of inclined

planes to be calculated. Gorishanski also gives a formula

for checking the resistance of such planes to bending in

the form of m=Z Fizi where F. refers to any bar crossing

the inclined plane and zi its distance from the centre of

compression in the concrete, which can be calculated by a

third equation involving longitudinal forces. It is also

pointed out however that if the main steel is well anchored

and has good bond qualities, and if the loads are applied on

the top surface of the beam, it is sufficient to verify the

54
bending condition in a vertical section.

The main criticism of this method is its implied general-

ity in
expressing the shear strength always in terms of
s/
equation of transverse forces; although this may be correct

in some cases it is not a rational solution to the longitud-

inal crushing of the concrete involved in the shear-com-

pression failures of the majority of rectangular beams. An

exception to this criticism is of course the case of con-

stant moment relationship mentioned previously and in gener-

al any case in which no c value within the limits of the

dQ
shear span satisfies the condition = 0. (It must be

pointed out however, that in the case of beams without rein-

forcement this constant moment relationship, while rational

in shorter spans, is not substantiated in longer spans

(a/dl >' about 3.0) whore failure occurs at loads very close

to cracking under almost constant shear. )

shear

It is also surprising that in an empirical equation for

the of the concrete the influence of the main

resistance

was not included among the parameters affecting it.

steel

Tests show that this parameter does have an influence,

particularly in rectangular beams. Actually the influence

of the main steel is admitted by Borishanski but he was un-

able to evaluate it in his tests because all the beams that

did not fail in bending were heavily reinforced with mild

steel. One can see here one of the justifications for the

use of high yield steel as main reinforcement in the tests

reported in this thesis.

It remains however that this is a very interesting

55
 to the Introduction
approach problem of shear and the of a

systematic method for determining the variable crack length

a significant step ahead especially, as has been found.

through the tests, in the case of shear compression failures.

Also in favour of this method is the fact that the contri-

bution of the concrete is not independent of the amount of

shear reinforcement. This also agrees with the test results

of rectangular beams failing in shear compression where it

can be observed that shear reinforcement hinders the propa-

gation of shear cracks, thus leaving a larger area of con-

creto intact and capable of resisting shear.

HILLERDORG43

This method is similar to Borishanski's but takes account

of the influence of the ratio of main reinforcement on the

shear strength of the beams; on the other hand the contri-

bution of the shear reinforcement-on any given plane is

about half of that assumed by Borishanskl, as Hillerborg con-

siders as active only the shear reinforcement crossing the

critical plane in the flexural tensile half, assuming that

the bars crossed by the remaining portion of the crack are

not significantly stressed. This of course is not a very

valid assumption according to test results. However ade-

quate predictions are made by this theory and this may be

due to the skilful selection of the empirical constant in-

volved albeit a slight overestimate of- the effect of the

main steel ratio.

The derivation of the formulae in the case of vertical

56
stirrups is given briefly as it is only numerically different

from the previous method (kg and cm units are used)

Q=Qc"*Qw
100A dl
Qc = 4.2(Uw)3 (b cst brdl
c6r1
Qw = rgfYq 2r

d=0 c2
the condition gives
4.2(Uw
bd)
r

Q 100A 1
whence
8.4(U rg fyq
budt wb dSý)
r1r1

if d is than that to the longest

greater corresponding

possible inclined plano (i. oa if c>a in a point loaded

beam) then the lattor is critical. (In a point loaded boam

then c=a).

This method is not applicable as such to the analysis of

diagonal tension failures of beams failing upon shear crack-

ing, and in general the same basic criticisms apply here as

in the previous method.

2.6 Regan 's Theory94

Generally acceptable compatibility and ultimate strength

precepts are used to establish a physical theory of shear in

reinforced concrete. Account is taken of the shear resist-

ance of both uncracked concrete and shear reinforcement.

Dowel and aggregate interlock actions are not taken into

consideration.

The problem of shear is envisaged in the following

57
terms: Shear failure is one that occurs at an inclined

crack and is associated with an ultimate load influenced by

shear and generally less than that corresponding to flex-

ural failure. In analysing a beam in shear one must deter-mine

whether shear cracking would precede flexural failure

and, if so, find the residual strength of the beam contain-

ing shear cracks; if the residual strength is less than the

shear cracking resistance failure would occur suddenly when

the shear cracks formed. If it is greater an increase of

load would take place between the shear cracking and col-

lapse. Depending on the details of the beam this increase

could be sufficient to allow a flexural failure; if flex-

ural failure were not attained then a shear-compression

failure would take place involving crushing of the concrete

at the reduced compressive zone. Concerning this last point

tests have shown that the shear failure does not always in-

volve crushing of the compressive zone, particularly in T-

beams, but also in some rectangular beams with heavy com-

pression steel and light shear reinforcement, and this con-

stitutes by far the major criticism of this method.

The assumption regarding the stress-strain relationship

of concrete in flexure and the failure criterion of con-

crete subjected to compression and shear are those commonly

in use. The most significant differences in Regan's approach

as compared to others aro:

(1) The combination of the actual fibre stress and not

the average block and shear stress is used in the failure

criterion determining shear cracking and crack propagation.

58
 I

(2) The extent of the critical shear crack is governed

by the crack propagation condition and is therefore not a

fixed value of any one parameter.

(3) Two failures of concrete are considered:

(a) a tensile failure determined by (1) i. e. crack-
-

ing - of the lower part of the uncracked concrete

which does not generally constitute collapse

but a crack propagation condition.

(b) a primarily compressive failure of the whole of

the compressive zone leading to the total collapse

of a statically dotorminato member. An average

stress block factor of 0.667 and a cylinder

strength compression criterion are used in this

case.

(4) The collapse load in shear compression is determined

by the crack propagation conditions, compatibility condi-

tions and minimum strength considerations.

Shear crackincy:

Considering that cracking resistance is concentrated at

the compression zone, using the assumed local failure crit-

erion and adopting an approach similar to L. Jones51 - see

section 2.7 - regarding coincidence of actual and critical

shear stresses, the following simplified approximation to an

exact solution was developed

qer = qor no b dl (2.6.1)

.............

where qcr = average shear stress when critical condition with

respect to shear is reached in a fully developed

z
stress block = approx Ucyl in lbf/in2 units.

59
 no = neutral axis factor corresponding to fully deve-

loped stress block with the main steel elastic.

For the case of T-beams, because of the largo compress-

ive area involved, conditions become critical in the web

before the shear force predicted by equation (2.6.1) is

reached. By considering Kani type concrete "teeth" and

further amplifications, the shear cracking load of the con-

creto teeth between flexural cracks is given as

Qcr = 0.6 qcr br d1 (2.6.2)

.........

Shear failure

The above equations give the predicted shear force at

the onset of shear cracking. As the shear crack propagates,

the extension of the cracks can be expressed in terms of

the external shear force and the stress in the shear rein-

forcement, at the ultimate load stage in shear compression

failure

Qult qcr nb d1 + rqfyq bc....... (2.6.3)

`

To establish an expression for a compatibility con-

dition is formulated. It is based on the fact, supported by

the present tests, that the inclination of the compressive

force above the shear cracks causes the extreme fibre

strain of the concrete to decrease more rapidly than

assumed in flexural analysis. By considering the total de-

formation of the extreme compressed fibres and the longi-

tudinal reinforcement along the length of the shear crack -

in type cc n`9
expressed a Walther manner as = the
st 1-n
following compatibility relationship was obtained by Regan

60
 2
n
0 (2.6.4)
n .... "......
Qc
nO +m

By considering equilibrium condition with an external

moment m, the following expression is derived for the ulti-

mate moment at shear compression failure.

2
ä+ 22
G1u1t =3u b d12 n (1- rgfyq b
..
(2.6.5)
cy1

The final solution is given by the least of two sets of

values: One set of values is derived by the simultaneous

solution of equations (2.6.3p4p and 5) and the other is ob-

tained by minimum strength considerations giving the criti-

cal length c for the condition in a manner similar

-=0
T'"

and with the same maximum crack length - shear span limit..

ation as Borishanski or Hillerborg. If of course a solution

less than the shear cracking load is obtained failure will

occur at the latter upon occurence of shear cracking.

To facilitate design, charts for obtaining these shear

compression solutions are given, an example of which is

shown in fig. (1).

Based on considerations of instability of the compressive

zone Regan suggests, in the case of T-beams, a procedure

similar to that for rectangular beams described above by

considering the T-boams as rectangular beams of breadth such

rf
that the value exceeds 0.02 (minimum breadth = web
cyl
breadth).

When compreosion reinforcement is present this is taken

into account as affecting the rotational equilibrium equat-

ion but not the compatibility conditions. Even so there is

61
 U'l5

fyq Ucyl
q

0.20

° I

0.15 N%

a
a°

ý.- o
- os

0.10
CP

060.5 0.4 G3 O.2

Qd1 /M

1
70,01I.
0 3.0

0
'dr- v
00.1
0 2.0
0.03
0 0-04

0 1.0

0 rq fy q
u
cyl
0.6 05 0.4 03 02 015 0-5 0.4 03 0.2
Qd1/M Qdl /M

Fig. (1) Design charts for nD- 0-6 according to Regan

a tendency to overestimate the strength of beams with heavy

compression steel and low degree of shear reinforcement.

In general Regan's method thus enables calculations to be

made for failures occurring in a critical region at the head

of a shear crack. Stress conditions in other parts of the

beam cannot be directly analysed but this is of no practical

significance and the overall structural behaviour implied

can be viewed physically as being similar to that of a truss

with a sloping compression chord. The inclination of this

chord and that of the concrete compression members of the

lattice are unknown except at the critical region where c

of
and n can be calculated. For the condition application

of this method however these are never critical.

Compared to the other shear compression methods Regan's

theory reflects much better the actual behaviour and com-

portment of a beam cracked in shear; and provides a rational

link between shear cracking and propagation and ultimate

failure. Its predictions are substantiated by about 1000

test results. The basic criticism of this method lies in

the fact that it predicts compression failures at all times

and thus does not describe accurately the physical behaviour

of beams failing in other modes - T-beams, beams with light

stirrups and heavy compression steel -. The overestimates

in the case of the latter, mentioned previously, can be

to this general criticism. It remains however

attributed

that in the case of shear compression failures of rectang-

ular beams - i. e. the majority - this method has been found

to be the most rational. There exist of course some criti-

63
cisms involved even in this case; They are not fundamental

however and are doscribed in Appendix A where the necessary

modifications, resulting in their elimination in any prac-

tical cases are also proposed.

2.7 Other Shoar Theories

In addition to the shear theories described there are

some other methods that present a particular interest in

that they introduce some different concepts or bring out

more correctly the influence of a given parameter. They are

of limited or of no practical application as presented but

represent some other possible approaches to the problem of

shear. Their basic points are discussed briefly in what

follows:

R. TAYLOR116

This theory is intended to form the basis from which an

design method for proportioning stirrups in a

empirical

beam can be developed, as presented it can only be used in

design and it is difficult to verify rationally the empiri-

cal coefficients suggested since beams designed according

to this method should always fail in bending.

Basically Taylor considers that the contribution of the

concrete is small compared to that of the stirrups and

furthermore that stirrups beyond the lower end of the cri-

tical crack should be taken into Consideration since they

are brought into action by dowel action. It could be said

64
 that Taylor considers the excess strength of the beams, over

that predicted predicted by the Morsch truss analogy as be-

ing due to the dowel action rather than the contribution of

the concrete compressive zone., The present series of tests

show however - through strain measurements - that the force

in the concrete does slope downwards towards the support;

they also show that the number of stirrups stressed by dowel

action beyond the end of the crack is considerably smaller

than that given by application of the empirical formulae

suggested by Taylor, in that the excess strength above

a/
norsch's prediction is primarily of function of the com-

pression zone characteristics (area and strength).

PETERSSON87

Petersson has proposed an empirical equation for the

'" diagonal tension failure of reinforced concreto beams which

is simple and seems to take better account of the influence

of the ratio of main steel than the ACI/ASCE Code.

Tho proposed equation applies to beams having a/d1 1 3.5

and is

qcr loop
1,3 (u ýf St
bd1 w brd1

when kgs and cms are used.

Although the cube root expression of the concrete term

seems justified by test results there is some danger of

overestimation involved in the use of only the square root

in the case of the steel 'term, especially if the ratio of

main reinforcement is important. Although this may not be

65
likely to occur in rectangular beams, heavy main reinforce-

ment may be present in T-beams. In the region of application

of this method, tosts suggest not only a reduction in the

amount of influence of the main steel suggested by this

method but also the necessity of an upper limit especially

if such an approach is to be used for the shear cracking of

beams having a tension flange.

It remains however that this method contains the main

parameters involved in the shear cracking of longer beams

although the limit of application could be lowered to

d
1
about 3.0 without serious underestimates of strength.

L. JONES51

Jones' theory applies only to rectangular beams without

shear reinforcement. Its scope is thus very limited but it

is based on actual stresses and strength and not on a se-

lection of empirical constants; the writer does not agree

with such an attempt in view of the importance of the neg-

lotted dowel and aggregate interlock actions but the inter-

est of this method lies in the theoretical approach adopted,

which could servo as a basis for a theory when more is

known about the two actions just mentioned.

This theory is based on the following assumptions:

the strain diagram is linear, the steel strain being

-

some constant fraction of the concrete strain at the

same level

- the concrete stress strain curve is parabolic

- Mohr's failure hypothesis for concrete, with a para-

66
 - bolic failure envelope to the stress circles, is the

failure criterion.

Jones also considers that the use in the hypothesis of a

tensile stress obtained from flexural tests represents

better the condition in the compression zones of beams sub-

ject to bending and shear.

From his assumption of a parabolic stress block in the

concrete, Jones derives the shear stress distribution ana-

lytically from the horizontal equilibrium of elements a

distance dx apart and the consideration of a variable lever

arm over, the length dx. The actual shear stress distribut-

ions are plotted against critical shear stress distributions

derived from the longitudinal stresses and the failure

criterion and a cracking condition is reached when the two

curves coincide at some point.

The theory was initially applied to the case where the

steel had yielded at the failure of the member. In

main

for it to apply to the case where the steel is elastic

order

and the beam cracked at failure Jones had to assume that

the bond between steel and concrete at the cracks had

broken down.

This method is of course impractical because of the

numerous stress distributions that must be plotted. Tth this

-

and the criticism concerning the. dowel and interlock action

two other drawbacks intrinsic to this method shculd. be

pointed out.

1) It is only applicable to rectangular beams since in

T- and I-beams the area of compression and initial crack-

67
ing are vastly different.

2) Since the coincidence of the critical and actual shear

stresses occurs in the central part of the compressive zone

Jones assumes that failure follows immediately upon the

attainment of cracking conditions. This of course is not

substantiated by the tests of short beams.

KANI56

The latest of Kani's ideas on shear is contained in an

article entitled "A rational theory for the function of web

reinforcement". Actually no theory as such is proposed on

which calculations can be based but the author attempts to

bring out what according to him is the basic function of the

web reinforcement and goes no further than more generali-

ties.

Beginning with a study of the stress trajectories, the

principle of multiple internal arches is developed and this

loads to the conclusion by Kani that the purpose of web re-

inforcemont is to produce supports for the internal arches

and that no direct relationship exists between the shear

force and the requirement for web reinforcement.

Basically this multiple arch suggestion involves arch

action at every stirrup. This is not borne out by the tests

which show one main arch action above the shear crack, while

the arch actions that could possibly butt against the indi-

vidual stirrups are considerably loss important, In Kani's

graphical representation of the multiple arches, the hori-

zontal projection of the critical shear crack is shown as

68
 d
2
about which is very far from tho truth in the case of

light shear reinforcement and distorts the actual image of

a cracked beam.

Kani also suggests that no stirrups are required near

the point of load application and the support region. The

former is of course also implied in most theories in the

sense that the shear crack never really reaches the load

plate until failure, so prior to failure the stirrups in-

volved are at some distance from that point. Similarly in

the case of the region near the supports a reduction could

be made provided enough stirrups are left beyond the ex-

pected critical shear crack to take up the dowel action;

this latter point is not considered by Kani. These reduct-

ions in shear reinforcement are thus more of a design quest-

106'
ion and a very recent CEB proposal not yet agreed toi

attempts such an approach.

To substantiate his ideas Kani carried out a number of

tests with shear reinforcement present only in a portion

of the beam. They cannot be considered a success since when

flexure failure did not occur reductions of up to 45% in

the shear strength were observed. This may be one of the

reasons why no concrete proposals were made.

2.8 Review of some Research on T-Beams

As has been seen in the critical appraisal of existing

theories, none of the codes and only two methods, consider

the shear strength of a T-beam as being anything else but

69
that of the rectanglo formed by tho web and its projection

to the top surface. Unfortunately these two methods - Laupa

and Regan - predict shear-compression failure, implying a

beam behaviour not substantiated by tests. Ramakrishnan91

also noticed that the failure of T-beams was not accurately

described by shear-compression conditions - failure due to

shearing of the compression zone took place on all T-beams

with shear reinforcement while equivalent rectangular beams

failed by crushing of the concrete in. the compressive zone -

but did not produce any theory to explain it and the numer-

ous empirical formulae of limited individual use that he

produced, reflecting his tests on small scale members, can-

not be thought of as a method of calculation. - Qualitatively

however some of Ramakrishnan's other observations agree

with the present series of tests. These are that in longer

spans, rectangular and T-beams without shear reinforcement

fail in about the same load but that in all other cases the

strength of T-beams exceeded that of rectangular beams;

this increasein tho shear resistance of beams due to the

addition of a flange was not uniform at all shear spans and

seemed to depend on the magnitude of the shear span and the

mode of failure.,

That the failure of T-beams is not adequately represented

by shear-compression consideration was also pointed out by

Adepegha and Swam

y6 who observed that in T-boams the com-

prossivo strains at failure are never critical; although

they tested T-boams without shear reinforcement and this

finding doos not contradict shear-compression theories in

70
the case of beams with long spans 'failing upon shear crack-

ing, the validity of shear-compression is questionable in

those of the beams that, having low valuos, withstood

d
1
increased loads between shear cracking and ultimate failure.

The very few comparative tests performed have shown that

gonerally the shear strength of T-beams exceeds that of

rectangular beams although no rational attempt has been

made to justify this. (Richert's97 tests show increases of

the order of 17% in short T-beams (a/d1=1.71) with and

without shear reinforcement. In the case of T-beams without

shear reinforcement Petersson's87 and Adepegba's5 tests

show that in the diagonal tensile failures at largo spans

no significant difference exists, while Adepegba's tests on

spans (= 3.0) show an increase in the case of T-

shorter

beams. ) The number of tests however, especially those with

shear reinforcement is very limited for any conclusion to

be drawn and a rather large number of T-beams with stirrups

wore tested as part of the present investigation. Use was

made however of the available test results of longer T-beams

reinforcement (a/dl' 3.0) Ferguson and

without shear -

ThompsoR7, Adepegba, Al Alusi7 «ý in the establishment of a

diagonal tension failure formula, and the proposed forwula

for T-boams with shear reinforcement has been compared with

all the available test results of normal-size T-beams

(Chapter 8). Furthermore, when examining the effect of any

given parameter affecting the shear strength of T-beams, re-

ference is made, whenever possible, to the available test

results including the well known Stuttgart tests. Some para-

71
meters were not investigated because tests had shown that

they had no effect; a case in point is the question of com-

pression rbinforcement in T-beams. Since no effect whatso-

ever was observed in either Ramakrishnan's or U'ilby's122

tests this parameter was not varied in T-beams.

A review of existing tests was also interesting from the

point of view of the beam's behaviour in respect to crack

opening and inclination. Evans, and ßorishanski's observ-

ations are discussed in Chapter 7 in connection with the

shearing mode of failure.

2.9 Ddwel and Ag regate Interlock Actions

The various shear theories described although they differ

in the concept of the mechanism through which shear is

carried in a reinforced concrete beam (truss, tied arch,

arches etc. ) generally agree on the assumption

multiple

that the only contributors to shear resistance are the con-

crete and the shear reinforcement. As a result the possible

contribution of the shear forces carried by dowel action of

the longitudinal steel at the base of the crack and by

interlock of the aggregate has been ignored. Even if to

the writer's opinion their importance is of a primary im-

portance only in beams without shear reinforcement failing

in diagonal tension and in short beams their existence

should be acknowledged and their effect should be ration-

ally explained to the extent that present knowledge of the

problem makes this possible. In establishing the various

72
equations of Chapter 7 these two actions are considered and

an attempt is made to explain their effect on the beam's

behaviour. In the present section the concept of the two

actions is introduced and some relevant experimental invest-

igations reviewed.

2.9.1 Dowel Action in Reinforced Concrete Beams

(a) Beams without shear reinforcement

Dowel action is the name given to the transfer of force

to a block of concrete by a reinforcing bar embedded in the

block and projecting from it (fig. 2). In the past the main

significance of dowels has been in highway work where they

are used to connect the concrete on the two sides of con-

struction joints. More recently attention has been paid to

the existence of dowel actions in beams and to their con-

tribution to the transfer of shear forces across cracks.

Most of the work on dowels in beams has been concerned with

beams without stirrups in which the tearing action of the

dowel is resisted by tension in the concrete fig. (3).

-

The failure of this type of dowel action takes the form of

a horizontal tearing of the tensile zone at the level of the

main reinforcement i. e. dowel cracking.

In many beams without stirrups shear failure occurs with

a more or loss simultaneous growth of an inclined crack up-

ward through the compressive zone and backward along the

main steel. The question of whether the dowel action carries-

a significant part of the shear force prior to failure and

whether the dowel cracking is the cause or effect of the

73
 Fig. (2) Dowel Action

tension

cornpression
ot

rýý
1 SECTION A -A.

Fig. (3) Dowel Action in a Beam

16
r-
x5
A

4-

3- shzcr force carried

rosette 4a by the do"ue)
L.
v
u2_
1M-
-ýi ý-3 14 °/000 °C Q
72"
1 °o
sh-ear force carried
°o by the concret:
yý °
Fig. (4) & (4a) Dowel test Beams 012345
H(refeld and Thurston V shear force carried by becm lbx 10

Fig. (5) Test results of beam with

rosettes Krefeld and Thurston.
beam's failure have not been yet fully answered. To do so

would require knowledge of the relative stiffness of all

possible shear carrying components (dowel, uncrackod con-

crate zon©, aggregate locking across cracks) and their

strengths.

Some work has boon done on the strength of dowels formed

by longitudinal reinforcement. The experimental side in-

volves tests of a number of types of specimen in which the

shear in the concrete compressive zone is eliminated or

measured and that supported by interlock is removed by the

use of smooth sided preformed cracks.

Krefeld and Thurston59760 carried out tests on dowel

specimens in which part of the tension zone of the beam was

cast separately from the rest of the beam and was connected

to it only by the tensile reinforcement (fig. 4). Thy also

tested beams of the same dimension as the above but without

any prefabricated cracks. From an interpretation of the

J/

corresponding beam of the two series it appoars that at load

near failure (diagonal tension failure at shear

stages

cracking load or for slight increase of load) the dowel

carried one third of the total shear.

Krefeld and Thurston also carried out one test of a diffo-

rent typo of boom (fig. 4a); the latter had a preformed in-

elinod crack on one side only, whose width was 3" at the

main steal level. Delta rosettes were placed on the 2.75"

intact concrete zone at the head of the crack. The results,

shown in fig. (5) indicate that up to 60% of the total shear

was carried by dowel.

75
 Some other investigators approached the problem from a

different angle and making various assumptions tried to com-

pute the dowel force in normal beams. Watstein and mathey120

computed the dowel force from equilibrium conditions, the

measured stresses in the tensile steel near to the bottom

assumption/
of the diagonal crack and the that the compression

force acted at mid-height in the concrete compressive zone.

they concluded that the dowel force is a function of the

ratio of main steel and at loads representing 42 to 46% of

the ultimate load 3B to 74% of the shear force was carried

by dowel; furthermore they noticed that the maximum idow©l" per-

centage cotrespondod to the load at which the inclined crack ex-

tended to the point of load application (a/d1=2.2, so the

beam did not fail upon shear cracking). Following an in-

crease of load beyond that stage the dowel effect decreased

rapidly. Acharya and Kemp2 also calculated the stress con-

dition at the top of the shear crack assuming that the

dowel force was zero and concluded that only 40% of the

shear force must be carried by the concrete compression

block otherwise the stresses would cause premature failure

of the compression zone. Unfortunately in both the last two

cases mentioned the results contrary to the authors' con-

clusion, must represent the contribution of both dowel and

interlock action; and this can explain the reduction noticed

by Watstein when at increased load the crack widened.

Theoretical research has offered a number of "beams on

elastic foundation" approaches in which the dowel element

is viewed as a beam and the concrete supporting it is

Nt P
76 --ýp 77 uat
viewed as the elastic foundation. Two such solutions have

been proposed. by Jones52 and Regan94. Both take the dowel

element to consist of the reinforcement and its concrete

cover. They differ in that Jones assumes the foundation

modulus (support force / dowel displacement) to be inversely

proportional to the depth of the beam while Regan assumes

that local supporting stresses at the level of the steel

are dissipated on the way up into the compressive zone,

with the result that the effective depth is much loss im-

portant. Unfortunately the experimental evidence supporting

these theories is very meagre.

A very recont attempt at the question of both dowel and

(113)( 114)ý
interlock action has been the ono by H. P. J. Taylor

The theoretical basis for calculation is not yet developed

but tests of Krefeld and Thurston type specimens (fig. 4)

and normal beams have shown the dowel force to be about 20%

of the total applied force.

In one test of the present series (beam 01) the dowel

forces wore artificially reduced to negligible levels in a

beam whose normal failure would be by diagonal tension: A

25% reduction in strength resulted (fig 6).

To judge by comparing the strength of dowels in special

dowel tests with those of ordinary beams of similar geo-

metry and reinforcement, the part of the shear force sup-

portod by the dowel in a normal beam without stirrups

whose ordinary mode of failure would be by diagonal tension

at shear cracking - seems to be of the order of 25%. In

short beams, where the critical shear crack reaches the main

77
 U° = 3500
R3
rst = 1.46°%o

" a/d1 _ 3.36

I Quit _7570 lbs

Di
lt Quit 10100 ibs II

dowel action negligible

Fig. (6) Tests Indicating the Importance of Dowel Action

Fig. '(7) Dowel Action in Short Beams

EC

Fig. (8) Dowel Action in Beams with Stirrups

 steel very close to a support applied to the underside of

the beam, the dowel forces may be much greater since the

dowel element is supported not just by tension in the con-

crete above it but also by compression from the support

(fig. 7).

jj Beams with Shear Reinforcement

Near the ultimate load stage of a beam with stirrups, it

is the stirrups that resist the tearing along the main steel.

A part of such a beam containing a shear crack is shown in

fig. (©). Towards the ultimate load stage the principal

shear crack is so wide that interlocking of aggregates a-

cross it may be neglected. Shear can thus be transforred

across the line ABC only by the following

1) The concrete above the crack

2) The shear reinforcement crossing the shear crack

3) The dowel element below the shear crack.

An alternative view of the same system is afforded by

considering the line (A©DE). In this case the main elements

supporting shear are

1) The concrete above the crack

2) The shear reinforcement crossing the crack along AB

3) The shear reinforcement crossing the dowel crack 8D

4) The remaining dowel at DE

In solving the complex compatibility and equilibrium

conditions governing the above systems, most shear theories

neglect dowel forces. The one theory that does not, -

116 in
Taylor's _ has already been described and discussed

79
section 2.79 the main criticism being its n16rsch type app-

roach and the overestimate of the number of stirrups

stressed by dowel action.

52
R. Jones tests on beams with
. carried out preformed

cracks and inclined stirrups and found that the shear force

carried as a dowel was about 25-30% of the total shear force.

This seems an overestimate and is probably due to the ab-

normally high percentage of main steel used.

Dowel tests on Krefeld and Thurston type (fig. 4) beams

with stirrups have been carried out by Baumann12 in 1968.

The results are presented in the form of load deflection

curves where the load represents the dowel force and the

deflection the displacement across the preformed crack. At

the cracking stage the deflection generally increases

rapidly for no increase in load. The ability of the struct-

ures to sustain increases in load hereafter is dependent on

the presence of stirrups capable of supporting the necessary

vertical tension. The ability of a stirrup to do this is

increased if it is close to the preformed crack and (to a

lesser extent) if the dowel element is stiff. The second

stage of the load deflection characteristic, generally

following after a considerable increase of deflection with

the dowel force equal to the dowel force causing shear

cracking, was found to conform to a parabolic curve, func-

tion of the distance from the preformed crack to the first

stirrup. The addition of further stirrups was found to have.

no appreciable effect (see fig. 9).

The general overall conclusion that can be drawn from

80
 4

3
Dowel
load
V oo- Curve ern. approx
in tons

2
where KL
v
and f_ distance from preformed
crack to first stirrup
Jv _ moment of inertia of dowel
See, enlargement
(bar+ concrete cover)
below.

12345
A (vertical deflection across crack ) in mm

Overall" characteristic for a beam with stirrups

Dowel
- load
V
in tons

0.05 0.1
A in mm

Fig. (9) Load/ deflection curves according to Baumann

3!
Baumann's tests is that very large vertical displacements,

- larger than those occurring in normal beams - are required

in order for a dowel force of any appreciable magnitude to

be developed, with the exception of the force that can be

transferred to one stirrup very close to the start of the

crack.

The above indication concerning the dowel action in

beams with stirrups seems to be confirmed by similar dowel

tests by Arroyo9 as well as by the stresses in the stirrups

beyond the lower end of the critical shear cracks as mea-

sured in the tests of normal beams reported in this thesis.

2.9,2 Aggregate Interlock Action in Reinforced Concrete

Beams

Measurements reflecting the behaviour of cracks in teats

of beams subjected to bending and shear indicate that not

only do the cracks open progressively as the moment is in-

but that displacements parallel to the line of the

creased

cracks also occur. Depending on the roughness of the surface

of the crack, the vertical displacements are such that parts

of the opposite sides of the crack may touch and carry force

across the crack. Fig. id gives a diagrammatic representa-

tion of such a crack which would induce compressive and

shear forces.

The existence of aggregate interlock action in beams sub-

jected to shear was first demonstrated by Paulay and Fen-

wick86 through tests of a beam without shear reinforcement

provided with crack initiators and by direct aggregate

82
 N

DV

AN

Fig. (10) - Derivation of aggregate intcriock forces

I
ýý
interlock-tests on small concrete block specimens. It is of

interest to note that in the tests of the concrete specimens

a diagonal, crack developed when the shear force across the

initial crack was sufficiently increased. The authors point

out that similar diagonal tension cracks to those formed in

these test specimens may be seen in tests of normal beams

and are referred to by them as aggregate interlock cracks.

The results of these tests can only be termed tentative

but they indicate that prior to diagonal tension failures

in normal beams 50 to 60% of the shear force may be carried

by interlock actions. The authors consider that aggregate

interlocking depends on the following parameters, although

no attempt was made to investigate them:

(a) The shear force transmitted by one aggregate particle

is likely to be proportional to the contact area

between the particle projecting across the crack and

the concrete on the far side of the crack

(b) The contact area is reduced when the width of a

crack increases

(c) The number of significant particles projecting across

a crack must depend on the grading of aggregates

(d) The deformation characteristics of the cement mortar

matrix will be influenced by the strength of the

concrete and probably by the water and air content

of the mix
(e) Different responses must be expected with respect to

load application to individual particles, depending

upon the direction of the applied load and the posi-

84
 tion of the particle in the freshly placed concrete.

Of particular importance is tho water and air gain

under larger aggregates.

The only other attempt at investigating aggregate inter-

lock is currently under way at C. and C. A. Laboratories

under the direction of H. P. J. Taylor114. A different type

of specimen is used for the direct interlock test but the

results are only preliminary and have not yet been inte-

grated into the behaviour of actual beams. Taylor however

has obtained an indication of the magnitude of interlock

action in normal beams without shear reinforcement (a/dl =

3.7)(113) in an indirect manner. Fig. (11) shows the shear

forces carried in the compression zone and by the dowel

actions plotted above each other. The rest of the shear

force according to Taylor must be carried by aggregate

interlock. The dowel load was obtained by comparing the

displacements measured at the foot of the crack to that of

the dowel load displacement curve obtained in a Krefeld and

Thurston (fig. 3) type small scale beam having similar

characteristics. The shear taken up by the compressive zone

was derived from a great number of strain readings in that

rogion.

It can be seen from fig. (11) that the magnitude of the

aggregate interlock action can reach up to 50% of the applied

shoar; this value is not very different from the one arrived

at in the case of Paulay and Fenwick's tests.

Although the question of interlock action seems to be

quite important in the formulation of any rational theory

85
 hear force
10000 by beam

d Force carried
U
c.
0 iprcrsion zone
w
)wel
rn
Y
5000 force
U
v by
4° :5sion zone

v
V

5000 10000

Shear force on beam Kg force

Fig. (11) Shear forces in beam carried by "dowel and compression zone
according to Taylor

t
T-A,
for the calculation of beams without shear reinforcement

failing in diagonal tension the research done so far in

this domain does not really allow this effect to be taken

into account other than empirically. No research whatsoever

has been done in connection with beams provided with shear

reinforcement; it is the writer's opinion that in this case

the large crack openings of the failure cracks must reduce

considerably the interlock effect at load stages close to

failure.

87
 p

CHAPTER 3

DETAILS OF THE EXPERIMENTAL INVESTIGATION

This chapter outlines the general characteristics of the

75 beams tested, which formed the axperimental part of this

investigation. These tests were performed in order to obtain

both qualithtivo and quantitative information on the behaviour

of various types of beams which have a practical importance

and normal size sections were thus used in order to avoid as

much as possible the scale effects intrinsic to small beams.

The basic purpose of these tests was to investigate the

effect of given parameters on the shear strength of a beam.

To accomplish this it was necessary not only to vary one

parameter at a time but also to obtain shear failures

throughout the range of this variation. The latter was

accomplished mainly through the use for longitudinal rein-

forcement of a Swedish steel having a yield point in excess

of 90000 psi"

The results of these tests are presented and discussed

in terms of the influence of the parameter examined and the

general behaviour of the beams in Chapter 4,5,6 relevant

to rectangular, T and thin-webbed I-beams respectively. The

present chapter describes the general common characteristics

of these tests, since, to provide an opportunity of making

comparisons between tests in different series, a number of

parameters were kept constant. More presicely this chapter

deals specifically with the following aspects of the ex-

88
perimental investigation

1) Concrete used

2) Reinforcement used

3) Geometry of test beams and arrangement of reinforce-

ment

4) Casting and Curing

5) Tost procedure

6) Parameters studied

3.1 Concrete

It was originally intended to use concrete mixes of var-

ious cube strengths of 3000,6000,9000 lbf/in2 at 20 days

to represent low, normal and high strength concretes res-

pectively. With these in mind preliminary tests were carried

out on the approximate design of these concrete mixes.

Though the results were positive, precautions were subseq-

uently taken to give less vibration to the 3000 lbf/in2 con-

crete to avoid segregation, while greater vibration became

necessary for the 9000 psi concrete mix. The 6000 psi mix

was quite satisfactory and it was chosen as the basic one

for the majority of tests.

When changing the size of the main aggregate for beam D2

care was taken to modify the design of 6000 psi concrete mix

but maintaining the same strength. Shown in table 1 are the

proportions of the 6000 psi mixes with varying size of main

aggregate.

89
TABLE 3.1.1 Mix Proportions 8y Weigh 6000 lbf`in2
ýUýý=.
at 28 s
_d

Max. size of aggregate used

Particulars
3/4" 3/0"

Ordinary Portland Cement 1.00 1.00

Fine sand (100-25) 0.61 0.66

Coarse sand (25-3/16) 1.39 1.53

Fine Aggregate (3/16-3/8) 1.14 2.34

Coarse Aggregate (3/0-3/4) 2.21 -

Water 0.61 0.62

The following 3/4" aggregate mixos were found satisfactory

and their proportion - by weight - are given below in table 2

TABLE 3.1.2. G1ix Proportion ýBl Woightý_ Aggregat©

Mix Mix Mix

Particulars
3000 lbf/in2 6000 lbf/in2 9000
2
lbf/in

Ordinary Portland Cement 1.00 1.00 1.00

Fine Sand (100-25) 0.95 0,61 0.37

Coarse Sand (25-3/16) 2.20 1.39 0.88

Fino Aggregate (3/16-3/8) 1.90 1.14 0.74

Coarse Aggregate (3/8-3/4) 2.75 2.21 1.12

Water 0.95 0.61 0.42

The siove numbers given aftor the sands and aggregates re-

for to the sizes given in British Standard Specification.

Both the sands and aggregates were Thames river deposits and

wore batched dry.

S/
6"x6" cubes and 6"x9" split cylinder specimen werd tested

on the same day as the beam and so wer© 6"x12" cylinders in

90
all but the first six rectangular beams. In view of the sub-

sequent relation between cube and cylinder strength for the

normal concrete involved in these six first beams and the

57972'74
findings of other investigators it was decided to

consider the cylinder strength relevant to these beams as

being 80% of their cube strength; table 3 lists the cube,

cylinder and split cylinder strengths of the concrete in

each of the beans tested.

TABLE 3.1.3 Values of U Ucyls_ U of concrete in the

brams tosted.. (in, lbf in2

BeamI Uw Ucy1 Usplit ©oam Uw [u1 Usplit

R1 4730 3800 372 Ti 155001,4050 422

R2 4710 3800 349 T2 515714067 398

R3 4500 3600 375 T3 5300 3990 430

R4 6190 4950 467 T4 6290 4710 458

R5 6190 4950 467 T5 5685 4890 387

R6 6290 5000 451 T6 5090 3740 344

1
R7 5320 4070 395 17 5130 3970 450

R8 4870 3870 410* T8 58301 45301 436

R9 5380 42901 403 T9 39101 2930 328

R10 5520 4295 387 T10 5475 4090 403

R11 4765 3800 364 T11 6120 5360 452

R12 6510 4920 510 T12 5830 4450 392

R13 6160 4680 430 T13 12470' 18501 226

R14 5590 4210 380 T14 5975 4840 436

R15 5610 4330 470 115 6071 4810 444

1
R16 6040 4580 454 IT16 6426 4740 420

R17 2440 1850 225 T17 6471 4790 414

91
 Beam! Uw Uay1 Usplit Beam Uw Ucyl Usplit

R16 5780 4540 457 T10 5600 4120 460

R19 5755 4390 420 T19 5558 4340 428

R20 8150 6230 558 T20 5015 4655 442

R21 6980 566 T21 5935 437

118810 14700
R22 15120 4280 414 T22 6424 4980 410

R23 5640 4370 460 T23 5785 4460 450

R24 5400 4480 463 T24 6138 5140 4C5

R25 5880 4470 426 T25 9683 7841 602

R26 6370 5300 441 T26 9520 0260 543

R27 2550 1980 223 T27 2487 1742 225

R28 5450 4580 396 T23 5880 4617

0-1 4135 3300 338 T29 6266 5349 424

D-2 5405 4400 423 T30 5226 4160 410

T31 5256 4494 396

T32 5155 40001 425

T33 6800 5340 494

T34 5757 4920 422

T35 5754 4080 420

{
T36 4125 3500 368

T37 5500 4615 415

T38 5630 4380 415

92
 Beam Uw Ucyl Usplit

W1 5407 4332 436

W2 2438 1827 235

W3 8220 6605 533

{
W4 6690 4985 527

W5 6047 4970 470

W6 5948 4920 460

W7 5934 4980 486

,

Although it was measured in every case the split cylinder

strength was used only as an indication of the actual tensile

and in that respect it was useful in the understanding of

the diagonal tension failures. As it can be seon, it is

rather erratic but, more important, it is based on plastic

conditions; the assumption of the theory of elasticity not

being applicable to concrete up to failure the tensile

strength given by the split cylinder test does not corre-

spond to the real strength'of concrete in tension. It is

even difficult to associate this strength to the uniaxial

tensile strength of concrete as tests in which both methods

were used give varying results:

Tests of R. Baus13 Lie - Tensile strop týh of Concrete

IUdiCt1
Ago Ui it
it
U=1,
direct
kgf/crn2 kgf/cm2

3 months 32.6 35.4 1.085

E ýý ! 33.0 45.0 1.36

5 years 35.5 41.0 1.15

14 days 13.4 17.9 1.33

18.6 23.4 1.26

93
 Age Udirect Osplit OMlit
Udirect
kgf/cm2 kgf/cm2

28 days 25.9 29.4 1.13

12.9 15.6 1.21

20.0 29.5 1.41

32.7 40.0 1.22

21.0 120.3 1.34

Tests of "Laboratoire EDF"34 Aix an Provence, France

Each figure is the average of 6 specimens; the coefficient

of variation is also indicated

i

Age Odirect u iss

split
u -1i
kgf/cr2 kgf/cm2 direct

7 days 32.4 130.5 0.94

4.6% 9%

28 days 38.4 36.4 0.95

2.9% 13.6%

90 days 38.7 41.6 1.07 !

7.1% 2.9%
?
210 days 43.9 37 0.845
tr
2.73% 11.45% i
r--
Because of the manifest inconsistency of the split cylind-

er test it was decided to base the tensile strength of the

concrete on the relationship between the uniaxial tensile

strength and the cylinder strength of concrete on the basis

of relevant tests as indicated in chapter 7.

Because of the great number of beams involved and the

intrinsic storage problem, the majority of the beams togeth3r

with the relevant spcimens were tested between 14-17 days.

94
 bUoc
.

400C

3000

N
G

N
J)
G

4,
N
2000

1000

20 0 400 600 800 1000 1200 1400 1600

. O/in
r Strain in in x i0

Fig. (12) Stress strain curves for normal strength concrete

cý v:
000

POO
º

P00

boo

boo

500 1000 1500

sthain in in x 0-61 in

000

1.
$00

'co

'1500 WOO
500 1000 .
in in x 1061 in 0
strain . 5 Ei
 4000

3000

N
C

4-

E
a

L
4J 2000
(1)

1000

1000 2000 3000 4000 5000

5000 6000 7000
in in x 1Ö 6/
Strain in

Fig. 0 5) Typical stress strain curve for the concrete

J
Fig. (12) shows characteristic stress-strain curves obtained

for the normal strength concrete, while fig. (13) & (14)

show the stress-strain curves obtained when low strength and

high strength concretes ware used. On one occasion the "fall-

ing branch" the

part of curve wos also measured and this

is shown in fig. (15).

3.2 Reinforcement

The, various types of reinforcement, their qualities and

their function in the tests are decribed in the following

paragraphs. Their stress strain characteristics were deter-

mined by tension tests of samples on which strain gauges

were placed, in an Amster machine equipped with a graph re-

corder.

(a) High yield deformed steel L Ks6)

This is a Swedish deformed steel with circular ribs at

right angles to the direction of the bar. The spacing of

the ribs is equal to 0.5 times the bar diameter. The bond

qualities of this stool seem to be adequate as no problem

in that respect was encountered, but it must be added that

a ono foot extension beyond the support, containing heavy

transverse reinforcement in the form of loops, was pro-

vided in every beam. '

Those Swedish "naturally hard" steals yield sharply at

the yield point stress and there is usuQiiy a definite

yield plateau. Following such yielding an ultimate strength

of about 50 per cent greater than the yield point in deve-

98
loped. It is a hot-rolled material in which high strength

properties are imparted by small amounts of alloy compo-

nents (C, Mn, Si)47

Four sizes of that steel were used: 16mm, 12mm, 10mm and

Smm.

The 16mm size of RS60 steel was the only type of longi-

tudinal reinforcement used in the tests. The high strength

steel was used in order to obtain shear failures using only

a moderate percentage of longitudinal reinforcement and to

enable shear failures to take place at longer spans, parti-

cularly in the case of T-beams. In many shear tests per-

formed before, the percentage of tensile reinforcement has

been unrealistically high, presumably in order to avoid

flexural failures; As it happens that shear strength is de-

pendent on the ratio rather than the strength of the main

steel, the present arrangement was, as expected, conducive

to shear failures. In general the use of this steel enabled

the present investigation to be based on actual shear fail-

ures of beams having realistic ratios of main reinforcement

and was instrumental in increasing the range of a/dl values

at which shear failures were obtained.

The other three sizes of Ks60 were used as shear reinforce-

ment in the web crushing tests.

To obtain the stress strain and yield point character-

istics of these steels six specimen of 16mm and four of the

other sizes were subjected to normal tensile tests carried

out in an Amslor testing machine of 50 ton capacity. Typical

for the in fig. (16) to (19).

curves reinforcement are given

99
 N

100000

Yield stress

90000

80000

N
S 70000

60000
U
L
N
N

50000

p. s. i.
40000

30000

20000

10000

1000 2000 3000 4000 5000

Strain in in x 106/in

Fig. (16) Stress strain curve 16 mm dia. I(S60

100
 100000

Yield stress

90000

80000

70000

60000
N
C_

50000
E
N

40000
0 5
p. s. i.

30000

20000

10000

1000 2000 3000 4000 5000

Strain in in x 706 in.

Fig. (17) Stress strain curve 12 mm dia. KS60

101
 110000

Yield stress
100000

90000

80000

.,

70000
c

60000
U)
'! )
d
L
'I)

50000

c 106 p. S. i.
4 0000

30000

20000

10000

1000 2000 3000 4000 5000

in in x 10 6/in
Strain

Fig. (13) Stress strain curve 10 mm dia. KS 6O

to?
 N

1000001- Yield stress

90000

80000

70000

60000

t' 50000
c!)
lbs / in2

40000

30000

20000

i
10000

1000 2000 3000 4000 5000

6/
Strain in in x 10 in

Fig. (19) Stress strain curve 8mm 'dia. KS 60

. tai
 Yield stress

40000 Yield stress 4COO0=

30000 - 3=
:v

ý rJ c E:=30x106 lbs/in2 20000 E= 29.4x 106 lb/in

" G_

N N
U
ý- a

stress strain curve stress strain curve

"
10000 *j 2 dia. mild steel 10000 I/4 dia. mild steel
3

1233 133 2
%
Strain in in x 10 /in Strain in in x 1Ö in

Yield stress

40000 Yield stress 40000

30000 30000
CV

E_ 29.6 x 106 Ibs/in2 20000 G E= 30x10' lbs/in2

0000 N
N

W 4j

10000 stress strain curve 10ý Stress strain curve

3/g dia. mild steel /4 dia. mild steel
.1

1231 2 3

Fig. (20) Stress strain curves for mild steel

I C=
 0
0
U)
r

0
8

o
o

SE
qo
oý
x
O.
O c.
C
hd

o
J0

OU
.60

N
N
N
()
b
4,

(r

S
ON
i/.^ ''"

N
th
lL

öö0000
00op
t0 NM-N

zur/sqi ul ssaJ; s
I
 The strength data given below relate to the averages of the

strengths obtained in each group of tests. No marked variat-

ion existed in any given group of specimens.

Size of Ks60 bar$Cross sectional Yield stress Ultimate

Area in in2 in lbf/in2 stress

in lbf/in2

16mm 0.312 90 000 140 900

12mm 0.176 92 000 i 142 000

10mm 1146 600

0,122 102 000

8mm 0.078 100 300 120 000

In figures (16) to (19) only the stress strain curve up to

the yield point is shown. The complete stress strain curve

has the following general form in all cases. E is approx-

imatoly 29.5 x 106 lbf/int.

St

failure

90

30 u-3.5 strain in inxlu din

106
(b )_Plain round mild steel

This was used for a variety of purposes but mainly as

shear reinforcement of the rectangular and T-beams. The

specific usage of each bar size was as follows

7/32 - Shear reinforcement

I/4 - 11 11 and transverse flange reinforcement

3/8 - 11 it Of compression reinforcement

3/4 - Compression reinforcement

The relevant average characteristics deduced from tension

tests are given in the table below and fig. (20) gives

typical stress strain diagrams of every size used. -E was

6t
approximatoly 30 x 10 lbf/in

Size of bar dia. Cross-sectional Yield stress Ultimate

in Area in2 lbf/in2 stress

lbf/in2

7/32 0.0375 39 000 61'500

1/4 0.049 41 500 64 000

3/8 0.11 39 000 63 000

3/4 0.44 44 800 70 000

1
Deformed Hard steel

Only one size of this type of steel was used on some

occasions, exclusively as compression reinforcement. The type

used was the N° 1110 bar having a cross sectional area of

0.791n2, a yield stress of 62000 lbf/in2 and an ultimate

stress of 98000. The relevant stress strain curve is shown

(21). 5
on fig. The value of E was approx. 29.5 x 10 lbf/in2.

107
3.3 Geometry test beamsandLarrangement of reinforcement

All the beams tested had the same total height of 12".

The rectangular and T-beams also had the same web breadth

of 6", and the thickness of the flange in the thin-webbed

I-beams was the same as the standard 3" thickness of the T-

beam flanges. The geometry and reinforcement of cach section

is discussed separately in subsequent paragraphs but the

following comments are of a more general nature.

1) In order to provide anchorage conditions for stirrups,

similar to those encountered in practice, nominal compress-

ion steel had to be included in all the beams with stirrup

reinforcement; to avoid an unnecessary variable the same

compression reinforcement was also provided in some of the

beams without shear reinforcement.

2) To avoid splitting of the concrete at the level of the

tensile reinforcement, under the high bond stresses the

deformed bars might develop, the prescriptions of CP11421

were followed and a 1" clear cover to the main steel was

provided.

3) It was anticipated that the formation of shear cracks

would place incroaeod demands on the anchorage of the main

reinforcement. Therefore, although non curtailed deformed

steel was used throughout the tests as longitudinal rein-

forcement, it seemed advisable to provide adequate anchor-

age in the form of extra length of 1111 on either side of the

effective span of the beams. The beam was thus extended 12"

beyond each support and heavy transverse reinforcement in

the form of 1/4"ý loops 2" crs placed in those extensions,

108
n
to improve resistance to splitting at level of main steel due

to the transverse tensions involved in the anchorage. This

arrangement proved to be satisfactory.

4) The position of each stirrup was marked on every beam

and the longitudinal arrangement of the stirrups is thus

apparent on the beam photographs.

5) When more than one layer of bars were necessary they were
99
normally grouped in groups of two. Robinson's tests proved

this to be a satisfactory arrangement in the sense that the

group possesses bond qualities equivalent to those of the

separate bars, provided the concrete could be vibrated

around them. In the present tests not only this arrangement

gave positive results but even when three bars were grouped

together in some instances, no adverse effects resulted.

The various details pertinent to each section used will

now be examined.

a Rectaular Beams

Two typical cross sections indicating the arrangement of

the reinforcement are given in fig. (22). The various do-

tails particular to each beam can be described in terms of

those cross sections and the following comments.

(1) Shear reinforcemont: The shear reinforcement was uni-

formly distributed throughout the whole of the span.

The various degrees of shear reinforcement used corre-

spond to the following sizes and longitudinal arrange-

ments of the stirrup loops.

rqfyq=83.5 lbf/in2 corresponds to No 7/32ý, M. 5. stirrup loops

placed at 611 crs

109
 C 'i
-i-_ --- _ ---- -,

i;
c :: 21 ý:ý

`N Stirrup loop has the scam irisi.!?

of th-- bar

I)
ý 3 No 31--Ii-n CASK530
J

2 rho 1" S' Hcrd steel

cä: 10

rr
cý"
Ito i rr;rýý J E::: >o

Fig. (22) Crocc, s cctiorn of Ruct r,:rulc! r b^rrr_ý

r
rgfyq=167 corresponds to No 7/32"9 M. S. stirrup loops

placed at 3" crs.

it it No 1/4"'0 M. S. stirrup loops

placed at 411 crso

rgfyq=326 N° 3/8"1 M. S. stirrup loops

placed at 4k" crs.

The above arrangements also apply to the case of in-

clined stirrups if it is kept in mind that the crs. distance

is always measured perpendicular to the direction of the

stirrups.

As mentioned previously the position of the stirrups

was marked in each beam prior to the test. It follows that

the photographs depict quite adequately that aspect of the

shear reinforcement.

2 fain reinforcement

The various arrangements of the tensile reinforcement and

the resulting effective depth d1 corresponding to the

different ratios of main reinforcement used are summarized

below:

rst_0.97% corresponds to 2 N° 16mm O bars in one layer and

d1 0.7

rst_1.46% If It 3 N° 16mm bars in one layer and

dl = 10.7

rst=1.95% 4 N° 16mm ý bars in one layer and

dl = 10.7

rst=4.16% 0 N° 16mm 0 bars in two layers and

dl = 10.

111
 (3)_Compression Reinforcement

As can be soon from the cross sections of fig. (22) a

concrete cover of 1" was provided in all cases. The different

ratios of compression reinforcement used correspond to the

following one layer arrangements.

rsc = 0.34% corresponds to 2 N° 3/8" j bars

rsc = 0.37% ,º ºº It ºº

rSe = 1.49% ,º2 N° 3/411

rsc = 2.98% r. ,º4 N° 3/411

rso = 2.61% It 2 N° 111 O ºº

(b) T-Beams

Fig. (23) shows the various cross sections used in the

case of T-beams. The arrangement of the reinforcement in

the web is in all respects similar to that of a rectangular

beam with the exception of the compression reinforcement.

To illustrate the latter as well as the arrangement of the

transverse reinforcement in the flange, the typical cross

section corresponding to all but four of the T-beams tested

is shown in greater detail.

All the comments made in the preceding paragraph (a) con-

cerning the arrangement of stirrups and main reinforcement

apply in the case of T-beams. To this however must be added

the detail of beam T7 which had a tensile reinforcement con-

sisting of six No 16mm 0 Ks60 bars arranged in two groups of

three bars, (oö). The relevant details in that case were

rst = 3.0% and d1 = 10.4""o,

1c Thin-webbed I-beams

The longitudinal aspect of these beams, together with a

112
 -ýypicý7jl

I 1f

F7 I

u
ý_.. I! ý; ca~ý in b,: r.rn
--- " ----- -
2) T21 C::, T 4
c r
1 `. 7114' V,: _' L:

d,.. 1 ^ýry
in li) ý-ýQrl {I tý_ l. I/ )'ý L"ý .. .ý, {rte Iýýý Ci
. _\ý1

11
`rn ýýN

- --- _-

T0

f'1'"ß ) iý_ c i" 6.'ýclýos

. 7.,
 i. e

04 C)
.-

JE u)

-H
f: L"' LL U c
4ý L
r C_-

ý. l "j
C:
Ld
'
t7a.
S_ O
_y

'ti CL
J-

F.
iJ
c
E

_ ., N
U
O
a9
i
11
ý-.____ r-ý
c
Iý
Ö
ti
ýý
. :..
_Li'2 S
.:
£

f-.

I! L... /i//
L
a.
C'

__--- --
CS

r7
1 ,,.
/I
a
r:
cý f
I
`
-n-- ----..
cý 'lf) ý l7
Li
r c'j

a)
C'
`CV
('4

--- ýý

..
ý,,
c)

Elcva".; cri c;*-,, c i ic ii-Hin d 1 l,:7.0t S

-! .... ý.,.._.
typical cross section in the shear span are shown in fig.

(24). It can be observed that the web width was increased'

in the region over the supports from 2.5"x'to 10" in order

to simulate conditions encountered in practice and also to

ensure that the failure took place in the shear span. In a

manner similar to Leonhardt's tests but unlike Robinson's

practice, no buckling diaphragm was provided in the region

of application of load as no important effects of that

nature were anticipated.

To render the arrangement and position of the reinforce-

ment as clear as possible, the cross section shown in fig.

(24) contains only one of the two similar stirrups used to

form the equivalent of a stirrup loop. These two similar

stirrups having the shape indicated in the above

- each

figure - forming one loop were placed symmetrically on

either side of the beam axis. The transversal distance

separating them in the web was that provided by the l6mm

tensile reinforcement bar positioned between them; longitud-

inally they were in contact.

No 10mm 9 Ks60 was used as shear reinforcement in all'

beams except in W5and W7which were provided with N°12mMVKe60

N° 8mm V Ks60 respectively.

and

3.4 eCasting and : Curinýc

The reinforcement for each beam was made up as a cage

which was located in the moulds using plastic spacers.

Steel moulds were used for the rectangular beams but wooden

115
 shuttering was used for the more complex shapes of the T-

and I-beams. In the case of the thin-webbed I-beams chamfers

had to be provided to facilitnte demoulding,

The concrete was weight batched in

and mixed a4 cu ft.

Eirich counter-rotating pan mixer. From two to five batches

were used for each beam, and the accompanying control speci-

mens. The control spocim ns per batch consisted of one 6"x9"

cylinder for split cylinder test and two 6u161lx6" cubes and

one 6"x12" cylinder for compression tests. In 25% of the

beams a 4x4xl2 prism was also-obtained from the batch going

to the top of the beam - in the I-boams the batch corre-

sponding to the web was used - for determining the stress

strain charactoristics of the concrete.

After mixing each batch of concrete was poured in a uni-

form layer in the moulds and was vibrated until full com-

paction, then the top of each beam was levelled and finished

with a steel trowel. To ensure correspondence, the same pro-

cedure was followed in both beams and accompanying control

specimens.

Following casting the beams and control specimens were

covered with wet sacking and polythene shbeting. On the

third day after casting they were removed from the forms and

then cured a further two days under wet sacking and poly-

thene. Thereafter they remained in the normal laboratory

atmosphere - relative humidity 50%, temperature 68 to 70°F -

until the'time of testing., The vast majority the beams

of

were tested at a total age of 14-17 days; the accompanying

control specimens wero tested on the same day as the beam.

116
3.5 Test procedure

The general arrangement of a beam in the testing frame

is illustrated in fig. (25) and (26). The only to

exception

this arrangement of central point loading were the short

beams R4, R5, R6 which tested under two point loads.

were

In general the load was applied by a 50 ton capacity Amster

testing machine through a 61Ix4l xl" steel plate. The beam

supports were 12" wide bearing plates carried at one end on

a 1" diameter transverse steel roller resting on the hori-

zontal plate, and at the other end on a ball setting; this

gave the beam freedom of horizontal movement and simply

supported conditions. Both the support plates were fixed to

heavy concrete blocks resting on the concrete floor.

Loading was applied in stages and the load incroasement

corresponded generally to an increase of 2 tons in the

applied shear. Near load stages at which a significant

change in the beam behaviour (e. g. shear

was anticipated

cracking) the increment in the shear force was reduced to

1 ton, or even less, until the parameters affecting the be-

haviour at that stage had been investigated. A1 ton incre-

ment in shear force was also used near failure, to enable

readings and observations at a level as close to failure

as possible.

At each load stage the following measurements were taken:

(1) Readings of Democ points on the concrete surface
(2) Readings of strain gauges positioned either on the

concrete surface or, more often, in the steel rein-

forcement

117
 c,

C.R0j I -" cý IhH

j

I. c> t1 (l C) I
;,!;

4' i.

Vn c; c1
ö

C)
JJ
I,

c)
v E
u
F- eý
0 f.,
Lf) Lý
Lei
L t)

Vi
E.,
E

j
U

f. 4)
.
'. " ý. - , ', - .? ((1
1

X ci
L.
Ci-
t..

C7

I. y_,
()_
C) Ui
'I)
U
I_
cv
U
F 0
L)
I'. 0 ý';
0 L n:
.4.J G]
C)
Vf,
C
U C G1
L fL
C'
0 Ou
U
O
Ct:

\-
ý_ýý.. fi.
.s!

I-

II

ýi

""

r3 I3
--""""

Fig. 26 : Test arrangement

li9
 (3) Deflections

(4) End rotations

(5) Cracks were drawn and their width measured

These various measurements which are presented and dis-

cussed in the relevant chapters 41 5i 6y and 9 will now be

described briefly

Demec points:

In the majority of cases the concrete strains were deter-

mined by placing demec points on the concrete surface and

measuring the strains between them with a mechanical extens-

ometre. The distance between the demec, points was either44""

or WW' and they wore used both the top,. surface and the
l -on, .

side of the beam. On the top surface they, were placed above

the centre of the web and they holped,, to determine the,

longitudinal top fibre strain profile and thus to study the

influence of shear, on the beams' internal structural behav-

iour. (If. the uncracked concrete carries any. shear after

shear cracking-it would seem that it must do so by an in-

clination of the thrust in it; this would cause a-more--rapid

reduction of top fibre strain towards, the supports than

would otherwise be obtained. ) A, similar task was performed

by demec points placed on the lateral -sides of the beam, -

0 including the sides of the T-beam flanges, in rows parallel

to the domec arrangement along the centre line.

Strain gauu a

Those were used mainly to measure the strain on the tens-

ile and shear reinforcements. On the tensile reinforcement
two 10mm gauges were placed in diammetrically opposed sides

120
of the bar to avoid readings resulting from the bar's own

bending. In the case of stirrups one 10mm gauge per stirrup

was used and its position was determined by the expected

shape of the main shear crack.

On occasions 30mm and 60mm gauges were also used on the

concrete; they were either placed between demec points, sub-

stantiating the latter's reading but also providing readings

up to the instance of failure, or placed on sides of the

beams, horizontally or at varying angles; the purpose of

this last arrangement was to measure the strains in the web

due to the inclined compression there which is a necessary

part of the 'truss action" of the stirrups. However, because

of the vast number of such measuring points necessary in

the study of web crushing failures, this function was per-

formed in that case by 4" demec points.

Depending on the number of gauges involved readings were

taken either on a 50
manually on a Peekel strain reader or

channel automatic strain recorder.

Deflections: These were measured by dial gauges measuring

to 001". Seven dial gauges symmetrically

.

placed were used per beam.

Rotation: Twoelinomoters were placed at each end of the

beam in the region beyond the support.

Cracks: After each load stage the cracks corresponding to

that load stage were drawn on the beam and their

width measured through a telescopic type crack

measurer graduated to . 001". The maximum widths of

the wider shear and flexural cracks were measured

121
 on each occasion but in some beams a more complete pict-

ure of the crack openings was obtainod by measuring the

widths at various height levels of the crack.

When failure had taken place the horizontal projection

and height of the failure crack was also measured. This was

done in order to substantiate the relationship between crack

lengths and ultimate neutral axis depths in shear compress-

ion failures, and as it turned out it was instrumental in

determining the more or less unique inclination of the

crack in the shearing type failures. In the T-beams the

crack was considered to extend from the level of the main

steel to the point at which it entered the flange when

failure was imminent. It is the horizontal projection of

this length that, together with the height up to the web-

flange junction, is reflected in the reported inclination

of the critical shear crack (chapter 5).

The control for tested

specimens each beam were on the

same day as the beam in a 300 ton Amsler testing machine.

3.6 Parameters investioated

The various parameters that were varied and their in-

fluence on the shear strength of the beam tested are pre-

and discussed individually in 4,5,

sented chapters and 6.

A synopsis of the extent of the range in the major

which

parawoters wore studied is presented in fig. (27).

Naturally not every parameter affecting the shear strength

122
 100 As `/b d1 tq fyq in lbs/in2 a/d1 100 A5c Ibd1

1.0 1-5 12-0 4-2 0 60 85 170 133012--3 3.5 4.5 5.0 0 0.4 1.5 2.6
X X x x X 2000 Ucyl
X X Fx X X 4500
X X X X X X X X X X X in 1-in
X X X X 3: J7500
100Ast j
X X X X X X 1.0
X X X X X X X X X 1.5 b d
r 1
X X X 2.0
XI X X X X X X X X 4.2
X X 0 f I
r q yq
60
in
x X X x X Y. x 85 2
170 lbs/in
'X X !X X

XX 23 °/d1
XXXX3.5
Rectangular beams X 4.5
X 50

ý: t/ci1 Uyl Ibs1in2 100 Ast/ brdl rc, fyqin lbs/in2 a /d1

3 0.6 000 1.2 1.5 2.0 3.0 4.2 0 60 85 170 3, 2.3 3.5 5.4 7.2
ýý
:. ý X X 2.0 b/ br
ýý X
ý. z
ýý x X X X X x x X X X X X X X X X X X 4.0
? S 7.0
x x x x
x`
X X X X X X X X X X" X X X X X X 0.3 t/d
_ 1
X X X X X 10.6
X X X X x 2000 Uc l
y
X X x X X : x X X X X X X 45'°°
in lbs/in2
x X x X X 7500
x x 1-2 100A
st
X X X X X X 1.5 b d
x x X 2.0 1 1
X X 3.0
= X X X X X X X X 4.2
X X 0 r q f yq
x X 60
in
X X x x 65
x x X x 170
1 1bs/in2
arns x 330
-1 -P-

/d1
100 rg s sin; tz
2.2 24. 3.3 0.3 45 06 45° 40°
X X 2000 U
CyI
x x x X X x X X 4500
x x 7500
.x
X x 'X x 2.2 100r
q
X X X 2.4
X 3'3
X 0.3 S511)ýy
x- X 0.45
beams
0.6

Fig. (27) Ranges of variation of the main parameters investigated

ýý,
could be fully investigated, even with tho testing of 75

beams. According to Leonhardt, in a lecture given at Imperial

College, there are 22 parametors that can affect the shear

strength of reinforced concrete beams. Based on a study of

relevant literature and test results the parameters investi-

gated were chosen as being the most significant, in the sense

that if a theory can reflect their influenco its predictions

4calc
will be within the same normal values intrinsic to
test
reinforced concrote members, -even in, the case of extreme

variations of tho parameter.

124
 CHAPTER 4

INVESTIGATION OF SHEAR FAILURE IN RECTANGULAR BEAMS

In the opening section of this chapter the general behav-

iour of rectangular beams failing in shear is described, based

on observations during the tests performed. The subsequent

sections are concerned with the actual tests. The limited

number of tests on beams without shear reinforcement is de-

scribed fully in section 4.2, while sections 4.3 to 4.8 are

concerned with tests on beams with stirrups. These have been

divided according to the parameter studied, so that a clearer

picture of its effect can be obtained. The final section is

devoted to the deformation characteristics of concrete and re-

inforcement during the tests of beams with web reinforcement.

To avoid repetition, numerical values of deflections and end

rotations aB well as full dotails of. crack width measurements

are not. roported here but in chapter 9 which deals specific-

ally with serviceability criteria.

4.1 Behaviour of rectangular beams failing in shear

At early load stages the behaviour of such beams is similar

to that of beams failing in flexure. After the tensile resist-

ance of the concrete at the extreme fibre has been overcome

flexural cracks appear first in the tension zone at the regions

of maximum moment. As the load increases these cracks rise

vertically to a height near the neutral the same

axis, while at

time, (provided the beam is well bonded), similar cracks appear

nearer to the supports, and secondary cracks form between the

125
 I-
R7

-4
nt

rf=O Diagonal tension o/d1 = 336

q YQ
failure

n R6_
n - ...
ýr_ iý

Inclined action ald1= 2 24

rq tyq =U
failure

R13 n
10

i
37°/°
rgfyq= 1671bf/in2 a/d, =36 sc-0
shear compression
failure

R16- n

rqfyq: 167 Ibf/in2 a/d1= 36 rsc =298

shearing failure

R23
B
17 B
4168 ýI @ MS
%4ý Ký"'gýýc

5 aI d1 -2 24
rqfyq : 83 inclined action failure

Fig 28 : Behaviour of beams failing in shear

rectangular

126
existing ones..

When the load, is further increased the above process con-

tinues; but at some stage an inclined crack appears. This

crack, referred to as_a shear crack normally starts at the

head of an existing flexural crack, but can also appear inde-

pendently. It travels in an inclined direction upwards and

towards the region of maximum moment.

The behaviour of the, boam from that point on depends mainly

upon the length of the shear span, -and the presence or ab-

sonco of shear reinforcement.,

In the case of beams without shear reinforcement having

shear to depth ratios, (d ) greater than about

span effective
1
3.0 (see fig. 28: beam R79 a/d1=3.36) and subject to point

loading, the shear crack opens up rapidly. At this stage or

at a load only in oxmss of that at which it first

slightly

appears it reaches the level of the main steel, if it has not

done that before. It higher than the flexural cracks,

rises,

its direction becoming flatter when progressing into the com-

pression zone. It should be added here that the boundary

d is
value of not a fixed one but tends to vary somewhat,
1
being lower for smaller ratios of main longitudinal reinforce-

ment. The opening of the crack results in a tendency of the

side of the beam where the load acts to, move downward in res-'

pect to the support side of the beam. Consequently the main

steel exerts on the concrete of the support side a trans-

versal farce in a downward direction which tends to tear away

the cover. This is a manifestation of the dowel effect. At the

same time the upper end of the crack extends right through the

127
compression zone and the beam collapses in a sudden manner.

This type of failure is usually referred to as Diagonal

Tension.

In shorter beams collapse does not occur at or near the

load causing the first shear crack, the progress of the crack

in the compression zone being much slower. It is difficult

to establish exactly in what manner the final failure takes

place. In the present tests it took the form of a "shearing"

of the region between the load plate and the shear crack,

(see fig. 28: beam R6, a/d1=2.24) although failures by

crushing of the concrete have also been reported (94) (32)

(100) in other tests. Since however the region of comprossive

strain at the top fibre is very small this crushing takes

place along an inclined direction and may be much the same as

shearing. To avoid too many separate definitions this type(or

typoo) of failure typical of short bomms will be referred to as

"Inclined Action" failure.

In the case of beams with reinforcement the latter's

shear

presence does not seem to have any significant influence on

the loads at which the first flexural and shear cracks occur.

Its influence on the subsequent behaviour of the beams is

however very important. In general it allows collapse loads

higher than those at which shear cracks appear, to be with-

stood. In physical terms it tends to restrain the progress of

the inclined cracks into the compression zone and to oppose

the tearing action of the main steel. In this manner even

small amounts of shear reinforcement are instrumental in pre-

venting the occurence of diagonal tension failures.

128
 In the case of beams with d (the
ratios greater than 3.0
1
reserves expressed previousl; - concerning this boundary value

are still valid) containing only small amounts of compression

steel and well detailed main stool, provision of shear rein-

forcement will ensure that if shear failure occurs this will

consist of crushing of the concrete at the head of the main

shear crack. The latter by its progress above the flexural

cracks reduces the available compressive zone to a fraction

of the flexural compressive zone and causes higher stresses,

and thus failure before the flexural capacity of the beam is

attained. This is the shear-compression (S. C) mode of failure

(see fig. 28: beam R13, a/d1=3.6, rgfyq=167 lbf/in2, rso -

0.37%).

If very light stirrups are used, but mainly if the amount

of compression stool available is excessive, then a shearing

type of failure (see fig. 28: beam R16, same as R13

occurs,

but rsc=2.93%). In the between the load platten and the

area

upper po-ti,:,. of the shear crack, though not exactly the head

of the latter high and forces

shear crack as the rises very

must be transmitted through some of its upper portion by

interlock the concrete is sheared through at an inclined

plane. A vertical movement seems to accompany this last action.

The term excessive used in referring to the amount of com-

prossion steel can be construed to mean that the compression

steel present is considerably in excess of what is required

to ensure that flexural failure take
will not place at the
load at which shear failure occurs. This is the case of beams

failing in shear and having heavy compression steel and light

129
or intermediate shear reinforcement. If heavy stirrups and

light compression steel were used in these beams then a flex-

ural failure would occur. The use of heavy compression steel

with heavy stirrups could make a shear failure possible before

the increased flexural capacity was reached. In that case a

shear compression failure could take place - see section 2.3

beam R28.

In the case of shorter beams although loads higher than

the ultimate loads of equivalent beams without shear rein-

forcement are withstood, the beams' behaviour is not very

different from that described for that case (see fig.

already

28: beam R23, a/d1=2.24, rgfyq=83.5 lbf/in2).

Some additional comments on the way shear is carried in a

beam can also be made at this point on the basis of deform-

ations measured on the concrete and reinforcement during the

tests. Strain measurements on the stirrups show that they are

virtually unstressed until shear cracking occurs. Following

shear cracking the stresses in them augment rapidly and in

the case of the ones crossed by the main shear crack, yield

is attained by the time failure occurs. Strain measurements

on the concrete at the top and-sides of the beam indicate

the existence of inclined compression. The more or less uni-

form strains measured on the main steel along, and sometimes

beyond the length of the shear crack, are an indication of

the existence of a degree of tied arch action. It can be seen

that in a beam shear can be carried by a combination of four

possible actions of varying importance: -

1) Truss action involving tension in the stirrups and in-

130
 clined compression on the concrete web.

2) Arch action involving tension in the main steel, and

an inclined compressive thrust in the concrete above

the cracks.

3) Interlock action allowing shear to be transferred

across the cracks.

4) Dowel action by the main steel.

The adjoining sketch on page 132 illustratos these

actions.

4.2 Rectangular beams without shear reinforcement

Because of the largo number of test results available only

a very limited number of tests were performed on rectangular

beams without stirrups, the main subject of this research be-

ing the shear strength of beams with web reinforcement. These

tests were carried out mainly in order to provide comparisons

with the tests of otherwise similar beams containing stirrups.

They help to illustrate the Diagonal Tension type of failure

as well as the difference in behaviour between short and long

beams. Table 4.2.1 and fig. 29 present a summary of these

tests which will also be discussed individually.

131
 2
1,1

i:., Gll iý. U"L1ý"ýýii, J 'iii: ' 'ýil.: ýt' i'ý`_->l:, lily CCýi! uýý":ýý Ifl Q
Sc
ý1ý"_'i.

('c: Ci'CCG CC:i"tClWIýe ý:: C1iT1

 rRi A NEN
ýý9G9

o/d1 =336

n
n

o/d, : 3.36

ý1 R4_ý.
n_... 4

a/d1 -- 2.24

02

fit

Ito

ý, t
a/d =336

III 111

Fig 29 Shear failures

: of rectangular beams without shear reinforcement

063
TABLE 4.2.1. TEST RESULTS FOR BEAMS WITHOUT SHEAR REINFORCE-

mENT

Beam Uc a/d1 Ult. qult Failure

1 st%
lbf/in2 Load
lbf/in2 Mode
tons

R1 3800 0.97 3.36 9 157 DT

R2 3800 1,46 3,36 9,5 165 DT

R3 3600 1,46 3,36 9 157 DT

R4 4950 1.46 2.24 17 296 IA

R5 4950 0.97 2.24 9.5 166 DT

R6 5000 1.46 2.24 14 244 IA

R7 4070 1.46 3.36 11 192 DT

D1(1) 3320 1.46 3.36 6.75 110 DT

D2(2) 4400 1.46 3.36 10,5 183 DT

(1) Artificial beam with preformed cracks

(2) Maximüm size of coarse aggregate was 3/8"

in all the other beams 3/4" was used.

DT = Diagonal Tension, IA Inclined Action

In general it can be soon that in long beams the failure

mode is one of Diagonal Tension, the load at failure depending

on the ratio of main steel and the concrete strength. In the

shorter beams the mode of failure changes for the same (a )

dl

value from an inclined action failure in beams R4 and R6 with

a ratio of main steel of 1.46% to a diagonal tension failure

in the case of beam R5 with a ratio of main steel of only

0.97%. In the case of beams R4 and R6 much higher loads were

registered than in the case of the otherwise similar but

134
longer beams R7 and D29 thus substantiating the findings of

other investigators - Moir ow and Viesio Leonhardi6 Beam R5

on the other hand failed at about the same load as its longer

counterpart R1, provided an allowanco is made for the higher

concrete strength of beam R5. The 20% difference in load bet-

ween the two similar beams R4 and R6 is a good indication of

the scatter of results to be found in tests of short beams.

No significant difference in behaviour resulted from the use

of a different size of coarse aggregate in beam D2 while

beam D1 provides an indication of the effect of dowel and,

to a lesser degree, aggregate interlock action.

Behaviour of Beams in tests

Beam R1 The beam failed in diagonal tension at the load at

which the first shear crack appeared. Collapse occurred

suddenly, approximately five minutes after this load

(9 tons) had been marking of cracks

applied, while

was in progress.

Beam R2, Diagonal tension failuro occurred at 9.5 tons, the

first inclined crack having appeared at 9 tons. Upon

its appearance the shear crack opened up to a width

of about 0.015" along the central three quarters of

its length petering off slowly towards the extremities.

The widths of all the other cracks remained quite

small (of the order of 0.002" to 0.003"). A small in-

crease of load after that brought about the failure of

the beam.

Beam R3 Failure at 9 tons was preceded by the rather early

appearance of an inclined crack at 7 tons. This crack

135
 progrossed comcwhat at increasing loads before

cutting through to the top surface. The shear crack

did not however open up until the failure load was

attained.

Beam R4 Shear cracking occurred at about 10 tons and the

length of the shear crack increased slowly, at subse-

quent load stages in the direction of the 'support to

loading plate' axis becoming flatter at both ends. The

crack width increased attaining values of about 0.03"

to 0.04" along most of the crack length at 13.5 tons.

Three minutes after the load reached 17 tons a crack

appeared extending from the*top of the beam at the

edge of the load plate, towards the shear crack, and

after an additional two minutes failure occurred at

that crack.

Beam R5 Shear cracking appeared at 9 tons and a very sudden

diagonal tension failure occurred at 9.5 tons.

Beam R6 Small, thin inclined ceacks appeared at 10 tons

and by 12 tons the main shear crack extended almost

the whole length of the beam. Failure occurred 14

at

tons and was morn sudden than that of R4, but not as

sudden as the diagonal tension failures.

Beam R7 Shear cracking in the form of small inclined cracks

occurred at 10 tons and when the load was taken to

10.5 tons nothing changed visibly. However when trying

to increase the load a sudden diagonal tension failure

took place at 11 tons.

Beam D1 In this artificial beam containing preformed cracks

136
 and crack formers a flexural crack daVclopod at 3tona from

the preformed crack, while no flexural crack had ap-

peared in the region of maximum moment. This crack

started to extend in an inclined direction from that

load onward, and a diagonal tension failure took place

at 6.75 tons. It was noticeable that the major cracks

originated at the preformed cracks and the failure

crack followed the direction of the crack former in

its flexural part. This is an indication that some of

the interlock action across the crack was also removed

by the presence of the crack former. The crack formers

wore thin 61"x6l' pieces of acetate paper placed vertic-

ally above the middles of the preformed cracks, They

were not removable after casting so obviously some

interlock action still occurred through them by virtue

of their Unintentional curvature.

Beam D2 Inclined cracks appeared at 10 tons but did not ex-

tend significantly. However at tho next 10.5

stage of

tons a sudden diagonal tension failure occurrod.

4.2.1 Strain measurements on concrete and reinforcement

Typical top fibre strain profiles for two values of (ý )

dl
are shown in figs. 30,31, and 32. It can be seen that up to

the shear cracking load strains follow a flexural pattern, in

that they are compressive and increase in magnitude from the

support to the point of application of load. At loads higher

than shear cracking however, while the strains near the load

continue to increase the ones near the support start decreasing

137
 0

CL
41 41
COiý (DP ýs N

"ýI

nI

(uoissaadwoo)

O
O U)
U11.9-OL X UI ut Suloa}$

F1q. (30) Load beam R3

strain curves- profiles of top_ fibre strains

t7^
 _`vl I

,..
ý-'-

0_

Nt

l7
ca

ä-
II
ci
r
sý
i-ý

C]
.
CU
tL t_L
to 1 C)
rý
0
c
_!
r' "y
.ems C)
0.. Cd r- C7
CT) LO
0'. ý t., J)
-ý0,

G
N L
Ct.
v : `ý ö

_)
L

1i 1 C
1 C)
L

.ý
\
El
ýý U

_1

ýý
A ýb

c
UL

( toy:
c) I_
- O-
ý> c:ý c>
C)
QOGC,
UOOOOU
ý C% C;> G' _)-T
ýy
C:) OO
(lJ to ` CV ý- Cr3

i
 1
2000
1 I

15001 i

R",
F-.
1ooc -- 24 24 24 -

'"ti

°z.,50C

f "° " "". 11tons

,P-. -0
pl --- '-0-- -----. &- at

0
0
c

500 12t

Fig. (32) Load strain curves - profiles of top fibre strain beam R3

strains in main steel

-152
0
'"ý----. P _4 tons
ýý'ý

} 6t

11t
O. ,A
1000 .
'1:
' L7
..

12 t

Fio, (32a)Load strain curves - deformation of main reinforcement beam R,6

%4.O
and can become tensile. Upon further increase in load the

region of tensile strains increases till at failure only a

small area of the top fibre near the load is still under com-

pression. This is an indication of the inclination of the

main compressive force in such a case.

Figure 32a shows the main steel deformation along part of

the length of the shear crack for a short beam (R6) that did

not fail immediately upon shear cracking. It shows a tendency

for the deformation of the main steel to become uniform under

the shear crack.

Figure 33 shows the compressive strains in the side of a

short beam (R4) and provides another indication of the in-

clination of the main compressive thrust. The strains at three

different distances from the load are shown. To render the

presentation meaningful the exact measured strains arö plotted

for the two positions furthest from the load considered; only

the increments from the previous load stage are shown in

connection with the section closest to the load. The demec

points used for these measurements were positioned after

shear cracking had taken place.

Figures 34y 359 and 36 are concerned with the measurements

taken on beam 02 which failed in diagonal tension at 10.5

tons total load. The positions of the measuring points and

the crack pattern are shown in fig. 34. Figure 35 illustrates

the strain across five sections along the beam at tho 10 ton

load stage, at which the first inclined crack appeared. The

strains seem to be linear from the neutral axis to the top

surface even at the head of the small shear crack although by

141
 P12
0 C

(3060)
in in x lÖ 61 in
total total strain in in x 106/in
1 13tons
(2200)

(1240)
2

1 (658)

I'i ck

KEY

CL b11
5-0
41
E
P
li
12 tons

6s
w
I' Successive
24"

(
247

)9tons A
247

13.5 tons
c

U
I' Load (P)
Stages o 12 tons a 15 tons
e
crack The same scale is used in section C
a
as in section B and C but in this case
the readings represent increments of
strain in respect to the previous load stage.
() Actual compressive strain
8
in in x 1Q61in at the level considered
corresponding to the 9 ton load stage.

ti

Fig. (33) Load strain curves. Profiles of strains (compressive) above the critical shear crock
at various sections along the shear span of, beam R4
1 %4Z
 2ýý 4-5 13-25 12.5" 31"
=Jý

P
_.., _. _ 12 - -13 -29 -J -3
-2© 3 -36 failure crack
13- -19 -27 - 37
-26 -N 38
K
IS- -20 -25 7.5
-24 6 7.5 "5 7.5
16- -_21 775 --23 7. 6
566
12 55
10ý II 7.5 5
14 4
5(/5 47
44

4
-; --- - --
4- locid stage (tons) indicating
extent of crack

H 11 so
1 ll-5Qýý_ t 1225 17.75
f-
___F

scale lM ,1

Fig. (34) Polition of meosurin9 points in beam D2

145
 "ý
x13 C"A
/ KEY
///E
A 1b a section A
g///
a' B
4J e x ,.
see Fig.
EdC

2I C(J
/
13

4.

5't
Fig. (35) Strain profile across various sections along
i
the shear span of beam D2 at the 10 ton load stage.
I'D

100 200 300 400 500

in 6
inx 1O /in
strain
x
" r
x

x
x
x

x KEY
x o section 0 -1
oD x 11 2-47 sze Fig.
ox A 3--4
o
x
oA

Axo Fig. (36) Main steel deformation in terms

o
o of applied moment at various sections,
o along the shecr span of beam D2
, .
Go
xo

AO
0
O

100 200
applied moment at section considered in ton - in
the next load stage of 10.5 tons the beam collapsed in dia-

gonal tension at that gack. The disturbances at the lower

ends of sections C and D are probably an indication of the

transfer of some forco across the teeth formed by the flexural

cracks.

Strains measured on the main steel are shown in fig. 36.

The steel deformations at three different sections are plotted

against the external moment acting at those sections. The

relationships are linear up to values of applied moments

corresponding to a load of approximately 8.5 tons. From that

load stage on, corresponding to the appearance of small shear

cracks in the form of inclinations of the flexural cracks,

the strains in the steel tend to be in excess of the values

corresponding to linear flexural behaviour. Table 1 appendix B con-

tains all the measurements taken for the complete range of load stages.

4.3 Influence of the doe of shear reinforcement

Six tests were carried out to investigate the influence of

this parameter on the shear strength and general behaviour of

rectangular beams. These tests can be divided into two groups.

One group had a low ratio of main steel (1.46%) which limited

the amount of shear reinforcement that could be used if a

shear failure were to take place. The next group had a high

of main steel (4.16%) and allowed the influence higher

ratio of

values of shear reinforcement to be investigated. Both groups

were tested with practically the same a/d1 ratio and the con-

crete strengths did not vary significanly. The only other dif-

ference was the use of 2 No 1" diameter bars of compression

steel in beam R28, necessary to ensure a shear failure at a

145
high level of shear reinforcement without changing any of the

other parametors. To render the comparisons more meaningful

the result of beam R16 tested in connection with tho study of

the influence of compression reinforcement is also included

in this section.

Table 4.3.1 lists the relevant data and results for the

beams concerned. Photos of these beams are grouped in fig. 37.

Table 4.3.1 Influence of the degree of shear reinforcement

j
Beam Ucyl rst% Shear rgfyq (a/dl rSc% Ult. quit Failure
Load
No lbf/in Reift. 1bf/in lbf/in mode(2)
tons

R7 4070 1.46 - - 3.36 0.34 11 192 DT

R14 4210 " 7/3211VO9" 55.7 " It 18 314 S

RB 3870 " 7/32"0@6' 83.5 " " 16 279 Sc

R9 4290 " 7/3211Q3" 167 " It 21 366 Sc

R12 4920 4.16 7/3211j'Q6" 83.5 3.6 0.37 22 411 Sc

R13 4680 " 7/3211'03" 167 it it 30 560 Sc

R28(1 4580 3/81104ý' 326 " 2.63 36 671 SC

R16 4580 " 7/32"03" 167 It 2.93 28 522 S

(1) Compression steel fy 60 000; all other compression steel

used f ti 40 000
Y
(2) S= Shearing, SC Shear-Compression: Indicates relati)e

predominance of a given mode of failure.

_
In fig. 38 the ultimate shear strength of the beams tested

is plotted against the degree of shear reinforcement. The

former is expressed in terms of the nominal shear stress

qul and the latter in terms of the product rgfyq which

bdlt
1

146
Ru n
,
Ie wI".
\"K
ný
n i2

rgfyq 557 Ibt/in- st 46 °/°

'1

n Rs n
ý-
_.

2
r Ibf /in
q yq

%7 R9

Tq'z
rq 1671bf/in2 rst 1.46 °/°

r R12

I
rq fyq: 83.5 Ibf/in2 rsc z O37/ rst =4 16 °/o

R28
.
1ý

ý fl

iI
INA
q tyq: 326Ibf/in2 rst :4 16 °/a
rsc = 2.61°/c

Fig 37 : Influence of degree of shear reinforcement : rectangular

beams foiling in shear

47
 600

400

.S

ift
N

zoo

ä
100-

d 100 150 200 250 300 ä5^

50
Degree of sheor reinrorcernent rq fyq in Ibi /iri2

a tests by Sund nand rst= 378 and 4.51°1°

o/d = 3.95

Fig. (38) Influznce of sheer reinforcement on ultimcte shear strength

.

f
-.4-
0

I C{
 0
ý'
O
ry

Cl)
LO
auO
vO
LN LN %T

p
LO

0
NN
0
LW LN

oD

O
0

o0

cl'

n
Ln

0
O
0

L7

ljgl 4Inb ssaJls ünags alowIlln

2ui U!
LIl1., 1
8000 --
0 8 00
O0

Fig. (39) Influence of shear rcinforcement on ultimate sheer strength

tests by Sunderland (a Id, =4 0)
also appears in table 4.3.1 where
A
r= ratio of shear reinforcement =p bs and
q sin
Aq = area of the two sides of a stirrup loop

s= spacing of stirrups along the axis of main steel

fyq = yield stress of stirrups.

In fig. 39 the results of tests by Sunderland(110) are

plotted. These beams had ratios of main reinforcement similar

to the higher ratio used in the present test (from 3.78% to

4.51% to 4.16%) d
compared and an almost equivalent value
1
(4.0 compared to 3.6). In Sunderland's tests compression steel

was also necessary in order to obtain shear failures at higher

rgfyq values. In both those figures lines representing the

marsch truss theory are drawn taking La=0.85d1, as well as

lines parallel to the forsch lines drawn from the ordinate

corresponding to the shear strength at zero shear reinforce-

ment.

Basically it can be stated that the influence of the shear

reinforcement is to increase the shear strength of a rectang-

ular reinforced concrete beam. However the rate of increase

is greater at low values of shear reinforcement than at high

ones. This effect seems related to the physical behaviour of

the beam as will be discussed subsequently.

It can also be seen that the addition of even a small amount

of shear reinforcement in a beam, that would otherwise fail in

diagonal tension, results in a failure not only at higher

loads but of a different type as well.

Flexural cracks appear earlier and are more numerous at

lower load stages in the case of beams with low main steel

150
ratios. However the web reinforcement does not seem to have

an influence on the load at which the inclined cracks - that

would have become critical in its absence - appear. On. the

other hand these inclined cracks appear at higher loads in

the case of beams with more main steel. In brief it can be said

that the inclusion of stirrups in a beam has no important

effect on its behaviour prior to the formation of diagonal

cracks. Consequently in the case of normal beams (a/d1 > 3)

reinforced in shear, the shear cracking load Qcr can be con-

veniently defined as the load at which an otherwise similar

beam without shear reinforcement would fail in diagonal

tension.

The above would tend to indicate that the shear reinforce-

ment does not become operational till after shear cracking.

This is also supported by strain readings on stirrups which

(see section 4.9.1) that the latter did not become

show

to any appreciable extent until a diagonal crack had

strained

Subsequently stirrups not crossed by any diagonal

appeared.

crack or its extension did not register appreciable strains

the ones crossed by the critical shear crack reached

while

yield stresses at failure.

The behaviour of beams following shear cracking will now

be examined by considering, progressively, beams with in-

creasing shear reinforcement. The pattern is generally simi-

lar except in the case where excessive compression reinforce-

ment was provided but this case will be examined in section

4.7 which deals with the influence of compression reinforce-

ment.

151
 Once shear cracking has occurred shear reinforcement be-

comes effective and hinders the opening and progress of the

shear cracks. The failure mechanism loses the selfpropagating

character evident in beams without stirrups. The load must be

increased in order to obtain progressively the opening and

extension of the inclined cracks. Although more than one

shear crack appears their number actually increases with the

degree of shear reinforcement - one crack becomes increasingly

wider than the others, especially at higher loads, and is

associated with the failure of the beam. This is

ultimately

usually referred to as the main or critical shear crack.

load is increased the yield

When the stirrups eventually

(see emotion 4.9), the main shear crack becomes quite wide,

0.025" to 0.040" and failure takes place by crushing of the

in the intact zone existing at the head of the main

concrete

shear crack.

In this failure mechanism the role of the web reinforce-

ment is one of opposition to the relative displacement of

the two parts of the beam separated by the main shear crack.

When failure occurs the part of the shear force carried by

the reinforcement assumed equal to the yield stress

shear

by the sum of the areas of the stirrups crossing

multiplied

the shear crack is only a fraction of the applied shear.

main

Most of the remaining shear is supplied by an arch-like action

involving basically tension in the main steel and inclined

in the uncrackod section of concrete above the

compression

crack. Thit arch action is the contribution of the uncracked

concrete to the shear resistance of the beam, and is sometimes

152
referred to as the shear resistance of the compressive zone.

A major role of the web reinforcement is thus to preserve

this uncracked compressive zone from a premature destruction

by the extension of the diagonal failure of the web.

In the testing of beams with progressively increasing

degrees of shear reinforcement it was noted that this pre-

servation of the compressive zone was better for higher values

of shear reinforcement. In other words the reduced neutral

axis depth, resulting from the shear crack rising higher

into the compression zone than the flexural cracks, was pro-

portional to the degree of shear reinforcement.

It can thus be concluded that the shear strength of a

beam is an increasing function of the degree of shear rein-

forcement for a double reason: 1) increase of shear force

directly by the shear reinforcement; 2) increase

supported

of the shear resisting capacity of the compressive zone.

While the above statements tend to explain the basic in-

fluence of shear reinforcement it remains that if no other

intervened the increase in shear strength of beams

condition

due to this parameter would presumably be uniform for both

degrees However,
low and high of shear reinforcement. as

the rate of increase is greater at low values

stated earlier,
(100)
than high ones. Robinson stating that the efficiency
at

the shear reinforcement diminishes at higher rgfyq values

of

this to be due to the varying importance of the

considers

zone in the internal equilibrium of the beam in

compressive

respect to shear. Although this is true some more specific

reason is warranted; this is provided by the length of the

153
critical shear crack. Observations during the tests showed

that this length is a decreasing function of the degree of

shear reinforcement. If two beams are reinforced in shear

with uniformly distributed stirrups having the same spacing

but different cross-sections, the number of stirrups crossing

the failure crack will be greater in the beam with lighter

shear reinforcement. It is therefore the combination of

these increasing and decreasing functions of the degree of

shear reinforcement that explain the shape of the curves of

figs. 38 and 39.

Furthermore the length of the failure shear crack in these

cases cannot be explained simply in terms of internal equili-

brium alone as various configurations of the crack can pro-

duce equilibrium conditions. There are two other factors that

may affect conditions governing the length of the crack -

interlock action and compatibility of deformations. The

question of interlock action has not been yet fully investi-

gated and its influence is not known precisaly; None of the

existing shear compression theories takes it into account.

It has been discussed in chapter 2 section 9 and assumptions

concerning its influence are made in chapter 7. In terms of

existing shear theories itic only compatibility that can have

an influence on the crack length. Walther's theory by assuming

a 450 inclined crack ignores this question and this consti-

tutes one of the major criticisms of that method. Bjuggren

does admit that the deformation in the concrete is influenced

by the exact shape of the shear crack but besides the fact

that the solution must be obtained by trial and error his

154
compatibility condition has the shortcoming of predicting

neutral axis factors greater than the flexural ones for

slightly inclined cracks.

Regan, by basing his compatibility on the total deformation

of the concrete block above the failure crack, automatically

takes into account the length of the shear crack in the com-

patibility condition. His solution charts are relatively easy

to use, no inconsistencies such as Bjuggren's exist, and the

crack lengths predicted are indeed a decreasing function of

the degree of shear reinforcement. The only criticism of

Regan's method in this respect is that it does not take into

account the actual contribution of stirrups in determining

the deformations. This question and a modification proposed

are discussed in Appendix A section 3.

Another effect of the stirrups manifest in these tests is

the better preservation of the concrete at the lower end of

the crack. Concrete splitting and dowel cracking are hindered

by the presence of stirrups. In opposing the splitting of the

concrete the stirrups act in a manner similar to the binders

(102)
used in bar pull out tests By opposing vertical crack
.

movements at the level of the main steel they tend to take

upon themselves most of the shear force carried by the main

stool acting as a dowel. Thus one way of taking account of

the dowel action in beams with shear reinforcement would be

to consider that more stirrups carry shear than those actually

crossing the failure shear crack. Since the dowel crack in

those cases does not extend very far beyond the main shear

crack and in view of the limited importance of dowel action

155
(see chapter 2 section 9 and reference (12)) its main effect

in a shear strength equation can be adequately approximated

by considering that the number of stirrups involved failure

at

in one more than number actually crossed by the failure crack.

This can be conveniently expressed by taking

ra_c
S
where N= number of stirrups involved at failure

c= horizontal projection of the crack length

s= spacing of stirrups

Comparing the test results with the morsch truss theory

shows how conservative the lattor is. Even the lines parallel
to the marsch line originating at the shear cracking strength

are situated below the experimental results. These lines are

representative of the attitudes of various National Codes

(ASCE-ACI, CEB etc. ) that have transformed the classical

marsch approach by adding a concrete term, equal to the shear

cracking resistance, to a JThrsch type term for stirrups.

Provided conservative values are used in the codes for the

calculation of shear cracking loads it follows from figs. 38

and39 that they will be most conservative when the degree of

shear reinforcement in a beam is about one half to two thirds

of that required to ensure flexural failure.

4.4 Influonce of the ratio of main reinforcement

Six tests were carried out specifically in order to in-

vostigato the effect of this parameter in beams containing

shear reinforcement. Two of the beams had a higher degree of

shear reinforcement than the others. Table 4.4.1 gives full

156
details of the beams and test results. For the sake of com-

parison the results of two additional beams containing heavy

compression reinforcement are also included. Illustrative

photos of the test beams appear on fig. 40.

TABLE 4.4.1. Influence of the ratio of main reinforcement

lqult
BEAM Ucyl rstf Shear rgfyq a/d1 rJo% ULT Failure
LOAD
2 2 2
No lbf/in reift. lbf/in TONS lbf/in Mode

R10 4295 0.975 7/3211-0611 83.5 3.36 0.34 15.1 263 SC

R8 3870 1.46 16.0 279 Sc

R11 3800 1.95 18.0 314 Sc

R12 4920 4.16 3.6 0.37 22.0 411 Sc

R25 4470 4,16 2.63 21.0 392 S

R9 4290 1.46 7/321103" 167 3.36 0.34 21.0 366 Sc

R13 4680 4.16 3.6 0.37 30.0 560 Sc

R16 4580 11 " " " 2.93 20.0 522 S

The results are also presented in a graphical form in fig.

41" Included in that figure are some results relevant to the

cracking of the beams in one of the series.

shear

Fig. 42 contains relevant results of other investigators.

All the above results are presented in terms of the geo-

metrical (rSt) rather than mechanical percentage of the main

reinforcement. Since it is unlikely that the main reinforce-

ment will yicld before a shear failure occurs, the geometrical

percentage is a more meaningful quantity.

It can be seen from fig. 41 that the shear strength of a

beam is an increasing function of the percentage of main steel.

157
 Rio
n-
1i IZ
/a74 1q 1' ýýk 1. `
ýý ý
tl 5g5
'' ß\ý ýq
,
ýýý4
u/% ,4 8 yLý

rst =0 97 °/°

n Rs n

çfl:T
a

Jtl st = 1'46°/°

ist = 1.95°/°

Q
..,
zo

16 ý ý:
16
16
Wo

!t

r 4-16 °/o
St

Fig 40 Influence ofratio of main reinforcement : rectangular

beams failing in shear

Iss
 10-0000
(D
1ýý
. t

ýQ
OC
,.
ýrý
©

ýrýoc

tn9
r ýcoýý;
ýýýo, KEY

O: heavy compression
reinforcement
a /d, =-' 3.5

1234_. 5

ratio of main reinforcement rst (°h)

Fig. (41) Influence of the ratio of main reinforcement

present - series of tests

i5.
 13 ,

N
G

ýo-

G
._L

07

t
N

4J
0

234.5
ratio of main reinforcement rat
J

Fig. (42) Influence of the ratio of main reinforcement- other tests

3
1;
The rate of increase is not uniform but has a tendency to

diminish at higher rst values.

Similar conclusions have been derived for the case of

(77)
beams without stirrups by Moody and Leonhardt and Walther
(65),
Unfortunately these tests did not extend to very high

main steel ratios - maximum rSt of 2.25% and 1.85% in the

case of 1oody and Leonhardt

P: respectively. Because of their

increased range some relevant results by De Cossio(30) and

Laupa(61) are therefore included in fig. 42; these show a

tendency for an increased shear strength when higher ratios

of main steel are used which is not very different from the

tendency obtained in tests of similar beams containing light

shear reinforcement. It is evident though that the influence

of the main steel is more preponderant in beams without

stirrups as in that case the ultimate loads are lower and

the percentage increase in terms of these low loads higher.

Such a conclusion is also reached by Borishanski(17) despite

the fact that, as he admits, his tests concerning this para-

meter were not reliable. This can also be indirectly supported

by plotting the nominal shear stress q at which an inclined

shear crack appeared against the ratio of main steel fig.

-

(41).. This is not a very exact measure as it is subject to

human judgement, but the values obtained can be considered

Q
as shear cracking stresses (qcr= bdr) since they do not
1
differ significantly from the ultimate nominal shear stresses

of similar beams without shear reinforcement that failed in

diagonal tension - see section 4.2 and tests by De Cossio and

Laupa on fig. (42).

161
 Turning briefly to the case of short beanis, tests of such
(78)
beams without shear reinforcement by Moody and morrow

and Viest(80) tend to indicate that the variation of the

ratio of main steel has no significant influence on the shear

strength at a given value of a/dl. Other investigators have

(reported
quite the opposite - Kani54)ý

In the case of short beams with stirrups morrotols(79)

tests fig. (42) are further substantiated by Clark's(25)

tests of beams Cl-13 and C4-1 which showed an 18% increase

in strength between rst values of 2.07% and 3.10%; the other

characteristics of these beams were a/d1=1.56, rqfyq=163

lbf/in2, Ucy1=3600 lbf/int.

Strain readings along the main steel, reported in section

4.9 show that the strains, although increasing with increasing

load, tend to become uniform in the region under shear cracks;

this lends support to the concept that shear is carried

partly by tied-arch action. Since this arch action is the

contribution of the concrete in supporting shear it seems

that when increased loads are withstood, only by virtue of

an increase in the ratio of main reinforcement, the higher

forces in the main steel imply " increased inclined thrusts.

Provided that all the other parameters remain constant, this

can only result from an increase in the concrete zone in-

volved at failure.

The photographs of fig. (40) indicate clearly the validity

of the above considerations.

As discussed previously the compatibility of deformations

must involve the length of the failure crack. A solution based

162
 (94))
on, such concepts (Regan would predict longer
slightly

cracks in the case of beams with a higher ratio of main rein-

forcement. The photographs of fig. 40 indicate that even if

this is not as clearly distinguishable as the ultimate neutral

axis depths' such a tendency dons exist. Furthermore, in that

solution the dependence of the shear strength on the ratio of

main reinforcement is expressed in terms of a reference

neutral axis factor no representing the flexural neutral axis

factor at failure, with the main steel remaining elastic. De-ý

fined in this manner no is an increasing function of the main

reinforcement ratio whose rate of increase diminishes at

higher values of rst. All other parameters remaining constant,

the "shear strength ratio of main stool" relationship would

be eXpected to coMform to this pattern. rig. (41) indicates

that this is actually so.

Another reason for expecting higher strengths in beams with

increased ratios of main roinforcomont comes from dowel action

considerations. The shear force that can be carried across a

shear crack by the main steel acting as a dowel can be expected

to be an increasing function of the area of main reinforcement

crossing the crack(111)(120). In the case of normal beams

(a/dl > 3) fig. (41) shows that the increase in shear strength

is represented by a steeper line in the case of beams with

higher degree of shear reinforcement. Viewed from an overall

increase in strength however, a 60% increase is obtained in the

case of rqfyq=83.5 lbf/in2 as compared to a 46% increase in

the case of rgfyq=l6? lbf/int. All other things being equal

this is an indication that the relative importance of dowel

163
action diminishes with increasing shear reinforcement; this

was also noted by Loonhardt(63) It follows that the approxi-

.

mation suggested in soction 4.3 for the dowel action in boams

with stirrups, is not likely to lead to overconsorvative

estimates.

A. 5 Influence of the shear spanto effectivo depth rr3tio

(a /d1-

Seven of the tests performed can be used to investigate

this parameter. They are divided into two groups, differing

basically in the amount of main steel present. A further

difference between the two groups i's the presence of heavy

comprossion steel ih the group having a higher ratio of main

stool. This was added in order to obtain shear failures at

higher. a/dl ratio. The relevant details of tho tests are

shown in table 4.5.1 and photographs of the beams appear in

fig. 43,

TABLE 4.5.1. Influonce of the shear span to effectivo dgpth ratio

BEAM U l r t Shear r f a/dl r Ult. q Failure

cy s g yq sc Load ult
No lbf/in % reift. 1bf/in % tons 1bf/in G1odo

R23 4370 1.46 7/3206" 83.5 2.24 0.34 19.8 345 IA/S

R8 3870 " of 3.36 " 16.0 279 S . C.

R22 4280 " if 11 4.48 " 16.0 279 S. C.

R26 5300 4.16 " " 2.4 2.63 30 560 IA/S

R25 4470 If 3.6 " 21 392 S

R24 4480 " If of 5.4 " 1865 345 S

13.6
112 4920 " " " 0.37 22 411 S. C.

164
 R23

'ý $ý
ßl1 rj
i2

ý' ý1 Ile K Ws
20 62 8le ýýI

a/d1_2-24 rst =1 46°/.

R22
`gyp
,-µ
ýr 1 \vi y1
'. tl

a/d1 =4.50 rs t1 46 °/

r
ý-
Res-- ý.

a
L°ý 12 112
ý/ 2elop+ ýI [[ Ir (8
- 14
4pß/fý.. ý'k
TZG

c/ d1 : 2.40 rst :4 16°/°

rsc : 2.61°/

(R25 ýi
-,,,$.-.v__
ýý
---ý_
I
ltý

o/d1 =3.60 rs t= 4- 16 °/°

r5C =2 . 61 °%

-'N

A-1
a/d1 -5'4 rst - 4.16°/o
rsc = 2.61°%

Fig. 43 Influence of shear span to effective depth ratio

:
rectangular beams failing in shear

t64Q
 ö ö
CO
N
0
C N :- CO
Is
E If 11
. i. ý 41
.4.6 (Ya

O
"y C Q
,L4j
yL
q

x
G
ýl .2 x
0
V N
J w..
CJ ýuCI

0 9 o0
i" "
W
tý- NC
+
a.
v
F1--
ý
ö ^
5Uw 1
p t (D ." 1
ý t r' ;"r U-)

y"
IG
n
'C}
n u
Y - ýt)
d

N
I v
L. O
LLT
in k 1:1 -4 ý1
O
N

Gi

.
e I. :i

d
H
U

N
CJ

/ t.
U
U
O ý`ý r
t1

)r-' ' 0'

00 Uo
ö-0 0
In %t; ßn0 ty)
sdi>1 ü u; 6uaa; s aaays alow! lin

Fig. (44) Influence of the shear span to effective depth ratio

ä
 In fig. 44 the ultimate shear strength is plotted against

this parameter not only for these tests but also for some

relevant tests on beams without shear reinforcement.

It can be seen that the influence of this parameter is

quite important as the strength increases rapidly in the

region of low a/dl ratios. Actually it can be observed that

each of the series appearing in fig. 44 can be divided into

two parts. One part where the influence of this ratio is li-

mited and a second part where it is predominant. The exact

boundary between these parts varies somewhat. For beams with-

out shear reinforcement fig. 44 shows that in the case of beams

(75)
tested by Mattock this boundary is lower (smaller a/dl

value) for beams with an increased ratio of main steel. This

tendency however was not confirmed by the results reported in

section 2 or by Kani's(54) tests - see section 7.8 as in

-

both cases the opposite tendency was observed (lower boundary

for lower rst). In the case of beiams with web reinforcement

the present tests confirm the tendency for a lower boundary

value in the case of a lower ratio of main steel. Thove tens

dencies can only be considered tentative, in view of the

great scatter of results in tests of short beams. A great

number of tests is required at various low a/dl values, with

variations of such parameters as the main steel reinforcement,

concrete strength, shear reinforcement, and compression rein-

forcoment in order to have a precise idea of the behaviour of

beams in that region, since not only does the boundary value

fluctuate but the rate of increase of strength also varies.

It can be noticed however that the value of a/d1=3.0 can

165
 be reasonably taken as the boundary in all cases. The oresont

investigation is mainly concerned with the shear strength of

the longer beams, more commonly in use, in the

- especially

case of T-beams - and 90% of the tests were carried out to

investigate the parameters affecting the shear strength of

such beams. In chapter 7 general solutions are presented for

the case of these beams. In view of the few tests carried out

on short beams only a tentative evaluation of their behaviour

can be undertaken.

This division of the investigation of beams into two parts

is justified by the fact that if the strength of a beam is

expressed in terms of a single function of a/dl for all

values of the latter, the correlation between tests and theory

will not be very satisfactory. Theories whose "Qce, re-

d"
1
lationship is expressed in the form of a curve with a small
(? 8),
radius of curvature (Laupa(61), Moody Clark(25)) will

tend to be oversafe in the region of high a/d 1 values, while

in the theories where this relationship is represented by a

curve having a large radius of curvature (Krefeld and Thurs-

ton(60)I ACI-ASCE(4)) the shear strength does not rise

sufficiently in the region of low a/d 1 ratios.

(94))
Theories with intermediate curvatures (Regan cannot

do justice to both regions.

Returning briefly to the examination of the beams tested,

it was observed that the Inclined Action failure of the short

beams (a/dl , 2.3) took the form of a shearing failure. This

was true not only in the case of beam R26 where heavy com-

pression reinforcement was present but also in the case of

166
beam R23 whose failure presented the aspect of shearing of

the compression zone above the inclined crack.

Beams R8 and R22 failed in shear-compression and had simi-

lar ultimate neutral axis depths and similar critical shear

crack lengths. Although more inclined cracks appeared in the

longer beam (R22) it was noticed that the width of what proved

to be the failure crack was greater than the width of the

other inclined cracks, at all load stages.

Beams R24 and R25 whose failure is better described by the

shearing mode displayed critical cracks whose horizontal pro-

jections were of the order of 1.5d1.

In general, shear cracking did not seem to be affected by

the a/dl parameter, taking place at the 10-11 ton load (total)

stage in the case of beams with rst=1.46% and at about the

15 ton load stage in the beams with rst=4.16.

4.6 Influence of the Concrete Strength

To study this parameter six beam] divided into two groups,

were tested. The first group had more shear and main rein-

forcement than the second group. All six beams were tested

under the same shear span and cLthin each group the only

variable was the concrete strength. Table 4.6.1 gives a sum-

mary of these tests and photographs of the beams appear on

fig. 45.

167
 Ru
r-. KM ýý" Äßi
ebP
.!.

if 9ýt
ucyl= lbt)Ulbt/in- rgfyq: 83 5 Ibf/in2

st'1 46°/°

R2o
}1 1e 'ý
/ý ýt M1 I> 2
eKNf
12
14
Is

It
it

Ucy1_ 6230 Ibf/in2

rgfyq: 83.5 Ibf/ßn2

rst=146°/

ýý R27

3ý tlýl

Ucyl =1980Ibf/in2 rqfyq = 167 Ibf / in`

St =416°/°

rý,
_J21, s=
ýY

ýKý ýA
it
ý'ro al

ii.
lc h 0ý rR /1 Bö6 A
J4. ýlAaým ýy
ýe ape
r/
iiv,. l ?'ýyef\m

Ucyl : 6980 Ibf/in2

rgfyq
t
167 Ibf lint
--
r, :4 16°! °

Fig. 45 Influence of concrete cylinder strength : rectangular

beams failing in shear

68
 TABLE 4.6.1 Influence of concrote strength

BEAM Ucyl rst shear rgfyq (a/dl) rsc Ulte quit Failure
Load
2 % 2 % 2
No lbf/in reift. lbf/in tons lbf/in Mode

R17 1850 1.461 7/321106" 83.5 3.361 0.34 14 244 S. C.

RB 3870 16 '279 S. G.

R20 6230 18 314 S. C.

R27 1980 4.161 7/32"Q3" 16? 3.6 X0.37 19 354.5 (1)

R13 4680 30 560 S. C.

; R21 6980 30 560 S. C.

(1) Probably a flexural compression failure although it was

a shear crack that ope.'ed wide giving the ultimate

rotation.

Before examining these results relevant to beams with

stirrups a word must be said concerning the effect of this

parameter on beams without shear reinforcement.

Moody(78) tested 16 beams whose concrete cylinder strength

varied from 1750 psi to 6000 psi to study this parameter.

The shear span to effective depth ratio was 3.41 and the ratio

of main reinforcement 1.89%. He concluded that the shear

strength of these beams increased with the cube root of the

cylinder strength of the concrete. A similar conclusion can

(118);
bo derived by plotting the relevant tests of Van Den Berg

in these the shear span to effective depth ratio was 3.50

and the ratio of main steel 4.32%. The 19 beams involved had

concrete cylinder strengths ranging from 2000 psi to 6500 psi.

In both of these cases the final shear strength was either

the same or slightly higher than the shear cracking load so

they can be considered as having failed in Diagonal Tension.

169
 Ramakrishnan(91) arrives at this relationship as well, and

also concludes from his tests that the effect of concrete

propetties was independent of the shear span to effective

depth ratio. Since the mode of failure in shorter beams

changes to one of shear-compression or shearing the cube root

relationship should be expected to apply also to the present

tests where compression failures were obtained. The cube root

relationship for shear-compression is also proposed by Leon-

hardt and Walther(66) but it appears to the writer that this

proposal was based on Moody's tests described previously. An

interesting comment on the subject was also made by Kani(55)

who observed that if the cylinder strength is ignored in a

formula but the other parameters are properly taken into

account, the majority of results will fall within a narrow

± 14%. Finally Taub
range of and Neville(112) state that the

effect of this variable is the same in beams with or without

shear reinforcement. However this statement seems to be

based only on tests of short beams.

In fig. 46 therefore, the test results from table 4.6.1

are plotted together with lines representing the variation of

the ultimate nominal shear stress 4ult with the cube root of

the concrete cylinder strength. It can be seen that the cube

root relationship can be applied to shear-compression failures

of beams with shear reinforcement.

The failure of beam R27 occurred at a rather low load and

is subject to caution. When the load was 84% of the ultimate,

the compressive strains under the load plate, one inch from

the top fibre had coached 35x10-3 fns/in and small cracks

170
 2/9
qu: K4(Ucyl)
600

N
C

4q-

Er
400
N

0
V)

O
V
300
:KXu1 2/9
(Uc;
q 3K 1ý
. wº u CY
f
Cc.

:.
a

D
200
KEY
2;
o rq fyc 83.5 lbflin rat 1.46 01*
"0q
fY4" 167 Ibflin2 ; rat = 4.15°! 0
100
K; constants

1000 2000 3000 4000 5000 6000 70,-,: )

0
concrete cylinder strength UGyl in Ibfj in2

Fig. (46) Influence of concrete strength

%!
appeared. A flexural failure was expected-but the flexural

cracks did not open when the load was increased and the ulti-

mate rotation was provided by the excessive widening of the

shear crack at failure. This crack had progressively reached

a crack width of about 0.015" at 95% of ultimate load.

Comparing the other five tests with the cube root relation-

ship it can be seen that a somewhat better correlation would

be achieved if the lines were slightly tilted in a clockwise

fashion. This can be accomplished through the use of a power

slightly smaller than 1/3 in the relationship

qult (u1)"
-k
The use of n=2/9 is illustrated in fig. 46 where it can be

seen that it does give a somewhat better correlation. As will

be seen in chapter 7a simplification of Regan's theory re-

sults in the shear strength being expressed in terms of this

power of the concrete cylinder strength. The correspondence

between theoretical and experimental relation is thus quite

satisfactory. The cube root term in the case of beams failing

in Diagonal tension can be explained in terms of the relation-

ship between the tensile and compressive strength of the con-

crete - see section 7.1.1. In the case of Shearing failures

the cube root term is also justified in Chapter 7, section

7.9,2 in terms of a Coulomb type failure criterion.

Returning briefly to the general behaviour of the beams

tested it can be seen from the photos that the reduced neutral

axis depths were smaller in the cases of beams with high

strength concrete. This of course should be so - in view of

the small increases in strength obtained in order for the

-

172
 horizontal internal equilibrium condition to be satisfied.

Furthermore it was noticed that the failure was more grad-

ual and ductile in the case of beams with low strength con-

crete and the whole depth of the intact concrete zone seemed

to have crushed simultaneously when failure occurred. On the

other hand the failure of beams with high strength concrete

was more brittle and at failure the concrete crushed suddenly

in most of the upper part of the intact compressive zone and

then affected the rest of that zone (see fig. 45 beam R21)..

This implies a better distribution of stresses at failure in

the compressive zone of beams with low strength concrete. It

would also seem that these almost uniform stresses were all

very close to the maximum compressive stress of the concrete.

A more unfavourable distribution of stresses can be expected

in the case of beams with high strength concrete.

As a result a more accurate prediction in strength can be

expected if a different stress block factor is used for

different concrete strengths. This question is examined in

more detail in appendix A section 1.

4; 7 Influence of Compression Reinforcement

To assess the influence of this parameter tests were carried

out on beams having increasing amounts of compression steel.

Three of the beams involved had a lower degree of shear rein-

forcement than the other two, but the other parameters were

kept constant in all cases. The relevant details of the tests

appear on table 4.7.1 and fig. 47 contains photos of the beams

173
 R12

?4K/ý s7 Zp
! J/ 16
fI /4 1x"
° Isý MwK26} K!

.t It
,5c: 0 37°I ° rgfyq - 83 5 Ibf / n2

R25

16KnýN uýýK
N® ýg M
16 16 ýý yt 2ý ý
ýýf I% 12ýG li
Iý 12

61 °/o 2
r_2 rqfyq: 835Ibf/in
sc

Rvs n
rm
YS1 20\
Kwl
K
ro(pt 1ý

ý It

ru: 0.377 rq fyq = 16 7lbf/in2

R15, ýl
ýf+
ýj t0 foýý M
to ub äý 16 ý y..
I

24

t 'lt
sc- 1 49% rgfy9 1671bf /ln2

R16 ýý j_'<tý !
ýfyö ^1
ßr7

lrIv r"DI

rsc = 298°/° rgfyq = 1671bf/in2

Fig 47 Influence of ratio of compression reinforcement

rectangular beams failing in shear

%74
 tested.

TABLE 4.7.1. Influence of the compression reinforcement

BEAM Ucyl rst rgfyq a/dl Compress. rsc Ult. qult Failure
Load
2 % 2
No lbf/in Steel (1) % tons lbf/in mode

R12 4920 4.16 83.5 3.6 203/8" 0.37 22 411 S. C.

R25 4470 " it of 2Jt1" 2.63 21 392 S

R13 4680 " 167 " 2J'3/8" 0.37 30 560 S. C.

R15 4330 " 213/4" 1.46 28 522 S/S. C.

R16 4580 " it " 413/4" 2.93 28 522 S

, R28 4580 " 326 " 261" 2.63 36 671 S. C.

(1) steel f =60 000 lbf/in2; All other f =40 000 lbf/in2
'forX=1" y

It is however realized that shear tests on beams having

increased amounts of compression steel are not always possible.

Such comparative tests cannot be made when the degree of shear

reinforcement and/or the a/dl ratio are high enough to ensure

flexural failure in the presence of light compression rein-

forcement. In these cases a shear failure is possible only

when sufficient compression steel is provided to ensure that

flexure failure is avoided. In view of this the result of

beam R28 is also included in table 4.7.1.

The relation between this parameter and shear strength is

shown in fig. 48 for the case of the comparative tests per-

formed. It can be seen that no significant variation in shear

strength occurred. Thus at first sight it could be considered

that shear strength is independent of the amount of compress-

ion steel present. Actually this question is not so simple

and the opinions of various investigators differ greatly on

175
700

600

x
rgfyq: 167

500
C

C
rf: 33.5
400

300 - 3 _
KEY
x Beams having rq fyý : 1371bfJin2
tests
0u II d : 835 it
200 Regan rq fya : 83.5 ibf/ in2
" .' -1
2 167 11
try
3 Laupa rgfyq : 83.5 II
4 I. II r 167 it
100 -

1°1o 2°1o 3°/0

Ratio of compression reinforcement rc" (016)

Fig. (48) Influence of compression reinforcement

ICE
this subject.

Bernander(14) and Al-Alusi(7) indicate that compression

reinforcement has no significant effect on the shear strength

of beams with high tensile reinforcement and no shear rein-

(112)
forcement. Taub and Neville consider that compression

reinforcement has no effect on the shear strength of beams

while at the same time admitting that it is possible that

the effectiveness of the compression reinforcement in re-

sisting shear depends on there being a sufficient number of

well anchored stirrups to restrain the steel from buckling.

A conclusion of no effect in connection with beams having

vertical stirrups is also derived from Larsson's(62) tests.

Laupa and Regan in their theories, described in chapter 29

suggest procedures for taking into account the contribution

of the compression steel to the shear strength of rectangular

beams failing in shear-compression. The tendencies predicted

when these theories are applied to the case of the beams

tested are illustrated in fig. 48t where it can be seen that

the predicted increases did not materialize. It is interesting

to point out here that an examination of the 186 tests on

which Laupa based his formula revealed that only thirtythree

of these beams contained compression reinforcement. On the

other hand five out of the six beams displaying abnormally

high values had compression reinforcement, and the two

Qclc
test
most unsafe results had an rsc value of 3.14%.
(52)
R. Jones performed tests on beams with inclined stir-

rups provided with an artificial inclined crack in the posi-

tion of the expected failure crack and evaluated the contribu-

177
 f

ap plied s hear
measured shear (tc tai)

shear t'. ken by stirrups

shear taken by the shear täi >n by the

compressio
main reinfo rcement ,n reinforcement
ýý
"..... ......: Shear take!n by the concrete
, ýI- .. .
O 10.8 21.6 32.4 43.2 54.0 64.8 75.6 86.4 97.2
load applied expressed as percentage of ultimate load (°l°)

Fig. (49) Contribution of the various elements of a beam to its

shear resistance - according to R. Jones52

t76
tion to the shear resistance of all elements, including the

compression reinforcement; fig. 49 illustrates his findings.

It can be seen that according to Jones a considerable pro-

portion of the applied shear is carried by the compression

reinforcement, which of course implies that the latter con-

tributes to the shear strength of a beam.

From Wilby, s(122) tests it can be concluded that when

heavy amounts of compression reinforcement are present the

shear strength of beams with very light or no shear rein-

forcement decreases with increasing area of compression steel.

If in the same beams small amounts of compression steel are

present a tendency to the opposite exists.

Another general observation concerning the contribution

of compression reinforcement is the fact that heavy amounts

of the latter are necessary if shear failures are to be ob-

tained in beams with higher degrees of shear reinforcement.

This is illustrated by the case of beam R28 of the present

series of tests as well as by Sunderland's beams having higher

rqfyq values sea fig. (39) section 4.3 Fig. (39)

- -. also

shows that in the region of lower degrees of shear reinforce-

meet, where comparative tests were possible, doubling the a-

mount of compression reinforcement did not seem to affect

the shear strength of the beams. Finally Moody(78) observed

that in simply supported beams without stirrups the influence

of compression reinforcement was negligible; on the other hand

in the case of beams with heavy shear reinforcement it was

found necessary to take account of this parameter in order

to obtain satisfactory agreement with test results.

179
 It can thus be seen that the effectiveness of the compress-

ion reinforcement seems to depend on the the

ability of shear

reinforcement to ensure flexural failure of the beam in the

absence of compression reinforcement. An empirical formula

can be developed which will give more importance to the com-

pression reinforcement in the presence of higher degrees of

shear reinforcement and will tend to neglect it or even attri-

bute to it a negative effect in the presence low

of amounts

of shear reinforcement. Such a formula would satisfy most of

the numerical tendencies of existing test results and the

possibility of negative effect would accord with Wilby's ob-

servations. Such a formula was actually developed and is given

below merely as an illustration of the effect of the com-

pression reinforcement, However, although numerically adequate

results are obtained, its application is not recommended. As

will be discussed subsequently, the mode of failure can change

in the presence of compression reinforcement and formulae de-

rived for one mode of failure cannot be rationally applied

when the mode of failure is different.

The empi'ical formula illustrating the effect of compress-

ion reinforcement is:

shear streu th of beam with compression reinforcement

bhQar strength of same beam without compression reinforcement
K+mrr 0i
(ý
sc - 0.2)

K
I(rstrn)2
whore K= + 2r m- rst m
st
10-
and m=5+
cyl

As has been shown quite conflicting in

views exist connect-

180
ion with this parameter. It is the writer's opinion that any

rational attempt to explain the influence or lack of influ-

ence of this parameter observed in tests must take into account

the actual mode of failure of the beams tested. The beams

with low ratios of compression steel - R12, R13 - failed in a

shear-compression manner, while beams R16 and R25 having

heavy compression reinforcement failed in the manrrr described

as 'shearing' in section 4.1. In the latter case the failure

shear cracks rose high almost reaching the tcp surface of the

beam before failure; at loads close to failure horizontal

cracks appeared at the level of the compression steel. This

phenomenon was also noticed by Wilby who subsequently ascribed

the decrease in shear strength to the high horizontal shearing

stress at the level of the compression steel, with consequent

rapid horizontal splitting of the concrete at that level and

release of bond. It was further noticed in the present tests

that when this cracking occurs there is a tendency for the

strains in the compression steel to stop increasing and even

(52) in the (50) (52)

to decrease - see fig. series to showing

the strains in the compression reinforcement of some of the

beams tested. This may be an indication of transfer of forces

from the stool back to the concrete. The length of the hori-

zontal projection of the critical crack was in both cases

approximately equal to one and a half times the effective

depth.

In the case of beam R15, which had an intermediate amount

of comprossion reinforcement the type of failure is not abso-

lutely clear. Beam R28, provided with heavy compression rein-

181
 F
3-

1.2 \compressior
-ý-T-rý- rei2 0 rc em2nt

Position of strain 1,2 3

- gauges and

compression
ent
top fibre

0 10 20 30

Fig. (50) Strains in beam R13

compression reinforcement and concrete.

192
 p

2,3 1,4-,
compression
`---
sk T- reinforcement
f2 4,2 2 0341

position of strain gauges 1,2,3 and 4

10 20 (20) 30
Total load in tons

Fig. (51) Strains in compression reinforcement beam R15

tG3
 P

1,23.4 \cornpression
-v-fib reinforcement
42 b- 4ý 40 4

position of strain gouges 1,2,3 and 4

Ö
r-
x
c

.9

5 10 15 20 25 (28) 30
Total load in tons

Fig. (52) Strains in compression reinforcement beam R16

is
 forcement but having a high degree of shear reinforcement as

well, failed in a manner associated with shear-compression

failures.

It sooms thus evident that the behaviour of all the beams

tested cannot be explained in terms of one failure mode and

this may explain the different views on tho effect this

of

parameter held by a number of investigators.

In chapter 7 equations are developed for shearing and

shear-compression modes of failure of beams with and without

compression reinforcement. When they are applied to the beams

reviewed in this section - chapter 8- the governing equation

in the case of beams R16 and R25 is the

clearly one based on

a shearing typo failure and relevant failure criterion. On

the other hand a shear-comprossion type failure is predicted

in the case of beams R13, R28, and R12 although in a loss

definite manner as far as R12 is concerned. Interestingly,

both equations predict the same shear strength at failure

for beam R15. These predictions and their correspondence to

the physical behaviour of thu beams are an indication that the

procedure developed in chapter 7 not only gives satisfactory

Qcalc
(average
strength predictions -------
Qtost = 91%, coefficient of

variation = 7% as will be seen) but also enables the exact

manner of failure to be reasonably predicted.

4.8 Influence of the inclination of stirrups

Tests were carried out on two beams having stirrups in-

clined at 45° to the direction of the main steel. These tests

185
 can be compared to those of two similar beams with vertical

stirrups. It should be pointed out that, the same stirrups

in the 45 °
and spacings were used corresponding and 90° cases,

the spacing being measured perpendicular to the direction of

the stirrups; thus if the degree of shear reinforcement is ex-

A
sip
pressed as rgfyq; where rq= b-S and s is the stirrup
vc
spacing in the direction of the main steel it can be seen

that the comparative beams had the same degree of shear rein-

forcement.

Table 4.8.1 lists the characteristics of tho beams com-

pared whose photographs appear on fig. 53. A theoretical

treatment of this question is presented in Appendix A section

3.

TABLE 4.8.1 INFLUENCE OF INCLINATION OF STIRRUPS

BEAM Ucy1 rst Shear rf a/d1 r Ult. q lt Failure

q yG2 sc u
2 reinf. Load
No lbf / i n % ( 1) lbf/in % n
tons lbf/i Mode

RB 3070 11,461 7/321106, HIS 3.36 0.34 16 279 S. C.

090°
R18 4540 " 7/32"06" to 17 296 S. C.
1
0045°
R9 4290 " 7 32113" 167 If to 21 366 S. C.
C)90°
R19 4390 " 1/41104" 11 11 it 24 419 S. C.
X45°

(1) spacing measured perpendicular to the direction of stir-

rugs.

Beam R18 failed in a classical shear-compression manner

for a beam having a low degree of shear reinforcement. At

about 80% of the failure load the main flexural cracks in the

region of high moments and the two inclined cracks on the side

on which failure took place had approximately the same width

186
 n R8
,:. ,'T
z

I` ýýýý
ý.
"

ti rqfyq = 83 5 Ibf/in2

R18 n

tA

oC . 450 rq fyq= 83 5Ibf/in2

r7
_Rg

rq fyq = 167 1bf / in2

Rig
e»
/1p ý J/ eN 8 tý
C' ýý X10 ýB
ýýiý J (B 20
ýLr

Y,
t4

oC -45° rqfyq _ 167 1bfhn2

Fig. 53 : Influence of inclination of stirrups

beams failing in shear
rectangular

187
of 0.006" to 0.00? ". When the load was increased to about 94%

of the failure load the flexural crack increased only slightly

to about 0.008" to 0.01011 while what proved to be the failure

crack opened to 0.02011 indicating the start of yield in the

stirrups crossing it.

The behaviour of beam R19 was not as clear. At about 83%

of the failure load the flexural cracks (0.012") were wider

than the shear cracks. When load was increased a new inclined

crack appeared, which proved to be the failure crack. It

started widening quicker than the flexural ones and attained

and exceeded their width at about 95% of the failure load.

Although the failure does look like shear-compression, the

fact that the shear crack, at whose head the failure took

place, did not rise above the flexural cracks and the fact

that failing compressive strains were measured in the con-

crete under the load pattern at about 95% of the ultimate

load, indicate that in actual fact the failure was a combin-

ation of both flexural and shear compression.

Table 4.8.1 indicates that the inclined stirrups improve

the load carrying capacity of the beams. This is in agreement

(78),
with the findings of other investigators (Moody G1onetto
(? 9),
Wilby(122)),

Besides the increase of load, it can be noticed that this

increase is more significant in the case of a high degree of

shear reinforcement (increases of 6% and 11% were obtained in

the cases of rqfyq values of 83.5 and 167 respectively). In

the theoretical solution developed in appendix A section 3 the

increase in shear strength due to the inclination of stirrups ,

188
is also an increasing function of the degree of shear rein-

forcement.

It can also be observed from the tosts that the failure

crack length is longer in the case of beams with inclined

stirrups. In absolute terms this difference in length is more

important in the case of beams having a low degree of shear

reinforcement. This is also implied in the solution of

eqn. A. 3.1 Appendix A section 3.

4.9 Strain measurements

4.9.1 Strains in the shear reinforcement

The stress in a stirrup is at a maximum within the width

of the shear crack and is reduced to zero within the concrete

by bond, end anchorage, or a combination of the two. Since

only one gauge was attached to each stirrup no data was ob-

tained on the distribution of strain along individual stir-

rups. However since the gauges were placed at different

levels, along an inclined line from the support to the sides

of the load platten, it is believed that the recorded strains

were close to the maxima.

The load-strain curves for beams R12 and R14 are presented

as fig. 54 and 55. The position of the stirrups in respect to

the crack pattern of the beams can be obtained from figure 37.

In both cases the strains on the stirrups were negligible

until the appearance of an inclined crack. Application of

load following shear cracking produced rapid increase of

strains. In the case of beam R14 the two stirrups on which

189
 ýýF
ý_
ýý
t- 1
feilure side
dP
,; f
ýi

8 .I
9.
1

ý-
A
710
position of strain-gauges
-.
`T

i
r

to
r-

N
C
O

U
0

12
O
6-0
O
H

ALo

0)

I3 01
 failure side

position of strain gauges

OD
r)

n
r

0
.-

Lt)

Fig (55) Strains in the stirrups of beam R14 (cf table 4.3.1)

%5t
 gauges had been placed were near the middle of the'shear span

and reached yield strain at failure. In the case of beam R12

gauges were placed in all stirrups. It can be seen from fig.

37 and 55 that stirrups 8,9, and 10 were crossed by the

failure crack. Stirrup 8 did yield but after 20 tons the

gauges in stirrup 9 and 10 were destroyed. However judging

from the rate of increase of strain in these stirrups at pre-

vious load stages it can be reasonably assumed that they

would have reached yield strain at failure. Stirrup 7, beyond

the head of the failure crack did not reach yield, and the

stirrup at the middle of the beam remained under compression

throughout the test. It can thus be reasonably that

assumed

all stirrups crossing a critical crack will yield prior to

shear failure involving the opening of the crack.

The behaviour of the gauge on stirrup 11 is also quite

interesting. Lying outside the failure crack, near the support

about 111 above the top of the upper layer of the main steel,

it registered quite high strain giving support to the idea

that the dowel action can be replaced by taking into account

the contribution of one more stirrup than those strictly

crossing the failure crack.

4.9.2 Strains in the main reinforcement

Strain gauges were placed 0 911 intervals on one of the bars

constituting the upper of the two layers of main reinforcement

of beam R12. (Ast=8j16mm). To avoid errors due to the bar's

own bending two gauges were used for each position, placed on
the top and bottom of the bar. The resultant load strain

curves appear in fig. 56 where the horizontal line represent-

192
 0 C "U
N
CD ut/9 pý x U! w su! DJ; s

Figs. (56) Load strains curves a deformation profiles of main reinforcement

beam R 12 (of tobe 4.3.1) P 22 tons
ult
a

ýQý
ing the longitudinal profile of the main reinforcement is

also the axis of zero strain.

It can be seen that up to loads between 12 tons and 16

tons, the profile was approximately linear, with the strains

increasing from the support towards the point of application

of load. This pattern however changed, following shear crack-

ing, just as the behaviour of stirrups changed upon attain-

ment of that load. The strain profile ceased to be linear as

the deformations exhibited a tendency to become uniform along

the length of the main shear cracks, before decreasing toward

the supports.

4.9.3 Concrete strains

Concrete strains were measured by demec points along the

length of the top fibre in all the beams tested. In some

cases strains were also measured along the side of the beam.

Typical top fihra ocraln profiles ara drown in fig. (57),

(59) to beams R13, Ra, and R16. In

'(58). ý, i corresponding

general top fibre strain profiles showed that the appearance

of a diagonal crack and its penetration into the compression

zone produced the following changes. The strain above the

diagonal crack seemed to increase at a higher rate; this was

more evident at higher loads and at the of the beam where

end

the development of the shear crack was more pronounced. At

the same time there seemed a tendency for all the compressive

strains to become concentrated over the shear crack as at in-

creasing loads losses in compression took place in the region

outside the shear crack. These losses in compression which at

first were registered near the supports startod occurring,

194
 3500--

3000-

2500-

2000-

28t:

1500 /24t

1000 Arýw
ýý.

500- torte f\ ti

Fig-(57) Top fibre strain profiles - beam R13 (of table 4.7.1. and fig. 47)
Ast= 80 16mm rgfyq=167Ibf1in Pult: 30 tons

l9F
3000-

2501

200(

150c

100(

50

Fig. (50) Top fibre strain profiles - beam R8 (of. table 4.3.1. and fig 37)
Ast :30 16mm rq fyq : 835 Ibf 1 in Pint 16 tons
--

% G.
 Y

26t/

ý20t ý..
:

ý`"-`

Fig. (59) Top fibre strain profi'zs - beam R16 (cf table 4.7.1. and fig 47)
Ast 80 16mm ; Asc 2 01" 167 lbf /in2 pult 28tons
-- rfyq : --

t:,:
with increasing loads and-crack deyeiopment, at sections

situated at decreasing distances from the point of application

of load. Measurements could not be taken at the failure load

exactly but it-is expected that the tendencies described

above were more pronounced at failure, than would be indicated

by figs. (57) to (59).

The question of failure strains can also be discussed on

the basis of these graphs. In the case of beam R13 having

intermediate shear reinforcement it is obvious that strains

in excess of 0.0035 in/in were reached at failure. The beam

failed clearly in shear-compression and a shear-compression

failure is definitely predicted by the equations of chapter 7.

(- the load necessary to cause shear-compression failure is

about 10% less than the load required to produce shearing

failure -). On the other hand beam R16 containing the same

amount of longitudinal and transverse reinforcement as beam

R13 but having in addition heavy compression steal reached

strains of the order of 0.0019 at 93% of the ultimate load.

It would seem then that failure did not occur by shear-com-

pression as this would require the attainment - albeit at

lower loads, due to the rise of the shear crack above the

flexural ones - of flexural compression failure strains of

the order of 0.0035. Ramakrishnan(91) measured failure

strains of 0.0038 at shear-compression failures. As has been

described in section 4.7 the actual mode of failure of beam

R16 was not one of shear-compression. but what is being re-

ferred to in this investigation as shearing failure. Equations

presented in chapter 7 predict this type of failure in the

198
case of beam R16. (The load necessary to produce shearing

failure is about 10% less than the load required to produce

a shear compression failure). The behaviour of beam R8 seems

to be intermediate between the behaviour of the other two beams.

Higher strains were reached in this case than in the case of

beam R16 - 0,0022 at 87% of ultimate instead of 0.0019 at 93%

of ultimate - but they were not as high as those of beam R13.

The actual mode of failure of the beam itself is not as clearly

defined as in the case of the other two beams and although it

has been described as shear-compression in the presentation

of results it is probably a border case between the two modes

of failure. It is not known whethop the increase in strain

brought about when the load was increased to failure enabled

strains to reach shear-compression values. The tendency indi-

cated by the graph gives, through: interpolation, strain

,
values at failure of the order of 0.0031 in. This of course

is smaller than the generally accepted CEB(27) value of

0.0035 or the value obtained by the formula proposed by

(45) U
£_0.004- c---
Yl
Hognestad, Hanson and Mc Henry --"--.
cc 6.5x106
which would in the present case give a failing strain value

of 0.0034. On the other hand Walther adopts a value of 0.0003

in a shear-compression equation. Variations of this order

make little difference to the calculated failure load i. e.

either Regan's or Walther's ultimate load equation. It would

seem that the beam must have been near the boundary between

shear-compression and shearing failures. The equations deve-

loped in chapter 7 predict in this case the same ultimate

load for either a shear compression or a shearing failure.

199
 The strain profiles of all beams tested showed behaviour

similar to the typical ones described; although as expected,

larger and smaller strains were registered at equivalent loads

in the case of beams with low and high strength concretes

respectively.

Figures 60 to 63 show examples of the strain distribution

at sections through the centreline of the beams obtained by

demec points. As expected the neutral axis depth decreases

with increase in load in all cases with one significant ex-

ception, that of beam R25 having light stirrups and heavy

compression steel. At the last load stage the neutral axis

increased. A possible explanation of this is the transfer of

compressive stresses from the compression steel to the con-

crete as discussed in section 4.7.

Figs. 63 and 64 (the latter showing the concrete strain

distributions at sections other than the centre line-in beam

R28) give an indication of slope of the main compressive

force in the concrete. From fig. 63 it can be seen that in

the section situated at the head of the shear crack the top

fibre remains the most compressed at all load stages. At

section AA (fig, 64) situated just outside the shear crack

the top fibre strain ceases to be the most compressive strain

at higher loads indicating a slight inclination of the main

compressive force. The inclination could not be expected to

be very pronounced as the beam has a high degree of shear re-

inforcement, and as shown in Appendix A section 2, the higher

the degree of shear reinforcement the smaller is the inclina-

tion of the main thrust in the concrete block above the shear

200
 D
0
0
n0 cJ
CD

N N
E
0
E
t
ö ä
V .
r v
C
° ü

cO
U
i-+
0 V1
41 O
C '- N rÖ

L
r' U U
}

U
C
U
C
lfO

e.,
C
0 r
"c
O
1O Uo
c
C 0

G ° :3
i
C.
,N
p A N v_7
.IV
d N
C
_
aC
OL n
N C

c
O
Ua

p
ý G UI Oa
ý V
p
S O
4p U
V

rn
0 LL
lu
V1 O

N 0
C 0
It)

O "' N M qo*
(uff) DOoj ns do3 wo, i; aouI; sip (ui) 3oojuns dot wjJ aoua; sip

i
 8

N LO
N
m

0
O
c o
0 Üö
x.n Cl)
G
O ca is
L
v'1 41
C L. 0N
cC4 O ö' 0 öM cL ^
Co U LQ Uö
N
t-
D. "N
N
jd

IR

a' C_
O

to
Lv

c co
C3

C
r_.
'O
C ýO
a 10-
v x-
ýý c 0
X
L Ný G
a
c N Q,
'17 C
N
C
c
ö
Nd

ýý
\ý.
to
U
L
d
U
UcoU O 0,
O
U

X(S)..

C
O
u
Et rº ý

t111
Q "- N co c L( 0N co c to

-(uff) aoDJJnsdol wo.JJ aouoisIQ (uff) aoojans dol woaj. aouolsIo

202
 1
0 TI
'
tcc'%nO
C 4j
I 411 41 41

C
+. c' to q. CO,
cv
041i : cam

V
uvý 1
L
O
N

ß
O
M
2d' [
F Yr-p
el r"
O
L AB
w

d
u
c3
U jA
eJ IBZ;
t.
/ scction. A-A Puh`: 36tons
0 1

100 300 500 700

o
concrete strains in in x 106/ in

0 0.. 6% cý
C .
'1
G

U
"1/
v
'4-
L

0
+0
E
U
L

U3 section B-B
N \a
LD

0 100 200 300

concrete strains in inx1CS61
in

Fig. (64) Concrete strain distribution at various vertical sections - beam R28

2C
 C
lO
xc

le
U
L

"J
v+ý
pr
L.
U
G
0
U

Fig (65a) Strain profiles at the level of 1" below top surface

'0
100
-
K
C_
"C
IISt
. ý, Q ^\
`
'0 Soo - 12t
ý/

' 'irons\ý `ýý

i ;! ."
.""
p2ptý.
ü0
.. r...

Fig. (65b) Top fibre strain profiles

Fig. (65) Concrete strain profiles in a short beam

beam R23 (of section 4.5) a/d1 : 2.24
rgfyq 83.5 = ibf/in2 pult : 20 tons

ZO1H,
 crack. A very definite proof of its inclination is obviously

obtained by the distribution of strain at section B-B. A

further indication of an even more pronounced inclination of

the compression thrust is provided by the case of a short

beam (R23) shown in fig. 65. In it the strain profilos are

shown not only for the top fibre but also for a longitudinal

section 1" below the top surface. It can be seen that up to

the level of the first load stage after shear cracking the

top fibre strains were the higher. At the following load

stage however the situation changed; although the main shear

crack passed through the demec points and prevented measure-

ments being taken near the centre, it is nonetheless evident

that throughout most of the shear span the higher compressive

strains occurred at the lower level.

Similar indication3 concerning the position of the neutral

axis in the shear span can be obtained from tests reported by

(60)
Watstein and mathey(120), Krefeld & Thurston and Rasch

et al(105), Watstein and Mathey's test was of a point loaded

beam without shear reinforcement having an a/d1 ralo of 2.0.

Krefeld and Thurston's tests woro of uniformly loaded beams

d
without shear reinforcement with ratios of about 14. Rösch
"1
Haugli and Mayer's tests concerned a Lniformly loaded beam
1-
having shear reinforcement and an ratio of about 8. In view
1
of these tests and those reported here and in section 2 the

following can be cancluded concerning this question.

The main compressive force in the concrete above the cri-

tical inclined crack slopes downwards from the section at the

head of the crack in the general direction of the support.

205
Away from the head of the shear crack the position of the

neutral axis in the shear span is not precise but depends on

the crack condition and propagation. In general the neutral

axis has a tendency to move down towards the bottom of the

beam with increase in load and cracking. As a result tho

distance between the neutral axis and the top surface of the

beam is much bigger in a shear span than in an equivalent

span subject to bending only.

206
 CHAPTER 5

INVESTIGATION OF SHEAR FAILURES IN T-BEAMS

The results of 38 tests carried out in order to investigate

the parameters influencing the shear failure of T-beams are

presented and discussed in this chapter. All but three of these

tests were on beams containing shear reinforcement, and with

the exception of one intentional flexural failure, shear failures

were obtained in all cases.

In the opening section tho general behaviour of T-beams

failing in shear and their modes of failure are presented. The

five sections that follow are devoted to the investigation of

the influence of the parameters affecting the shear strength of

any reinforced concrete beam: Degree of shear reinforcement,

ratio of longitudinal reinforcement, shear span to effective

depth tatio, concrete strength and inclination of stirrups. The

next three sections deal with parameters intrinsic to T-beams:

Flange reinforcement, Flange breadth and Flange depth. In the

final section, the deformation of concrete and reinforcement

measured during the tests are presented and discussed.

5.1 Behaviour of T beamsfailinin shear

The behaviour of T-beams does not differ from that of rectang-

ular beams in the early load stages and the first inclined cracks

appear in the web at approximately the same loads as in equival-

ent rectangular beams. In the absence of shoar, roinforcament,

if the shear span to effective depth ratio is greater than three,

a sudden failure takes place when this load is reached or

207
 Lý
_

R -:.
_
_^. v`

;<

Tig cm
(
---o-O r.
atý: _.

Fig. 66 Propagation failure crack in T- beams

of a

08
 T4
,- -- ---. -f

0 T26

ýrý ...., -. _ _.. _ ,. ýý

20 20
4 ~*p
Jý ýr

oý
4.f* . ,PF
sk.
ý'( '±7 f
'ý
..

-
It Iji

ý' 7
1 -,

Fig 67 Aspects of shear failure in T- beams

-

209
 :.ý ý4
-
.ý ý .., ý.. '.
ss:. jý
dý C'RdA[ tom iii^'ý ..
.
/n) a.
.

_ :._ i

Fig 68 Appearance of a flatter critical shear crack at a

later load stage beam T8
-

fib 1R
,.

i ý6-ýa . ýýiý

Fig. 69 Detachment of flange in T- beams

beam T21

Fig. 70 : Flexural failure of aT- beam

Fig 71 : Failure beom shear reinforcement (a/dß_336)

of aT- without

MIIn
slightly exceeded. This is analogous to the diagonal tension

failure of rectangular beams. Beam T2 in fig. (71) provides an

illustrative example of this type of failure. The first shear

crack appeared at 9 tons and extended up to the bottom of the

flange. At 11 tons the crack widened suddenly, entered the

flange"and extended to the top surface. The corresponding rect-

angular beam R-7 failed in the same manner and at the same load.

In the case of short beams without web reinforcement failure

does not take place upon shear cracking, and, in a manner simi-

lar to rectangular beams, increased loads are withstood. The

mode of failure is not very different from that of a similar

T-beam with shear reinforcement although collapse takes place

at lower loads and will be discussed when reviewing the latter.

When shear reinforcement is used in beams having shear span

to effective depth ratios greater than 3, failure does not

take place upon shear cracking but at increased loads. Since

almost nine tenths of the T-beams tested fall in this category

their behaviour will now be described in some detail.

Following shear cracking the shear crack increases in width

and extends from the level of the steel to the web-flange

junction. There is usually more than one shear crack in a

given shear span, and generally two or mores depending on the

length of the spän, will widen more than the others at in-

creasing loads) without however differing significantly in

width from each other. (See figures 72 and 73). Furthermore

the widths of these cracks do not vary significantly along the

major portions of their lengths although they peter out at the

two ends (see figures 73 and 74).

211
 ca k

Lr LO LO

ONQ t0 t0 ce)
N

co U

ý0 ä
q- N
x
E
bN 0

N tIi iO co to C)
.Obi
r- . .
C
®
01 +' co L
V/ V1 y
a
tr) LO to ,C
ra N
.ý
a CV (') N U Iý
to E
O ý
c . o ý
0 (B x
L L
T CJ1
Q N ;
N NN" V V C
s
r II U pý
ý_
L E
N C
o ® U)
lo
C*) to al tf3 t! ) j
ý' v
E L U
E N
(D tf) LO C V
ý
o C1i.
Q to N CO tb co 0 E
O U C
++ +' E
Ü
d O
.C
d
.o C 3
ýO N 0 ® . - to to dN d
W N
ý
0 N
d
N ý7 V [t N
ýy a Ü N V
o1 N Lo fi ö
O Ný'ý Q.
Q t4 N to to L) Lf)
Ü
v
E
°
o
c

M Q
.. d
L] 0

C 4' 0
O Q' 4.3
ý' 0
O
41 O s 0 ä
tom-
©
u rN

U) LO U) or) d

O
J
C'In U-)
to Z
I 6!
IE I (U! ) ? U, 'H

Fig. (72) Widths of major cracks in beam T-4

I-
. 07
 IP

00
25 7M7

A-

KEY
20
1,2,3 : F1Zxural cracks in the central region.
M
1O
4,5 ? Major inclined cracks in one shearspan
r- ^//
X
C

15
,c
4J 5
ti

U
O

1O U
E

_E
X,
O
E
51-
2
....
T

12 16 20 24 28 32
Load (P) in tons

Note. Crack 5Q appeared at a later load stage and was more inclinad
than the othzr existing shear cracks. Failure took place when this
crack entered the flange at 41 tons

Fig. (73) Crack devc,iopment in beam T-6

P: 14 tons (81°/s of ultimate)

floh
I
20 I
22 ýc,
ýo
'II 22

Fig. (74) Width of critical shear crack along its length-

beam T-10 (in inx1Ö 3)

tiZ
 The flexural cracks in all cases remain much smaller than the

shear cracks (see figs.? 3 and74). Although at increasing loads

flexural cracks rise into the flange in the area of high bending

moments the shear cracks do not enter the flange but they may

extend along the web-flange junction. Failure seems to occur

when at a certain load one of the major shear cracks enters

the flange and travels in the direction of the loading platten.

On entering the flange the critical shear crack shows a tendency

to widen in a near vortical manner - see photo of beam T4 in

fig. (67).

This propagation of the crack into the flange does not al-

ways start at the point at which the critical shear crack

first reaches the web-flange juni tion although sometimes it

,

may do so - see beam T34 in fig. (66). In other cases the entry

into the flange may originate at some point along the already

mentioned extension of the crack in the web-flange junction

(see beam T19 in fig. (66)).

The entry of the critical shear crack into the flange seems

to result in a failure taking place along an inclined plane ex-

tending from the web-flange junction at the point of entry to

the top surface of the beam in the vicinity of the load platten.

The extension of this failure plane into the flange is not al-

ways the same; sometimes it reaches the edge of the flange -

fig. (66) and beam T4 of fig. (67) but in other cases a more
-

limited region is involved in the failure - beam T26 in fig.

(67) then takes the form of punching under the load

- which

platten beam T7 in fig. (67).

-
Inspite this difference in the appearance of the flange
of

14
A.
at failure it would appear that failure still takes place by the

sudden extension of the critical shear crack into the flange.

The fact that it is not always the shear crack closest to the

centre of the beam that proves to (Ea the failure crack - see

beam T19 fig. (66) indicates that the compression in the

-

flange does not play a determinant role. Consequently failure

is dependent on the conditio-wg governing the entry of the crack

into the flange. Furthermore the more or less sudden failure

along an inclined plane would indicate a failure of a tensile

nature rather than a compressive one. Not only is it not always the

inclined crack closest to the load that proves to be the failure

one but similarly it is not always the inclined cracks formed

at the first shear cracking that become critical. If these

original cracks are not sufficiently inclined new flatter

cracks are formed at later load stages. It is then the extension

of those later cracks into the flange, upon further increases

in load, that produces failure. Beam T-8 (fig. 68) and T-6

(fig. 73) provide examples of this behaviour.

It was noticed during the tests on beams with vertical

stirrups that if the failure crack is idealised as a straight

line extending from the level of the main reinforcement to

the point at which the critical crack enters the flange, then

the inclination of that line is approximately constant and

to tan J. In the present beams with 3" flanges this

equal arc

implies a horizontal projection of the crack of about 1.5d1.

When stirrups inclined at 450 were used, the crack lengths were

much longer, their horizontal projection being of the order of

2.2 dl.

215
 Strain measurements on the stirrups (see section 5.10) and

the widths of the critical shear cracks indicate that all

stirrups crossing the failure crack yielded at failure. Although

some of them, especially those situated in the central part of

the crack, may have yielded at earlier load stages, failure

seems to occur only when yield has been reached in all the

stirrups affected by the critical crack.

The physical diffezonce between the shear and flexural

failures of T-beams is illustrated by an examplo of flexural

failure in fig. 70,

If no transverse reinforcement is present in the flange then

the flange is detached from the web at a certain load stage.

This detachment of the flange can of course take place only if

the shear strength of the beam exceeds that load (see fig. 69).

In the case of short beams (a/dl 4 3) the failure is more of

the punching type; the crack lengths tend to be shorter and

their inclination steeper than in beams with longer spans.

5.2 Influence of decree of shear reinforcement

A number of the beams tested can be used to investigate the

influence of this parameter: The pertinent details of these

beams are given in table 5.2.1 and typical photographs appear

in fig. 75" In fig. 76 the results are plotted for all cases

in the form of a graph representing the relationship between

the degree of shear reinforcement rgfyq and the ultimate

nominal shear stress galt. Both these quantities and the ratio

of main steel appearing in table 1 are expressed in terms of

the web breadth b

r

216
 T,1 MM

q fyq :0 rst c1 46°/°

T3
ýý

rq =83 5 Ibf/ in2 rst .1 46 °/°

Ts

tF

rgtyq= 167 Ibf/in2 rst .1 46 °%

TQ

rq fyq= 83.5 Ibt/ in? rst: 4 16 °/°

T6

rgfyq=326Ibf/in2 rs,ti :4 16 °/°

Fig 75 Influence of degree of shear reinforcement

T- beams failing in shear

2%7
 0
800

0 50 100 150 200 250 300

degree of shecr reinforcement in Ibf / in2

Fig. (76) lnflu2rcc of the cl: grec of s`; 2cr reinforcement

T- beams failing in shear

zle
 TABLE 5.2.1 Influence of Shear Reinforcement
iUoyl
Beam main 100Ast Shear rqfyq a/dl Ult. qult
Load
brd1
No lbf/in Steel Reinf. lbf/in tons lbf/in

T-2 406? 3016mm 1.46 3.36 11 192

- -

T-10 4090 " of 7/3209" 55.7 " 17.4 303

T-3 3990 " " 7/32@6" 83.5 " 210 366

T-5 4890 " " 7/3203" 167 " 28 488

T-18 4120 8Jt16mm 4.16 3.6 15 280

-
T-8 4530 " if 7/3206" 83,5 " 25 466

T-9 2930 " if 1/404" 167 " 31 578

T-6 3740 " " 3/8©42" 326 " 41 765

T-32 4000 4315 813

T-37 4615 42 784

T-16 4740 " " 7/32Q39" 55.7 7.2 18 336

T-15 4810 7/32C611 83.5 " 21 392

T-17 4790 " " 7/3203" 167 " 27 503

T.. 19 4340 7/32@6" 83.5 5.4 22.5 420

T-20 4655 " " 7/32©3" 167 " 30.6 572

T-28 4617 2.4 24 448

- .-

Tu29 5349 " If 7/3206" 83.5 " 31 578

4460 " " 7/3203" 167 " 39 728

I1-23
A
rqg

Qult
qult brdI
100A
at
ratý' bd

Examining first the shear resistance of the T-beam without

reinforcement it can be soon that, with the exception of the

219
 beams with small ('; 3.0) a/d1 ratios, they failed in the same

manner and at about tha same load as the equivalent rectangular

beams. Beam T-2 failed at 11 tons and R-27 at 10.95 and beam

T18 at 15 tons which is also approximately the load at which

inclined cracking appeared in equivalent rectangular beams.

The shorter beam T28 also developed shear cracking at about

the same load as the longer rectangular and T-beams having the

same concrete quality and ratio of main steel, but, in a manner

similar to the behaviour of short rectangular beams, it did

not fail upon shear cracking but only after the applied load

was substantially increased.

Even in the case of beams with stirrups, shear cracking in

general was not different from that of rectangular beams. It

occurred at approximately the same loads and was influenced

by the same parameters. Shear cracks in beams with rstz1.46%

occurred at a total applied load of about 10-12 tons while in

beams with rst=4.16% shear cracking took place at about 15-16

tons total load.

As far as thu ultimate shear strength is concerned it can be

seen from fig. 76 that it is an increasing function of the

degree of shear reinforcement. Furthermore the rate of increase

J,.
seems to be approximately constant for all the shear span to

effective depth ratios and main steel ratios considered. This

implies that the shear force carried by the shear reinforcement

is directly proportional to the degree of shear reinforcement.

If therefore it is assumed that the stirrups crossing the failure

shear crack yield at failure, then in order for the shear force

to remain almost directly proportional to the degree shear

of

220
reinforcement the length of the critical shear crack must re-

main approximately the same in all cases.

To ascertain the behaviour of shear reinforcement the

strains of the stirrups were measured in a number of beams

tested. These measurements are presented in section 5.10.1 of

this chapter where it can be seen that the stirrups crossing

the failure crack did in fact yield prior to failure which, in

this case, occurred with the pronounced extension of the shear

crack into the flange. Yielding of the stirrups is also ro-

ported by Ramakrishnan91 who states that T-beams failed after

the stirrups had yielded fully. One might be sceptical about

these tests because of the small size of the beams used

(br=3"9 d1=4.8) but yield strains in beams similar to the

present ones were also reached at about 93% of the ultimate

loads in the cases of beams ET2 and ET3 reported by Watther

(64)(67)
and Leonhardt. These beams had shear reinforcement of

rgfyq = 166 lbf/in2 and rgfyq = 249 lbf/in2 respectively. Yield

strains were also reported by these same authors in some later

(68)
tests on T-beams in the case of beams TA3,11f 15p and 16

having an rqfyq value of 355 lbf/in2 and TA4 and TA12 having
t
an rgfyq value of 219 lbf/in

The discussion on cracking that follows presently is re-

stricted to the first two series of tosts at a/d1 of 3.36 and

3.6. The question of cracking at longer spans will be discussed

in the section devoted to the influence of the shear span to

effective depth ratio.

Following shear cracking' the inclined cracks reach the

level of the flange but they do not go into the flange to any

221
appreciable extent. With subsequent increases in load the

shear cracks widen but it is not always the same crack that

widens most. Sometimes two cracks on the same shear span behave

in almost the same manner throughout the test and only at

failure does one of them widen excessively and enter into the

flange. In many cases before entering the flange, the critical

crack progresses horizontally along the web/flange junction

and then enters the flange either at the end of this extension

or at the point at which it first reached the flange. In eva-

luating the length of the crack it is the point at which the

critical crack enters the flange and produces failure that is

important. Its position is not always obvious from the photo-

graphs; sometimes the horizontal progression of the crack

mentioned above takes place at loads close to failure. At

failure destruction spreads over some distance and substantial

pieces of concrete in the under flange become detached. Thus

pieces of concrete had to be removed on occasion in order to

ascertain more accurately the horizontal projection c of the

length of the critical crack from the level of the main rein-

forcement up to the point at which it enters the flange. These

crack lengths are listed below where it can be seen that they

do not vary greatly, and consequently neither does the slope

of an idealised straight line critical crack.

Beam ic d1-t tan L4 Beam c dl-t tan ý-i

in
No in No in in where is the

T-10 15 7.7 0.51 T8 16 7.0 0.48 inclination of

T-3 15 7.7 0.51 T9 15 .7.0 0.47 the idealised

T-5 13 7.7 0.59 T6 14 7.0 0.50 critical crack

222
 These small variations in the crack length tond thus to sub-

stantiate the findings concerning the almost constant rate of

increase of shear strength mentioned earlier.

As in the case of rectangular beams, the development of

dowel cracking is hindered by the presence of higher degrees

of shear reinforcement but it was also observed that at inter-

mediate and low degrees of shear reinforcement the extent of

the dowel crack was more pronounced when heavy main reinforce-

ment was present.

5.3 Influence of ratio of main reinforcement

Tests on five beams having the same shear span and degree

of shear reinforcement were carried out in order to investigate

the influence of this parameter. The pertinent details of

these tests are given in table 5.3.1 and photographs of the

relevant beams appear in fig. 77. Furthermore table 5.3.1 also

contains some additional test results of beams which) although

tested for the investigation of other parameters, Can still

provide some indication concerning the influence of the tensile

reinforcement.

These results are plotted in fig. 78 in the form of a graph

showing the relation between the ultimate shear strength and

the ratio of main reinforcement.

It can be seen from that graph that in the case of beams

with longer spans, the increase in shear strength due to the

main tensile reinforcement is not very significant. On the

other hand this parameter seems to have an important effect

where short beams are concerned.

223
 TI

rst 2 1.25%

Ts
m

rst = 1.46 °/e

rst, 1.95%

rst = 3.0°%

st = 4.16°l.

T- beams
Fig. 77 : influence of ratio of main reinforcement
failing in shear

224
e\tJ
d'

a,
0.
w? '
J-r M

ö
y
c

E
cJ U
L
O
C

C_

\; U
E

W
O

0
L

000

Fig. (78) Influence of ratio of main reinforcement

T. -beam failing in shear
a

225
TABLE 5.3.1 Influence of Tensile Reinforcement

Beam Ucyl Main dl 100ASt Shear rgfyq a/d1 Ult. qult

load
brd1
No lbf/in Steel in reinf. lbf/in tons lbf/in2
216mm
Tl 4050 10.7 1.245 7/3206" 83.5 3.36 22.1 385
+i012mm
T3 3990 3016mm " 1.46 21 366

T4 4710 4016mm " 1.95 22 383

T7 3970 6016mm 10.4 3.00 3.46 22 396

TO 4530 8016mm 10.0 4.16 " " 3.6 25 466

T5 4890 3%16mm 10.7 1.46 7/3203" 167 3.36 28 488

T9 2930 8j16mm 10.0 4.16 1/494" If 3.6 31 578

T30 4160 3016mm 10.7 1.46 7/3206" 83.5 2.24 22 384

!
T29 5349 8116mm 10 4.16 " " 2.4 31 578.5

Examining the five tests carried out specifically to investi-

gate the effect of this parameter, it was noted that the small

increase in shear strength du© to an increased main reinforce-

ment did not take place until eight bars were used as main re-

inforcement. This indicates that this small increase may be

due primarily to the increased shear carried by dowel action

since not only was the number of bars increased in that case,

but the fact that they were placed in two layerq increased the

area of concrete surrounding them. The stiffness of the dowel

element was thus increased, and a somewhat larger force may

then be expected to be carried by the dowel according to

Jones' theory, Krefeld and Thurston's(59) tests and Taub's(Ih1)

study of the problem. Anothor indication that this increase is

mainly due to dowel action is the fact that in all cases the

critical shear crack does not enter into the flange prior to

failure which implies a rather similar crack propagation, in-

226
dependent of the main steel. There is of course some change in

the stresses in the flange but its effect cannot be important

since combined with the dowel action it gives only a very

small variation in shear strength. It is not possible to

evaluate exactly the effect of each of these actions but the

resultant effect in strength has been evaluated by considering

all the results of shear spans greater than three and is of

the order of

*(3 V1.5 ) Qc
rstý -

where Qc is the shear force carried by the concrete com-

pressive zone.

In the theory developed in chapter 7 this small increase is

neglected because of the arguments immediately following.

When longer spans are used there is a slight decrease

in shear strength due to the larger a/d 1 ratios as will be seen

in section 5.4. In order however to obtain a shear failure at

longer spans it is necessary to increase the ratio of main

steel, otherwise a flexural failure would occur. The two

effects tend to compensate each other and can be neglected in

all practical cases. They do not entirely compensate each

other when heavy steel is present and the a/d 1 ratio is between

3 and 4 but even in those cases crack length (and consequently

the force carried by stirrups) is expressed in terms of a

height of crack d%, which represents the distance between the

bottom of the flange and the level of the lower layer of main

reinforcement. This'approach tends to give a slightly longer

crack in the presence of heavy main steel than would be the

case if the crack length were related to dl. Inevitably there

227
is in this
some small error region of a/d1 when heavy main re-

inforcement is used, but it is on the safe side and is con-

sidered small enough to be neglected in order to avoid the use

of too many formulae for particular cases. The formula deve-

loped in chapter 7 contains a 9-10% safety margin. This safety

margin is probably due to residual interlock and dowel actions

in excess of those taken into account by considering the con-

tribution of a stirrup beyond the crack length. It would be

ill advised to attempt to remove this margin without knowing

more about the parameter that may affoct those actions.

The failure of all beams occurred in the same general

manner by one of the main shear cracks extending into the

flange. Sometimes this resulted in longitudinal splitting in

the side of the flange as in T4; on other occasions the

failure crack went to the load without spreading and produced

a failure very similar to punching (T7). The failure shear

crack widened considerably at loads close to failure, 0.05" at

91% of the ultimate load in beam T7j 0.02" at 82% in beam T4.

It was also noted that the point at which the crack enters

the flange is not fixed. Although once it enters the flange,

the failure crack grows in the direction of the load platten

the position at which this entrance takes place seems to vary

considerably. This is illustrated by beams T4 and T7.

The length of the critical cracks of the various beams does

not seem to vary significantly, and their inclination has much

the same constant value mentioned in the previous section.

This is illustrated by the failure crack characteristics given

below.

228
Beam c di-t tan iT

No ins ins

Ti 15.5 7.7 0.497

T4 16 7.7 0.482

T7 14 7.4 0.528

The crack characteristics of beams T3 and T8 have already

been given in section 5.2.

5.4 Influence of shear man to off ective depth ratio

To study the influence of this parameter threo series of

tests wer© carried out. In one series the ratio of main rein-

forcement was small (1.46%) while the other two possessed

series

heavy main reinforcement (rst=4.16%) and differed only in the

degree of shear reinforcement. The relevant details of the

beams and the results obtained are given in table 5.4.1 while

fig. 79 contains the photos of the beams belonging to one of

those series.
5
During the test the shear span was increased from 2' to 7

in the case of the beams with heavy main reinforcement. The

failure of beam T17 with a7 shear span and a high degree

of shear reinforcement was the limit to which the series could

be tarried as it was probably a case of simultaneous shear

and flexural failure. The same applies to T31, which was also

the limit of the series with a reduced amount of main rein-

forcement.

The test results are plotted in fig. 80 in the form of a

graph representing the variation of ultimate shear strength

with the parameter investigated.

229
 o/d1 =2 4

Ts 0

o/d1 -36

Teo

ý'
yý

a/d1=54

T
o/ d1 -7.2

Fig 79 Influence span to effective depth ratio

of shear
T- beams failing in shear

230
 P. 1%

I-
n

O
7

U
G

(1.

L
Q
U

NN

q-

Fig. (80) Influence of the shear span to effective depth ratio

T- becwns failing in shear

ý?
TABLE 5.4.1 Influence of the (.afd3 ratio

Beam Ucyl Main 100A Shear rgfyq a/dl Ult. quit

st Load
brd1
No lbf/in Steel reinf. lbf/in tons lbf/in2

T-30 4160 3116mm 1.46 7/32,6" 83.5 2.24 22. 384

T-3 3990 3.36 21 366

T-31 4494 4; 50 19 332

T-5 4890 " " 7/32@3" 167 3136 28 488

1
T-33 5340 4.8 22 400

T-29 5349 8116mm 4.16 7/3206" 83.5 2.40 31 576,5

T-8 4530 3.6 25 466

T-19 4340 5,4 22.5 420

T-15 4810 7.2 21 392

T-23 4460 " it 7/3203" 167 2.4 39 728

T-9 2930 " " 1/404" It 3.6 31 578

T-20 4655 " 7/3203" 1' 5.4 30.6 572

T-17 4790 " it 11 7.2 27 503

(1) Flexural failure.

It can again bo seen that in a manner similar to the roctang-

ular beams the graph can be divided into two sections. In the

cases where the a/d1 ratio exceeds 3 the variation in strength

due to this parameter is quite small up to the shear span at

which flexural failure occurs. If an average shear strength

were considered constant the variation obtained would be well

within 10%. On the other hand in the case of short beams the

effect of this parameter is very significant and must be taken

into account.

As in the case of the rectangular beams this change in

232
behaviour does not take place at the same (a/dl) value in all

beams. It would seem that beams with lower amounts of main steel

do not increase in strength until (a/d1)

show an appreciable

values much lower than three are reached. The shear strength

of beams with heavy main reinforcement seems to increase

significantly from an (a/d1) value of about three.

The behaviour of beams at longer spans did not differ

basically from the behaviour of the beams at (a/dl) ratios of

3.36 to 3.6 discussed in the previous sections. The shear

cracks were more numerous but did not extend into the flange

until failure. The width of these cracks was not significantly

different before failure. In beam T17 the widths of 5 cracks

on one shear span wore found to be between 0.01" and 0.015"

at 89% of the ultimate load. These widths were also measured

at several height levels of the cracks and in general the

crack width seemed to be uniform throughout most of its length.

It was also noticed that in the longer spans the failure

cracks did not always appear at shear cracking. They would

appear at a later stage and be more inclined than the shear

cracks already existing. Beams T-20 and T-17 provide a good

example of this behaviour. What proved to be the failure

cracks appeared at 20 tons and 24 tons respectively, while the

first significant inclined cracks extending to the bottom of

the flange appeared at about 15 to 16 tons in both beams. In

both cases however these now cracks did not become signific-

antly wider than the other major shear cracks until the loads

were increased to almost ultimate values.

A similar behaviour was also shown by beam T-19. In this

233
case instead of a new flatter crack forming, the upper part of

an inclined shear crack that had developed from an almost

vertical flexural crack extended in an inclined direction

behind the top of the flexural part. In this manner a long

inclined crack was formed at about 90% of ultimata load and

failure occurred when this crack extended into the flange.

The length of the cracks on which failures took place did

not differ greatly at different (a/dl) ratios and'their in-

clination seemed to conform to the pattern indicated in the

preceding, two sections.

Beam c d1-t tan

ins
No ins

T-31, 14 7.7 0.55

T-19 14 7 0.50

T-20 15 7 0.47

T-17 14 7 0.50

In the case of the short beams the first inclined cracks

appeared at approximately the same load as in the longer beams

(T29
at 15-16 tons when heavy main reinforcement was usod. and

T23) and at 11 tons when reduced main reinforcement was

present (T-30). Whon the load was increased the inclined

cracks reached the bottom of flange but did not enter into

the flange until failure. Although more than one shear crack

appeared in a given shear span, the width of what proved to

be the failure crack usually became larger than those of the

other shear cracks, at a load stage half way between shear

cracking and failuro. In general the widening of tho cracks

was slower than in longer beams, but widths of the order of

234
0.02" were reached at 90% of the ultimate load in beam T29 and

at 87% of the ultimato load in boam T23. Furthormore the

failure crack was always the most inclined of all the shear

cracks and was also situated noaror to the support than the

others.

Failure in these beams took place by the extension into the

flange of the critical shear crack and a form of punching under

the loading platten. The lengths of the cracks were generally

smaller in the case of short beams' resulting in a steeper

inclination of the failure crack.

Beam c d1-t tan ý"

ins
No ins

T23 12 7.0 0.58

T29 11 7.0 0.64

T30 13 7.7 0.59

5.5 Influence df Concrete Strength

Two groups of three beams were tested in order to investigate

the influence of this parameter. All six beams possessed the

same shear span of 31, and the difference between the two

groups lay in the amounts of shear and tensile reinforcement

present. The relevant details and results of the tests appear

in table 5.5.1 while fig. 81 contains typical photographs of

beams in which this parameter was varied.

The results obtained are plotted in fig. 82 in the form of

a graph showing the influence of the concrete cylinder on the

nominal shear stress at failure,

235
 T13

wm

W
- P,
Ucyl : 1850 Ibf / in2 rgfyq: 83 5 Ibf

st° 1.46 °%

T25

Ali

Ucyl= 7841 Ibf/in rgfyq_ 835Ibf/in2

r : 1.460/.
st

Ts m

.,cyl _ý,.,.,...., .. q yq
rý+. 4 16°/°

T26 0

Vsr
}yam

ttom.... _--tw
r...._t_- --r
Uýyl= 8260Ibf/in2 rgfyq = 167Ibf /in2

rs t=4 16 °/°

Fig 81 Influence of concrete cylinder strength: T - beams failing in shear

236
 0 8

vý __ _

Fig. Mý influence of concrete strength T- beams failing in shear

TABLE 5.5.1 Influence of concrete strength

Beam Ucyl plain 100A Shear rgfyq a/d1 Ult. quit

st Load
2 brd1
No lbf/in Steal reinf. lbf/in2 tons lbf/in2

T-13 1850 3016 1,46 7/3206" 83.5 3.36 18 314

T-3 3990 21 366

T-25 7841 23 400

T-27 1742 8 16 4.16 7/3213" 167 3.6 26.5 494

T-9 2930 " it 1/4@4" It " 31 578

T-26 8260 " 11 7/3203" 11 " 36 672

It can be seen that there is a small but steady increase in

shear strength when higher strength concretes are used. As

shown on the graph this increase can be approximately given

by an equation of the form C+k (Ucyl)' in which only the

constant C, dependent on factors other than concrete strength,

need vary. The differences between the lines representing this

equation test less than *5%.

and the actual values are

In general the beams tested did not display significantly

different behaviours lengths of critical cracks

and similar

were measured in beams with different concrete strengths.

Beam c d1-t tan -

No ins ins

T13 15 7.7 0,51

T25 14 7.7 0.55

T27 15 7 0.47

T26 14 7 0.5

Beams T25 and T13 provide further indications on the in-

fluence of concrete strength on shear cracking and they also

illustrate the formation at a later load the failure
stage of

238
shear cracks. Shear cracking in beam T-25 occurred at 12 tons

inclined the bottom the flange; this

and the cracks reached of

is higher than the shear cracking of similar beams with lower

strength concretes. In beam T-13 of the same group shear cracks

and extended up to the level of the flange at 8 tons.

appeared

The ratio of the shear cracking loads in these two beams is

0.67, This is almost the same as the ratio of the cube root of

the cylinder strengths of the concrete involved; this ratio

(1850).
is _ 0.62. This finding supports the statement made in
(7841),
section 5.2 that the shear cracking load of rectangular and

T-beams is the same and depends on the same parameters.

almost

In that tho effect of the main steel was indicated;

section

it can now be seen that it applies to the other parameter as

well, namely the concrete strength. The cube root relation

also the results of Ramakrishnan's(91) tests on

agrees with

T-beams without shear reinforcement.

The behaviour of the present beams following shear cracking

will now be described. In beam T-13 more shear cracks appeared

at the next load stage of 12 tons and were in general more

inclined than the previous ones although their widths did not

differ (about 0.01"). At the next load stage

significantly

three of the shear crackq formed at the previous load stage,

widened significantly attaining widths of the order of 0.04"

to 0.05". Failure occurred when the most inclined of these

cracks entered the flange at 10 tons. In beam T25 tho widths

of the shear and flexural cracks was the same (0.006" to

0.007") at shear cracking. At 16 tons the widths of the shear

cracks reached 0.01411 to 0.01511 while those of the flexural

239
cracks were only 0.010". The additional shear cracks formed at

that load stage did not differ greatly in inclination from the

their widths were rathor smaller (0.008" to 0.011").

others and

At the next load stage of 18 tons all the shear cracks widened

two of them reaching a width of 0.0181; while the

slightly,

flexural crack did not open at all. When the load was further

increased the cracks continued to widen in a slow

shear

till 22 tons a new crack was formed. This new crack

manner at

the two thirds height level of one of the shear

originated at

that had reached a width of 0.018" at the 18 tons

cracks

load the bottom of the flange

stage; It extended and reached

in direction than the original crack.

a much more inclined

Failure the flange at

occurred when this new crack entered

23 tons.

5_.6 of inclination of shear reinforcement

__Influence
Four beams with varying degree of shear reinforcement

having stirrups inclined at 450 were tested. The spacing of

the stirrups measured perpendicular to their. direction was

the same as in four otherwise similar beams with vertical

stirrups. In this latter case the spacing has been measured

the direction. of the main steel which in fact is also

along

a direction perpendicular to the stirrup inclination.

Table 5.6.1 contains the relevant details of the beams

tested whose photographs appear on fig. 83.

240
 T22
R

rgfyq= 83 5lbf/in2 rst= 1 46°/°

T, m
%

rq yq=83 5 Ibf / in2 St= 4 16%

Týý

ýý7n acs,

r fyq : 1671bf/in2 st' 4 16°/a

q

T14
smow

1-in

rgfyq= 326Ibf/in2 rst _416"%

Fig 83 Influence of Inclination of stirrups (45°)

T- beams failing in shear

241
60

50

KEY
40-
Q90 Q45 Ultimate shear strengths of
similar beams differing cnly in
30
aLn stirrup inclination

20-

10-
0 5°1a increase
0
0
0 100 200 300 400
degree of shear reinforcement in Ibf / in2,

Fig. (84) Percentage Increase in ultimate shear strength of T- beams

with stirrups inclined at 45°

Sý d
7dß ý`

cC

Fig. (84a) Stirrup spacing and critical crack length in

T- beams with inclined (45°) stirrups

244
TALE 5.6.1 Influence of inclination of shear reinforcement

beams inclined 45 °
- with stirrups at

Beam Ucy1 main 100A Shear s sin q rf a/d1 Ult. qult.

st q yq Load
2 bd reinf, 2 2
No lbf/in Steel r1 at in lbf/in tons lbf/ i n
G< =45°

T-12 4450 8j16mm 4.16 7/32} 6" 83.5 3.6 29 541

T-11 5360 1/4jä 4" 167 " 32.1 598

T-14 4840 if if 3/81 4k" 326 " 44 821

T-22 '4980 3016mm 1.46 7/320 6" 0315 3.36 22 384

In fig" 84 the results are plotted in the form of a graph

showing the percentage increase in strength due to the inclin-

ation of the stirrups; this has been obtained by comparing the

results obtained in this series with those obtained in tests

of otherwise similar beams with vertical stirrups.

It can be seen from this graph that beams with inclined

stirrups, attain higher shear strengths than equivalent beams

with vertical shear reinforcement. This increase seems rather

erratic and apart from one high result lies within a 5% limit

even for beams with degree An

a high of shear reinforcement.

even smaller increase in the load necessary to produce shear

cracking was also noticed. In a manner similar to the behaviour

of the beams with vertical stirrups failure in these beams

took place when the critical crack entered the flange. It was

however noticed that the final widening of the main inclined

crack prior to failure took place at 5 to 10% lower loads

than in the case of equivalent beams with vertical stirrups.

This agrees with the results of £oody's(78) tests on restrained

beams in which yielding of the inclined stirrups seems to have

243
taken place at 80% of ultimate load. However the widening of

the crack mentioned above affects primarily the stirrups

crossing it in its central part. It was therefore necessary

to increase the load in order to produce yield strains in the

stirrups situated at the extreme ends of the long failure

crack. Failure took place once this had been accomplished.

The failure crack length was found to be significantly

longer than in the case of beams with vertical stirrups.

Beam cI d, c dl-t
d1
ins in c
No

T-12 24 10 2.4 7/24

T-11 22 10 2.2 7/22

T-14 20 10 2.0 7/20

T-22 24 110.7 12,25 17,7/21

It can thus be seen that the crack length can be taken as

about 2.2d1 as compared with 1.5d1 for vertical stirrups.

Actually this value of the crack length combined with tho in-

clination of the stirrups may provide an explanation for the

lack of any significant increase in shear strength. To do this,

lot us consider the shear carried by the stirrups crossing the

failure crack, with the following notation.

c= horizontal projection of critical shear crack

Aq = stirrup cross section (one loop)

Fqv = vertical component of force of all stirrups crossing

the failure crack.

The number of 450 stirrups crossing the failure crack is

c+ 0*7d1
approximately
3

244
 Provided the stirrups crossing the failure crack yiold, the

shear force FqV they support is see fig. (84a)

- -
c+0.? dl
45 o
sin xAqf
q=sx yq
0
450 br
multiplying by sin .
sin 45 r

b c+0.7d1 2v
x (sin 450) Aq=F
br. ---- q yq q
rs sin 450

x (sin 45°)2 (c+0.7dl) brr Fq"

b s sin
si 450
r
A
but
q
bs sin 450
r

and (sin 450)2 = 0.5

rgfyq (c+0.7d1)b FqV

x 0.5 r=
In the case of vortical stirrups the force in the stirrups

crossing the crack is approximately

Fqý = rgfyq bxcc

In beams with c has been found to be

vertical stirrups

approximately 1.5d1 in all cases

Fqv = rgfyq bý 1.5d1 for vertical stirrups

in the case of inclined stirrups at 450

Fqv= rgfyq b 0.5 (c+0.7d1).

rx
However c in this case was found to be much larger than in

the case of beams with vertical stirrups and close to 2.2d1

Fqv= rgfyq rx0.5 (2.2d1+0.7d1)

= rqfyq xbx1.5d1

It can thus be soon that if the stirrups yield, which seems

to be the case judging by crack widths and the strain measure-

ment presented in section 5.10, then the contribution to the

245
shear resistance of vertical or inclined reinforcement is not

too different provided that in determining the degree of shear

a spacing perpendicular to the direction of the

reinforcement

stirrups is used.

Turning briefly to the case of rectangular beams failing in

the height of the crack may be expected to be

shear where

higher to 0.75d1 treatment, resulting in

and closer a similar

lengths of the
Fqv=rqf xl. 5bd1 would again require crack
Yq
2.1 to 2.2 dl. The only rectangular beams with in-
order of

that type failure con-

clined stirrups reached near-shearing

ditions its length was of the order of 2.0d1

was R19 and crack

that beams with stirrups failing in shearing

while of vertical

was about 1.5 d1.

51 7 Influence of transverse flange reinforcement

To investigate the influence of this parameter tests were

carried in the amount of transverse

out on beams differing only

flange reinforcement All the beams tested had the same

used.

standard cross section used throughout the tests and a 4.16%

ratio of tensile reinforcement. The degree of shear reinforce-

was 326 lbf/in2, but two of the beams had stirrups in-
ment

clined at 45°. The relevant details and results of these

tests appear on the table below.

246
TABLE 5.7.1 Influence of transverse flan e reinforcemont

Beam Ucyl Inclin. Transv, p% Load at Load at Ultimate

of which a which Load Vy
2
No lbf/in Stirrups Reinf, longitud, flange (T)
lip
crack was de-
appeared tacr. cd
in the (T)
flange
(T)

T21 4700 90° - 28 30 -

T37 4615 90° 1/41@911 185 32 - 42

.

T32 4000 90° 1/40©6" 1276 32 - 43.5

T6 3740 90° 1/40c3" . 556 36 - 41

T24 5140 45° 32 34 -

- -

T14 4840 45° 1/4j03" 556 36 - 44

.
tf
A
in the table p t--g-"-9 --
tf
tf
where A is the area of one bar of transverse reinforce-
9
ment and stf the spacing of the transverse reinforcement.

In the cases the flange became detached the load was

where

not increased that the beam

any further as it was considered

had ceased functioning as a T-beam. The photographs of fig. 85

illustrate the physical aspect of the flange when increasing

amounts of transverse flange reinforcement are used.

It can be noticed both from the table and the photographs

that the first longitudinal crack appeared in the flange at

approximately the same load in all cases. However the behaviour

of the beams was quite different after that. In the absence of

any transverse reinforcement the flange became detached from

the web, when the load was increased slightly above the level

at which the first longitudinal cracks In the case of

appeared.

247
 1
Mumm

p=0

/*

P= 0"1b3°%

-eu ""
rk

Iii'
ý. ;..
"" "___ 8x/ )
! ý__ - `
__ ýýýfý-ýý{.
-_ ..

p=02 8°/°

\..
-.

T6
4L4=6
ýý /
ýý,
ý,

ýA
p- 0-56%

Fig 85 Effect of percentage of transverse flange reinforcement

on the cracking of the flange

248
beams with transverse reinforcement this did not happen, and the

only result of increase of load was the formation of more

cracks and the extension of the existing ones from the centre

of the beam towards the support. The number, extent and width

of these cracks was smaller in the case of an increased ratio

of transverse reinforcement. In general the whole physical

of the flange top seems to be preserved better in the

aspect

of increased ratios of transverse reinforcement. Tho

presence

fact that the cracks extend from the centre towards the supports

is also indicated by the load strain curves of figs: 86 and 87.

These figures represent the strains in the transverse rein-

forcement of the flange. It can be seen that this reinforcement

is not significantly stressed till the load reaches 28 tons;

it would thus that the longitudinal cracking of the

appear

flange really took place at 28 tons that is at the beginning

-

of the load increment rather than at its end; this is exactly

the same load the first longitudinal cracks appeared

at which

in a similar beam without transverse reinforcement (T21).

Furthermore it is very clear from the positions of the bars

and the loads they register an appreciable increase

at which

in strain, that the reinforcing bars situated closer to the

load are stressed first. As the crack progresses towards the

support at higher loads it affects successively transverse

reinforcing bars situated at increasing distances from the

centre. Once a crack has crossed them these bars tend to be

stressed quite rapidly and the strains moasurod show a definite

tendency towards attainment of yield values. In the of beam

case

T32, where these strains were measured, the crack

main shear

249
 L7

wlý

N 0

0
Cl)

0
0 cv

c
0
U

C
CO c

4,1 tL v
1-1) v
0
O

0
O0

Fig. (06) Load strain curves beam T-32

-
strains in the transverse flange reinforcement

2 5O
 I)
M

)
)

c
0
4-3

a
C

v
v
G

ö
4J

0
t)

0
S
r
r

Fig. (87) Continuation of fig. (86)

8

2
entered the flange and occasioned failure, before all the

transverse reinforcement had yielded. Although a similar

failure took place in beam T37, the condition of the flange

before failure was worse than that of beam T-32. Considering

the fact that beam T-37 had only two thirds of the trans-
also

reinforcement of beam T-32 it is reasonable to assume

verse

that shear failure took place almost at the load at which all

the transverse reinforcement was yielding and that this per-

centage of flange reinforcement must have been approximately

the limit ensure that the beam would function as

which would

a T-beam up to failure.

Figures 88 and 89 show the strain measured at different

positions in the flange of beams T21 and T24 respectively.

These beams contained no transverse reinforcement. It can be

seen that the tensile to and at 45o to

strains perpendicular

the axis of the beam are very small before longitudinal crack-

ing. Once this happens the crack is wider in the region nearer

the load The same pattern of tensile concrete deform-

plate.

ation was also found in the beams with transverse reinforce-

ment. However longitudinal cracking occurred the demec

once

readings, after this stage can only indicate crack widths

which

rather than deformations, were much smaller.

It is interesting to note here that in the case of beam

T-20 the ultimate load was 30.6 tons tensile stresses

where

the order of 300-400 psi were measured at 29 tons near the

of

load platten, perpendicularly to and above the web flange

junction. This is an indication that longitudinal cracking of

the flange was imminent at a load not very different from that

252
 öÖ41 0
40
co
(D N 0
ýJO dfl

Hub
4
41
t°- 12 19
O
lý O

O£S OZ- O O
1g1
di
O q O
o 0
Ui to ý
C
`
30
\
096 OLSb OL- OZ
O O
0) t0l O
co 10
F)
O O

09:
N
0992
09;i PI °l
oý.
os' o
0
(o
P
0 0
ýO
ö)
.I
ý
"U
181
It
O
IäI co
O"
n
10 d"
c
Os OZgS- 0 1 oa
I
oi
2"o
C, I 0
I N.
O
rn /O "cýý

Oc Os- 03- OZ- OL os-

I
I rn

inns
U
izil ICII,
U r-

I- to

(88) Concrete 6/
Fig. strains (in in x 10 in) at the top
surface of beam T-21

h t. i
a---- ._-

v7 ýý
t:
c
Mo
co
N
O
II 0

P')

ioddnS

ý011
o
119 0 V
0
0.

U
oI 4J d
41
iý
FQ
rýJ pý

ca
0
cý
p
O
I
I
rýI r)
ýJC ewo
off'
N CV
(o

I
'0 IO
r

O
ro O1
/o O 111

' O
Iý In
'I

OO O 0
ö
pOO
0
rn rh co

IO'OCI o
F-Ill
!O

0
öý
"0
Lo
CO o2
ell
C
oý, ',
r
1

o
Iý
O U
Y0,

lry

OO
ýI

O Ici4ý
Nm\
U!
IC J

jq
ýj-j. II
co
ýO,

oddng Gallo

Concrete (in in x16 beam T-24

Fig. (89) strains /in) at the top surface of
which produced longitudinal cracking in the beams reviewed in

this section.

In general it can be said that, provided a minimum percentage

of transverse reinforcement is present - capable of ensuring

a shear failure before the flange is detached - the amount of

transverse steel has no effect on the shear strength of a T-beam.

It is only when stated in these terms that the writer can agree

with Ramakrishnan's(91) conclusion that transverse reinforce-

ment has no effect on the shear strength of T-beams.

508 Influence of flange breadth

To investigate the influence of this parameter two beams

having flange breadths different from the standard 24" used

throughout this research were tested. In one case a wide 42"

flange was used but this width was reduced to only 12" in the

case of the other test see fig. 90, Pertinent details of

-
these beams and of T and rectangular beams are
comparable

given in the table below.

TABLE 5.8.1 Influence of flange breadth

Beam d1 b br t Uoyl Ult. qult

Load
2 2
No in in in in 1bf/in tons 1bf/in

R-24 10.0 6 6" - 4480 18.5 345

T-34 10.0 12 " 3" 4920 22.5 420

T-19 10.0 24 " " 4340 22.5 420

T-35 10.0 42 4880 23 429

Note: In all cases: rst = 4.16%, rgfyq = 83.5 lbf/in2,

5,4; in the case of beam R-24 rsc

a/dl = = 2.63%

255
 b 12 t3

/ 135 N"
-I.

b= 42" t-3

T36

AWW
6" rgfyq1671bf/in2 b= 24
t_

T38

b= 24
t_ 6" rq fyq = 326 Ibf/in2

Fig. 90 Influence of flange geometry

T- beams failing in shear

256
 0

0
Cl

s
U
C

0
CV fl
.

t_
.44
V
U
L
A

e.ý
U)
r
U

0
I-

Fig: (91) Influence of flange breadth T beams falling in shear

-2f')i
 These results are plotted in a graphical form in fig. 91.

This graph indicates that there is an increase of about 20%

in shear strength between a beam having no flange (rectangular)

and a T-beam with a rather small flange breadth. However any

further increase in the flange breadth does not produce any

appreciable increase in shear strength. This would indicate

that the flange is in carrying

only a limited part of effective

shear, On the other hand the more or less uniform compressive

measured sections across the flange - see section

strains at

5.10.3 indicate that in all cases the whole of the flange

-

co-operates in resisting bending stresses.

Shear cracking and failure of these beams did not differ

from that 24" flange, failure occurring

of tho beams with with

the flange. The aspect

when the critical shear crack entered

that the beams presented at failure was slightly different.

In beams T-34 and T-19 when the crack entered the flange the

failure extended to the cantilevering parts of the flange

into two. In beam T-35t where b=42"9

which was thus separated

the failure was restricted to the region near the web and did .

not spread across the whole flange.

The failure crack lengths followed the established pattern

and did not differ significantly.

Beam c d1-t ?
an -
No in in

T-34 14 7 0.5

T-35 15 7 0.47

As in previous cases of longer beams more than one shear

crack appeared in the same shear span. In beam T-34 the load

258
was brought to zero after shear cracking in order to place

demec points around the shear cracks with the intention of

following the behaviour of these cracks. When the beam was re-

loaded and the load brought to 16 tons all the four major in-

clined cracks on both sides of the load had the same width

in their central parts. When the load was further increased

to 20 tons the failure shear crack was 20% wider than the

other major shear crack in the same shear span. The domec

points pcsitioned in the central parts of the four major

cracks showed negligible compressive strains parallel to the

crack. They also indicated that the crack opening was uniform

and that this opening was created by a more or loss vertical

movement as the vertical increment in width was twice as large

as the horizontal. Actually in normal T-beams more important

(35)
vertical displacements were noted by Evans and measured

by Leonhardt and Walther(68) in beam TAll. As far as the flex-

ural cracks are concerned, the width of the widest flexural

crack at 90% of the ultimate load was only 30% of the maximum

shear crack width.

5.9 Influence of flqpýqe dfe th

_ý
Two beams having flange depths of 6" that is 0.6d1 were

tested in order to investigate the effect of this parameter.

The geometry of these beams seems rather unreal for a practical

T-beam, but their behaviour should provide indications con-

corning the behaviour of beams with intermediate flange depths.

The relevant details of these beams whose photographs appear, in

fig. (90) and of comparable rectangular T-beams

and other are

259
 60

54

ß-
.

c40
i

3C

L.
a
c,
r_
N
2C
4º
wv
E
:ä

is

12 :S
flange depth (t) in inchzs

Fig. (92) Influence of flange depth T- beams failing in shccr

,

^. -#'
.
given in the table 5.9.1 below.

TABLE 5.9.1 Influence of flange depth

Beam br b t U-yl rsc rqr Ult. quit

Load
2 2 2
No in in in lbf/in % lbf/in tons lbf/in

R16 6" 6" - 4580 2.93 167 28 522

T9 " 24 3 2930 - " 31 578

T36 " " 6 3500 - "j 36 671

R28 " 6" 4580 2.63 326 36 671

-

T6 " 24 3 3740 - 41 765

T38 " " 6 4380 - " 48 896

Note: In all cases rst 4.16%, a/dl 3.6, d1 = 10"

These results are presented as a graph in fig. 92. As can

be seen from this there is a definite tendency for the

graph

to increase with increasing flange depth in

shear strength

the case of the two degrees of shear reinforcement used. This

implies that the shear resistance of a T-beam is affected by

the flange depth; it is difficult to explain this effect in

terms of a shear comprossion failure as in that case only the

upper portion of the flango depth would really be relevant

regardless of the depth of the flange below the shear crack.

Therofore an equation for predicting the shear failure of T-

beams must take into account the actual flange.

Although the failures of these beams had the aspect of a

shear compression failure, crushing at the top of the flange

did not take place until tho load stage at which the critical

shear crack entered the flange. Thus the general observation

that failure in T-beam takes place when the major shear crack

enters the flango is still valid, and an equation developed

261
to predict the conditions under which this happens is still

applicable in this case even though the actual failure that

follows presents a different aspect. A possible explanation

for this behaviour may be the following. Under certain con-

ditions the crack enters the flange and travels towards the

load; Because of the increased load carrying capacity of the

member compressive stresses under the load platten - maximum

moment - are quite high but they are not high enough to cause

failure. When the failure plane reaches that area after entering

into the flange it provokes a further concentration of stresses

and the failure takes the aspect of a compression failure at

the top of the beam. Since however this happens at the same

load stage at which-the crack enters the flange this writer

believes that failure can still be expressed in terms of the

conditions governing the entry of the failure crack into the

flange.

The width of the major shear crack in beam T-36 was 0.025"

at 89% of the ultimate load while the width of the major flex-

ural crack in the web at that load stage was about 0.005". In

beam T-38 where conditions are closer to flexural failure the

corresponding widths of shear and flexural cracks at 83% of the

ultimate load were 0.01911 and 0.010E'.

The lengths of the critical shear cracks werd much smaller

in this case than in the case of beams with less deep flanges.

Their inclination however seems to have remained constant.

g
Beam c d1-t tan
No in in

T-36 8 4 0.5
T-38 7.5 4 0.53

262
5.10 Strain measurements on concrete and reinforcement

5.10.1 Strain measurements on shear reinforcement

Strains measured on the stirrups of a number of beams are

shown in the form of load-strain curves in fig. (93) to (98).

The strains were measured by placing strain gauges on the

stirrups. Since the maximum strain on a stirrup inevitably

occurs at the point where it is crossed by a crack the gauges

in beams T3 and T12 were positioned on an imaginary line re-

presenting the direction of the expected failure crack. Because

of the greater number of shear cracks expected to appear in

beam T20, having a longer shear span, no such attempt was made

in that case and the gauges were placed at mid-height.

From these figures it can be seen that no appreciable

strain was registered before the appearance of inclined cracks.

This was true even in the case of inclined stirrups which may

be crossed by flexural cracks prior to shear cracking. It

would appear that th© widths of the flexural cracks at this

rather low load stage was not sufficient to provoke any appreci-

able strains in the inclined stirrups.

Following the appearance of inclined cracks the strains in

the stirrups started increasing rapidly. This did not happen

automatically in all the stirrups but as soon as they were

crossed either by a new shear crack or by the extension of an

existing one..

Examining'the behaviour of stirrups 5 and 6 situated near

the centre of beam T-3 it can be seen that they are the least

strained of all, especially No 5 which is situated just ahead

of the point at which the critical shear crack reaches the

263
 failure side

position of strain -gauges

a/ d1 --3.36
zp
rq fyq _ 55.71bf/ink - 9"crs

cf tab! e 5.2.1. and fig. (75)

leid strain of stirrups

r. i
t p"
o

000

X9ý

500

fait
0 5- 15 (174)
10
total load in tons

Fig. (93) Load strain curves

strains the stirrups of beam T -10
_in

ýý.
 jP
inure s-id-

2000

20 13 13 17
position
15 15
I
HtLLLtff
of strain - gauges
14 13 12 10 9

aid 1 5.4

rq fya 167 lbfl in2 -¢7" hD 3" crs

cf table 5.4.1 and fig. (79)

10iß

/" %
yield strain of stirrups 12
,,_,

7:... 13'
1000-C /
cj
41
rr
.r r
/
c rý
YDý i

r.

500 0

10 /

12, "ý'"fIý
fail
". LJ
0 10 20 30(30-9)
r-ý ýýý
c total load in tons.
0

ta.
E
0
....

r
Fig. (94) Load strain curves.
strains in the stirrups of beam T- 20

r., -,..
..,, -ý
 ' '19

17° 15

16 ý

1500 8

JO 7/
Id strain of stirrups

1000 /

'0
x
G
20'
//
1
G

tV
cn
Ü

N
Soo

e ý/

30
total load in tons

Fig. (95) Continuation of fig. (94 )

,

ýý ý- .
 no
NQ=.
/
®° 49
0
cv

47
'ýO c
O

O
O

F-'
tO
e

R_

----0--
N

Q- 2 =t4 1
N tý

to ý.
ei
s cv th ti.
U1
:

J)
cn 1
.
(L N
0- tn
N 0)
G ýd
CD
ci) .ä
0
to

ID T
I
411- :2
Ö
Ln
0)

O
--i}

ýOý

. -' (uoisuaj) ui / QL x ui ui suioJ}s

0
o
Oj

Fig. (96) Load strain curves

strains in the stirrups of beam T -3

o ? Gýýý
 _12
a
a
v
N

Co
0)

,
wa

to
u

F-

CL E
U

«ßi
Z
'l)
4J o"
UU tn
a
E. o v
UN rn
Ur

QL r_ c.
a
N
LC
ýQ O II '-
NU O
ÖN
r_
E 0
U.
v7
UUN
Gl p, 0
CL
ö

° cn
L. =
ce
,, º ,C N.
Q)
=o
1Iwü
th
lz.

26`
1000

fail
1 _. 1
0 (29)30
total load in tons

a/d1 : 3.6
83.5 Ibf/iný 3 ät 45°
rgfyq
cf table 5.6.1 and fig. (a3 )

Note : for the position of the strcin-gauges see fig. (97)

Fig. (98) Load strain curves

strains in the stirrups of beam T-12 - stirrups inclined at 65°

4ý'
under flange. The behaviour of this stirrup tends to indicate

that the critical shear crack does not extend into the flange

- as would be expected if the crack were to cause a shear com-

pression failure - until failure takes place. Furthermore it

can be noticed in this beam that high strains were registered

in the stirrups situated beyond the main shear cracks. These

stirrups are brought into play by the main steel acting as a

dowel. It'seems thus reasonable, in a beam with stirrups, to

take into account the force carried by dowel action by con-

sidering as fully active - i. e. stressed to their yield point -

in supporting shear more stirrups than those actually crossed

by the inclined failure crack.

During the tests some of the gauges became damaged and

ceased functioning at load stages before failure. A reasonable

estimate of their behaviour can nevertheless be made by examin-

ing their behaviour before they stopped functioning and com-

paring it with that of other stirrups that did not become

damaged.

As mentioned previously the various stirrups displayed

appreciable increases in strain once they had been successively

crossed by an inclined crack. The rate of increase of strains

seems to be steeper in the case of inclined stirrups, and the

stirrups situated in the central portion of the failure crack

seem to yield at an earlier load stage than would be the case

in an equivalent beam - same rgfyq - with vertical stirrups.

However the failure crack in the case of a beam with inclined

stirrups is longer than in an equivalent beam with vertical

stirrups and additional increases in load are necessary to

265
produce yield conditions in all the stirrups crossing this long

crack. In fig. 96 it can be seen that the four stirrups exhi-

biting a slower rate of increase than the others are those

situated at the upper and lower ends of the two major critical

cracks.

5.10.2 Strain on main tensile reinforcement

The strain profile in the main reinforcement was measured

by placing a number of strain gauges at regular intervals

along its length. These gauges were placed on one bar only

but in each measuring position gauges were placed at both the

upper and lower surface of the bar in order to eliminate the

strains, due to the bar1s own bending. Two examples of the load-

- strain curves obtained appear on figures 99 and 100 corre-

spending to beam T-8 and T-20 respectively. Both beams had

main steel ratios of 4.16io but the degrees of shear reinforce-

ment and (e/dl) ratios were different: 167 lbf/in2 and 5,4 in

beam T-20 but only 83.5 and 3.6 in beam T-8. Because of the

difference in shear span the maximum strain in beam T-20

should be greater at the same load stages. Although this is

true in the central region of the span, in the regions near

the supports greater strains were measured in the case of

the shorter beam which indicates that the arch action is more

important in that case.

Figures 99 and 100 suggest that the behaviour of the steel

before shear cracking is as would be expected of a member

failing in flexure. However when the load is increased beyond

that stage there is a tendency for the central part of the

266
 Yý

Fig: (99) Load strain curves,

Deformation profilcs of main reinforcement bccm T-a
-

nr y'.
2000

ý.

1500
4 1/ /H\\ \\
\\
1000

4J

500

C failure side
a 54N L

r&t 4.16 °%
-
rq fyq = 167 Ibflin2
Pult = O'9 tons

Fig. 000) Load strain curves

Deformation beam T-20
profiles of main reinforcement -

2 ä!.
strain profile to become flat. This tendency is more pronounced

in the shorter beam, but in both cases it is more significant

during the later load stages and in the shear span in which

the failure shear crack exists. In beam T-8 the failure crack

was formed at 22 tons in a more inclined direction than the

already existing shear cracks see fig. (68) the change in

- -;

the main steel strain profile in the span in which it occurred

is illustrated by the uppermost line in figure 99 which shows

an even more definite change of behaviour in that span at the

23 ton load stage. It can therefore be concluded that at loads

close to failure there is a marked tendency for the strains

along the main steel to become uniform under the critical shear

crack.

The behaviour of beams with inclined stirrups is similar to

that described above but the more or less flat part of the

curve tends to extend over a longer region' due presumably to

the longer cracks.

5.10.3 Strain measurements on concrete

Both demec points and strain-gauges were used in measuring

concrete strains at various load stages. The study of these

strains and their implications concerning the behaviour of

the beam can be undertaken by considering separately the two

types of measurement carried out. These were:

a) Strains on the top surface of the beams

b) Strains on the sides of the beams.

a) Strains on the top surface of the beams

Strain profiles were measured along the center line of the

269
beams in a manner similar to the rectangular beams. Furthermore

strains were also measured across the flange in order to in-

vestigate their distribution across it.

In trying to determine the strain profile along the longi-

tudinal axis of a beam it was of course impossible to have a

measuring point in the middle of the beams under the loading

platten. To infer this strain from measurements on the side of

the flange, as was done by Ramakrishnan(91)1 would not give

a true picture of the actual strain profile because at the

central region the strains tend to diminish progressively

towards the outer edge of the flange. The values representing

the strain in the middle of the span in the strain profiles

appearing in figs. 101 to 105 was thus measured 2" from the

side of the loading platten (5" from towards the edge of

the flange.

Figures (101) to (105) represent top fibre strain distri-

bution of four typical T-beams which failed in shear and of

one T-boam that failed in flexure. The general aspect of the

strain profiles corresponding to shear failures is not too

different from those obtained in the case of rectangular beams

in the sense that after shear cracking, the region of com-

diminishes progressively. (It is interesting

pressive stresses

to note that this behaviour is almost nonexistent in the case

of beam T33 - fig. (105} that failed in flexure. ) The strains

at the centre of the beams tend to be higher when more stir-

rups are present but this seems to be due to the fact that

higher loads are withstood and conditions approach those of

a flexural failure. Also higher strains at the centre occur

270
 i

C A

ald1 =3.36
Pufft 114 tans
_ 55.7Ibf/in2
gfY9
cf table 3.2.1. and fig.

/'\'1692d/oof '%v-t ultimate)

P =12tons

Fig.. (101) Top fibre strain profile (ý) beam T-10

/.
/
/ a/d1 : 3.5
rgfyq = 83.5 Ibf1in`
(91°1° of ultimate) 20j ý cf table 5.3.1. and fig.

/ Put = 22 tor S
P_16tons

/ a: 36" \
//
i

Fig. (102) Top fibre strain profile ((L',

) beam T-7

c? G"
 c
0
2000

CL
F
0
U

(91010of ultimate) 400t ., a/d, -3.6

1500 Z rq fyq : 326 Ibf/in2
cf table 5.6.1.and fig (83)
C s'
36t
r_
1000ý- u
c
Q
ült 44 tons
N

X4t
ýR.

500-5
0
0
ý2t ý"
Eton ý' "°---ý: =
p"
T
o-
0ý

C
O_

C
G

500
79
36"

Fig. (103) Top fibre strain profile (c")

becm T 14 stirrups at 450

a/ C-'
 o a3
E

v-
0

O)

cli
r' E
N
u
U

y-
0
Z-
O.

C
O) Ü
r C-
N
tT)
'o=
L
Cljc V
c
°
[1
F-
M
LO - co Lri
nn
.D
ff U
r
° v
1" 1"
. 1;
M
{ý S

(uoissa. ldwoo) u! /S p(, x u! U! su! oJ}s 3421J3uo: ) (uoIsua)

0
ooo0 o
o000
p In
I-
0
.-
Il'
127
 O
v
th

"ý /

'- /
.0

.. uI
CD
11
Lrr
d"

"r/

1/a
#1

uj -- -
".: "'
c

\\
v

a.. 1
"

01

1.

N1

44

u! susDJJS 3Iaa3uo3
(uoisszJdwoo) ui 14L X ui
00 C) 0 0
0
4')
cV r r
C\l

Fig. (105) Top fibre strain pr ofil, -s (cE) Beam T-33 : flexural-fa ilur"
.

6-
1 i/i_
 in longer spans but in that case also, flexural conditions are

responsible. It can be noticed however that even in the case

of long spans T19 fig. (104) high degrees

- - or of shear re-

inforcement (103)
- T14 fig. - the beams that failed in shear

were significantly less stressed at the centre than beam T33

which failed in flexure; this can be seen by comparing the

strain profiles at equivalent (same percentage of ultimate)

loads. Even if the strains in beams T19 and T14 are relatively

high in the centre it should be realized that they represent

the peak of the strain distribution at the top of the flange

along a line perpendicular to the axis of the beam. It has

been noticed that when a compression failure takes place

crushing occurs across the whole of the flange implying large

strains not only near the load but also across the whole

breadth of the flange. Large strains of the order of 2000x10-6

near the edge of the flange were measured only in two cases; in

bopm T-33 which definitely failed in flexuro and in benm T17 where

failure can be described as simultanoous shear and flexurc.. I n conclusion

assuming that strains similar*to those occurring in beam T33 must be reached

if a flexural failure is to take place, it can be stated that

oven in terms of strains, compression failures did not take

place in the beams tested.

The aspect of the strain profiles on either side of the

point of load application indicate that there is a progressive

loss of compressive strains at sections towards the support.

This implies an inclination of the main compressive thrust.

However it must be kept in mind that these strain measurements

were taken directly above the web and all that can be said at

276
this point is that there is evidence of an inclined thrust in

the web. To determine how important that thrust iss in com-

parison with the total compressive force in the flange, the

distribution of strain in sections across the flange at some

distance from the point of load application must be investi-

gated.

Figures (106) and (107) show typical distributions of strain

in the flange at sections close to the point of application

of load. The proximity of those sections to the centre implies

that they are situated in the region of increasing strains in

the longitudinal strain profile curves. As can be seen from

figs. (106) and (1071 corresponding to beams with high and low

degrees of shear reinforcement, the strains across the entire

flange increase at increasing loads and their distribution

is approximately uniform across its breadth. This agrees with

the strain distribution across the breadth of the flange re-

ported by Ramakrishnan(91); a perusal of his results revealed

that this finding was based solely on the investigation of

sections close to the centre of the beam.

measurements of strains at sections situated further from

the centre than is the case of figs. (106) and (107) show a

different pattern of strain distribution. This difference in

behaviour becomes apparent when the applied load exceeds the

value causing loss of compression in the flange region

directly above the web. In other words, the behaviour is differ-

ent if the section is situated outside the region of continuous

strain increases in the longitudinal strain profiles of figures

(101) to (105). Typical strain distribution across such sections

277
 o/d1 . 3.6
ra fyq : 326 Ibf/in2

ült 41 tons

strains across a section 6" from centre

c
0
' 40t 'üº
N

38t n

-1000
36t

32t

28 t
?.4t

20t

16t

12t

ßt

P: tons

Lo

,
rig. (106) Distribution of strains across the flange of beam T-6

A3
 I

-600
c
0
ül

20t ti
0.-
E
0
U
v
-600 17t '

b
16t r- 15t

12t
12t

P=8 tons R6 tons

____----°

' axis of ald1 = 3.3$

beam
rgfyq = 83.5 Ibf/ in2 rgfyq:: 83.5 Ibf/in?

Pult 22 tons Pult = 22.1 tons

--
strains ccroz-, a c:: ction
;, trains across a scction
O'1 from centre 4° from centre

Fig. (107) Distributions of strains across the flange of T- beams

4/?
 Yýp

id
Nä
v

oS

I
rý 1
C
wö
a
ä
sN
O '-
N
co
Q a- i.
Cti i
mil a3

C
0
.J

CL

" ý1
ýI
a+ CI
N
CQ
I,

" -0 aý a
bj

co
a II

(uoºssa. Jdu+.
oo) ul/ pt x Lit Lit , U!DJ S (uo1sual)
9
00000
000'0

Fig. (108) Distribution of strains across the flange of beam T,-10

oQ; -
 20 Z9

Aý P
54a 7Y77!

old1 = 5"4-
rgfyq 83.5 Ibf/in2

ui` . 22.5 tons

13 2t

P=14-5t
---- P=1a"2t
P=21-4t 300 '--

P=10.9t ca.
P_14.5t

., Px21.4t
-200 P_10.9t
P: 7.271 ".
10 ý''' """ P: 21.4t
ý .. '
x P: 7-27t "'""::
o ,ý..
c 100
P_ 3.65 torts
cI P_3.65tons

,c
o.

section A section ß

Fig. (109) Distributicn of strains across the flange of beam T-19

'1ri
 lp
,do-It t

Si S2 '

20" 20"

a/d1 _ 3.6
rgfyq = 1671bf/in2

Pult ` 36 tons
(high strength concrete)

"U
section S1 00 section 52

Ip
r

2ý3t 200
24t°".. ...
ze.ot 1st
ýý
12t -100 f'_12t ý ^ýý24t
o wo
ýP_32t *ZZ Bt

. ***ýv

o r--

100

II i

Fig. (110) Distribution of across the flange of b.-cm T-25

-strains

ý:
GCi G.
 S1
,P ISc
"' 20 20

aldl -3-6
rq 'yq _ 32615f/ in
Pult = ß'r6 tons
500

40t pý ö
section S1 section ä2
36t
` ü. 4C0

%, 0 p... 40t
`p c 36t " ýý.
ö
24t -300
2- t

X; 16t
15t p----. "ý _____- . _. p
-200 --. ___"p

12t p-- --ý- °-_' _""' -"--ý

.___ _ý
u7
8tonsý.
°-"" 100 a--P": "
P=ß tons

-0

Fig. (111) Distribution of strains across the flange of beam T"35

flange depth 6"

2týý
 are shown on figs. (108) to (111). It can be seen from these

figures that the strains cease to be more or less uniformly

distributed across the flange when the compressive strains in

the region above the web start to decrease with increasing

load. The rate of decrease of the strains is much more pro-

nounced in the part above the web. This implies that although

an inclined compressive thrust forms in the wob, its value

must be mainly dependent on the compressive forces above the

web which are but a fraction of the total compressive force.

A large part of the total compressive forces does not contri-

bute to this inclined thrust as it remains in the flange at

all times. The latter is further confirmed by strain measure-

ment on the side of the flange as will be seen subsequently.

The difference in behaviour at sections equidistant from the

centre but in different shear spans depends on the location

of those sections in respect to the main shear cracks. Since

in T-beams the position of these cracks is somewhat random

different strain values can be expected. It remains however

that the basic pattern of behaviour is not different and

agrees with the general case described.

b Strain on the sides of the beams

Longitudinal strains on the side of the flange wero measured

using Demec points. The position of the sections and the flange

depth levels at which those measurements were taken are indi-

cated in figures (112) to (117). These figures the load-

contain

strain relationships obtained in a number of typical cases.

Although democ points were also placed in the web, in sections

corresponding to the position of the points on the flange, they

284
 0-1
rw

O
3 Cýý

C
O

Q0

ä
aý
0

0
0 LO
_0 'N

ýý1
-!

L
r
10
ß au
O

E
O
v- 0
J. 1

C 3
Li

Fig. (112) Strains on the side of the flange

Beam T-22 = 22 tons
- ült

,.: c.
 w
C
U
,a+

'0
cý
0

U
a-+
0

N
N

CL
c
c
Ö

to
G
O
4
d
.o
v- C

a
0
+1
b

vO
0
a
A
aL
O
r v

MON
fý fý

Fig. (113) Strains on the side of the flange

Beam T-25 - pult = 23 tons

L.
vG
 N ýj

--

N
N
s
iý
N

to O
0
0
O
N

r-
N

Q
ýý r

Fig. (114) Strains on th-- side of the flang-

Beam T -26 Pult -- 35 tons

-

ýV/
 N
" C
_

CJ
IT Cn
C C
v- 3

4L ý0

C
a O
44 v;
I C)
a

N c1

ý.. NN

O
._N
cD
ct)

c
O

O-0
NO
IO

ýj Ö
0
.o
IR

(uoJsszadwoo) uff/ Qf.xu! uI w! oJIs

00 0
N^ °'

Fig. (115) Strcins cn the side of tha flange

Beam T- 36 Pult _ 36 tons
-

c. 0
 L

4i
C
O
^rý

ý
a.

E
n
a

O
Cn
C

v=
cr
L

C
O

tr1
C
v
r
N

ai
o8o
o
N
to

ui / 01 x ul ut , ulD.a3s
to
N Li-
d

I#.
0-
d
U
kký
L
N

aN

.G

aNý,

r 1ý
It
©ý
co

aaN
I _l. ____
a 86
o000
N T-
U! /, (PL X U! UI SUIDJ; S

28.
 V co
9 n

0
N

L
r-
N
0.

\
co
N

1
N
4 r

r_*
0
.J

c
WLU T3
t3
qo
r^ý

0
U
JN
to
Ut
to
-0
U

d ON I
G1
r- U O
V'I
ýJ d
"7

OO
--a
: -ý rr
0.0. l
.1
1

(uoissa. jdwoo) (uaisuzI)

.I
UI! Q1.x u! U! 5UJtJ is
0 0
o0
0 0 0
N '" R

(117) section , beam T-30 - Pu,,t= 22tons

Fig. Strains across a vertical

ao
were crossed by cracks at early load stages and do not appear

in these figures., One exception to this was the case of beam

T-30 where it was possible to read the corresponding strains

in the web under near ultimate conditions.

It can be seen from these figures that at early load stages

increasing compressive strains occur at all lovole of tho

flange depth. When the load is increased there is a tendency

for the compressive strains at the lowest level to exceed those

higher loväl, s in the flange depth. This implies that the'

at

inclination of the thrust, which in the central part of the

beam takes place within the webe is in this case restricted

to the flange depth resulting in compressive forces at the

underflange. Comparing the strains in the bottom of the flange

with those of the wob in fig. 117 it can be seen that although

thoy are much smaller they must be assumed to represent quite

a large compressive force since they exist over a relatively

large breadth. This existence of increasing compressive forces

in the underflange is also illustrated by the strain reading

on beam T-7. In that case strain gauges were used in the under-

flange and demec points on the top surface. Longitudinal

profiles were thus obtained along sections near the

strain

fig. (118) fig. (119) the flange.

middle - - and edge - - of

The load strain curves shown in these figures, where the full

lines represent the strain in the top surface of the flange

and the dotted lines the strain of the corresponding positions

in the undorflange, indicate increased compressive strains in

the underflange.

It seems therefore that the inclined thrust in the case of

291
 tons
E
-10
10 ,ý, '
7%/7

P 12 tons
--
-200

100
r_-, ý
P
ý ý /

P _16 tons

200

/m--
-100 "ý ý''ý" - -4-- -. " -4

c d i
-100. U

P: 20 tons

300

"--. °'----tip
-200

100
P
0
1 /
/
/
I
100 10

-I-511 KEY
top surface

strains on top surface

rAB bottom surface

I

i
Ordinate represents concrete
strains in in x 10 6/ in
The line of zero strain also
bottom j,

represents the length abscissa

Pult 22 tons
'

Fig. Longitudinal strain profiles in the flange of beam T-7

-(118) strain section A
at
 X00
P. ýS tons
1O0

0 -., P

P= 12 tons
`00
100

0
cr

P 13 tons

200

100 -ýýVý

r "ý
0

P. 20 tons
-400

-300
-10\ i
f% %k
? /
_00

100
P
O

100

Fig. (119) Longitudinal strain profiles in the flange on b_em T- 7

strain at section B

? 9:
T-beams cannot be as important in resisting shear as is the

case-in rectangular beams. In that case the whole compressive

force contributes to the formation of an inclined thrust in

the web and the ensuing arch action constitutes the so called

shear resistance of the compressive zone. In the case of the

T-beam it would seem that only part of the compression is

inclined into the web and the rest stays in the flange whore

the possibility of important inclination is limited. As a

result a smaller fraction of the total compressive force can

support shear through arching action and in this case the re-

sistance of the zone above the failure shear crack may be

limited to the shear resisting properties of the concrete as a

material.

Strains were also measured in the webs of beams T7 and T8.

The strain gauges used for that purpose were placed on either

side of the major shear cracks after shear cracking had taken

place. The strain obtained were similar in both cases and those

measured on beam T8 are shown on fig. (121). Figure (120) indi-

cates the position of the measuring points. The maximum strains

attained before failure were in the neighbourhood of 300x10-6

in/in which, although small when compared with the strains

attained in web crushing failures, are an indication of the

existence of compressive stresses in the web. It can also be

seen that at loads below 20 tons when the crack widths were

small (approx. 0.01") transfer of forces by aggregate interlock

enabled almost similar strains to exist on either side of each

shear crack. When the load was further increased the width of

the major shear crack in one shear span widened to 0.02" while

294
 C'ý% / ýý 1
ý IýJJ

(V
II
r
1..

`U

L'
U
ll

1.: -\\
Cý

IL

/
E
U

.n
ý> )
ý-ý/ý' 0
ý` /
ýý

týt
t'1

0. ý i.

C>

r.
0
F-
V1
O
0.

0
N

U_i

CJ

\
\v

ýý;
Cv

ý\ ý
ý".

CV; ý' 1

ý`ý
 L7
l..

it)
N
1 L ,
CL

E
C

cr

O
Gl
/
Lr) (0 e- r ``)
rr
N i° r

.. N p
.

'gy
po
U

76
O

U!/ x u! u! : U!Djls u, /90I ui u( Su!DJ; s

t i7
QQ 0 0
o
o N
I-

NLn r- N to (D
o. .. v
``

b
41

(uolsszjdwoo)
U! Ip (, X U! U!, SUIDAS U! I OL X U! Uf , U!DJ 5I
t I1 19- 1tp
OO 00 8
0 00
0 04
Cll

Fig. (121) Strain gauge readings in the web of beam T-8

for position of numb? red gauges see fig. (iao)

-p ýk=
a new more inclined critical crack appeared on the other span.

As a result all the strain gauges situated between these cracks

and the ends of the beam showed continuous increases. Of the

strain gauges situated between the cracks and the centre of the

beam only the gauges closest to the cracks showed appreciable

gains. It can thus be reasonably implied that the major com-

pressivo'thrust lies in the concrete block above the major

shear crack and that the widening of cracks, or the formation

of cracks with a more pronounced inclination, diminishes the

effect of aggregate interlock action in the sense of transfer

of stresses across cracks.

297
 CHARTER 6

INVESTIGATION OF WEB CRUSHING FAILURES

It has been seen in the preceding chapters that if the

degree of shear reinforcement is increased higher nominal shear

stresses can be withstood before failure. In normal rectangular

if the is increased
and T-beams, amount of shear reinforcement

sufficiently a flexure failure will occur before the shear

of the beam is reached. However in cases where the

strength

degree of shear reinforcement is vary high, as in the case of

thin webbed I-beams having very closely spaced stirrups or

stirrups with high yield valuos, a secondary failure may take

place before the flaxural capacity of the beam is attained.

As has been mentioned earlier, shear in a beam is resisted

mainly by a combination of an arching action by the main com-

pressive force cbovo the cracks' end a truss action below the

cracks involving tensile stresses in the stirrups and com-

pressive stresses in the concrete. In a rectangular beam the

critical zone of compressed concrete is the reduced area at

the head of a crack. This is not so however in T-beams because

the width of the compressive zone above the cracks is very

large due to the existence of the flange. However the situation

in the web may become critical whore the width of concrete in

compression there is much smaller. This of course is more

likely to occur in the thin webs of I-beams especially if the

truss action is increased by the provision of heavy shear rein-

forcement. In this case failure may therefore occur by crushing

of the concrete under and between the inclined cracks in the

298
web. Stich a mode of failure is commonly referred to as 'web

crushing'.

An interesting assessment of the conditions loading up to

web crushing is given by Robinson(101). In reviewing Leon-

hardt and Walther's tests of series ET(69) he states that:

"If the following series of beams with vertical stirrups

is oxaminod

a) rectangular beams

b) T-beams with low degree of shear reinforcement

C) un high n it rr of

d) I- "r n It to it n if

it can be seen that the shear resisting mechanism involving

the compressive zone is gradually transformed and becomes that

of the classical Ritter-1 rsch truss. The proof that this is

so, at least qualitatively, lies in the fact that in the last

case shear failure may occur by compression failure of the

concrete struts formed by the cracks in the web. "

The purpose of the tests to
r_ported hero was therefore ob-

tain failures by web crushing in

under varying conditions order

to investigate'the_r effects on web crushing Once

strength.

the basic trends relevant to this type of failure are estab-

lished use can then be made of the other available results for

tho purpose of arriving at a rational formula for predicting

web crushing failures. This is done in chapter 7.

To obtain wob crushing failures the web breadth of the T-

beam of chapter 5 was reduced to 2.5" and high yield stirrups

(fyq j 90000 lbf/in2) were used. To avoid flexural failures the

beams were also provided with a tension flango in which 8X16 mm

299
bars of high tensile steel (i' > 90000 lbf/in2) were placed.

Although only a limited number of tests were carried out an

attempt was made to investigate the following parameters:

1) Concrete strength

2) Amount of shear reinforcement

3) Spacing of shear reinforcement

4) Inclination of shear reinforcement

All seven beams tested had the same cross section and

tensile reinforcement. The shear span to effective depth ratio

was 3.5 in all cases.

In the opening section the general behaviour of the beam

is described and the results of the seven tests given. These

are discussed in the following four sections which deal indi-

vidually with the effects of the parameters listed above. The

final section is devoted to the presentation and discussion

of strain measurements on the concrete and reinforcement.

6.1 Behaviour of thin webbed I-beams

Of the seven beams tested in this series six failed by web

crushing. The only one that did not do so had stirrups in-

clined at 450 to the axis of the beam. Table 6.1.1 gives the

relevant details of these tests,

300
 TABLE 6.1.1 Web crushing failures
(Stirrup
Beam U AJ Shear r rf q q q
cyl cyl i q q yq ult ult ult
spac ng Ucyl
No lbf/in Reift. in (%) lbf/in lbf/in
'
1: -10
Wl 4332 165.0 , ä m 41 2.16 1950 1405 0.324 21.35
n 90
W2 1827 42.8 791 0.433 18.5

1113 6605 181.2 1755 0.266 21.6

{ 2ß(4 0m 1
I
i W4 4985 70. 5 4 2.44 2200 1975
045 5
W5 4970 2ý1O0m
70.5 3 3.25 2925 1670 0.336 23.7
90
W6 4920 ýý900m
70.15 6,5 2.16 1950 1360 0.277 19. '4
J
W7 Qýýý0
4980 70.5 3 2.09 1873 1417 0.281 20.1

Note: Failure by web crushing in all beams excepting W4

Loading of the beams was carried out at 5 ton intervals and

in all cases the first inclined cracks in the web had appeared

by the 10 ton load stage, usually in the middle of the shear

spans. When the load was increased more inclined cracks appeared

on either side of the first ones, and the majority of them ex-

tended ultimately from the tension flange to the bottom the

of

compression flang©. The shear cracks did not enter the com-

pression flange at any time during these tests. The direction

of the shear cracks in the web varied slightly from 35°

about

at the region towards the supports to about 400 in the region

closer to the load application point.

Additional inclined cracks appeared higher loads

at and the

web was then progressively transformed into a number of con-

creto struts. The widths of the inclined cracks were generally

quite small, compared with those in the T-beams of chapter 5.

and although all the cracks were not of the same width a sur-

prising number of the larger ones had similar widths. In beam

301
W1 for example the widths of the five major shear cracks

differed by only 10% at about 80% of the ultimate load. It was

only at loads very close to failure (95% of ultimate) that the

cracks situated in the quarter span nearer the support showed

some slight tendency of widening more than the others, but

their widths always remained quite small in comparison with

widths indicating yielding of the stirrups.

As mentioned previously the web was transformed into a

number of concrete struts. Strain measurements on these struts

showed that they were subjected to an inclined compression

which increased as higher loads were applied. The rate of in-

crease of the web strains was found to be smaller in beams

with higher rq values and in beams with the same rq value but

having inclined shear reinforcement.

In the case of vertical stirrups failure occurred when

this increasing compression in the in the

struts resulted

spalling and crushing of the concrete under and between the

inclined shear cracks. Although crushing occurred over quite

some length it was more pronounced in the quarter span region.

The strains in the stirrups remained below yield values in all

six tests.

When inclined stirrups were used the crack widths were

generally smaller. more importantly, the rate of increase of

the compressive strains in the struts was much slower than in

the case of beams with vertical'stirrups. Consequently higher

loads can be supported before the strains become high enough

to produce crushing of the concrete in the struts. This process

however allows the strain in the stirrups to become quite high

302
 Wi

w2

0 iý5

Ws 4 ý`

Fig 122 Thin - webbed I beams failing by web crushing

1503
 W!
f.

I I.

-S

beam wi

:.ý'
4r '

t/
/ýAý/

ý,
r,

3'

beam W`9

Fig. 123 Thin webbed I- beams failing by web crushing - closeups

.
of failure regions
'9o4
 Wx

ýýoý ýýo
3 I

14
ti 1fýtý

7-m
7----ý- %11 15

Fig. 124: Thin webbed I beams failing by web crushing -

close ups of failure regions

'o5
 W4

reinforcement cage of beam W4

Fig. 125: Thin I- beam inclined (450)

webbed with stirrups

wi
1?!
i

Fig. 126 : Typical aspect of the top of the flange in the tests
web crushing

306
and reach near yield values. Unfortunately one of the stirrups

in beam W4 broke at a sharp bend at 45 tons and the test was

not brought to its logical conclusion. It remains however

that a load almost one and a half times the web crushing load

of equivalent beams with vertical stirrups was withstood. No

great increase in strength could be oxpected beyond the load

causing yield of the stirrups as the beams would probably dis-

integrate at that load.

Photographs of the six beams in which web crushing took

place and close-ups of the regions affected appear in figures

(122) to (124). Figure 125 contains a photograph of beam W4

and its reinforcement; the latter provides a good illustration

of the arrangement of the shoar reinforcement in the beams of

this series. A typical example of the aspect of the top sur-

face of the flange during these tests is provided by fig. (126).

6.2 Influence of concreto strength

An indication as to the influence of this parameter can be

obtained by comparing the concrete strengths and nominal shear

stresses at failure of beams Wly W2, W3. These beams differed

only in the cylinder strength of the constituent concrete.

Furthermore beams W6 and W7 can also be used since their only

difference from beam W1 concerns the spacing of the stirrups.

As the latter does not seem to affect the failure load it is

considered appropriate to include these beams in-the present

comparison.

307
TABLE 6.2.1 Influence of concrete strength ;, pesent testsZ

a/dl = 3.5

Beam Ucyl qult qult quit

Ucyl 67.
No

11
W1 4332 1405 0.324 21.35
!
W2 1827 791 0.433 18.5

W3 6605 1755 0.266

121.6

W6 4920 1360 0®277 19.4

W7 4980 1417 0.281 20.1

It can be seen from this table that if failure occurs by

web crushing then the nominal shear stress at failure can be

reasonably considered to be proportional to the square root

of the cylinder strength of the concrete. The only other series

where such variations of concrete strength occur were those

of J. R. Robinson(103), partly reported in the last CEB plenary

session in Lausanne, but not yet published. However the writer

spent four months in the summer of 196? at the IRABA labora-

torfies near Paris whore these tests were carried out, and has

thus been able to obtain complete details of them.

TABLE 6.2.2____Influence of concrete strength l Robinson's tests,

,:
a/dl= 25

Beam Ucyl qu qu lt
2
lbf/in2 lbf/in
YUcyl

8Q12 2560 974 19.3

BQ7 4670 1320 19.4

BQ6 4900 1350 19.3

6Q15 5580 1450 19.5

30©
 These tests, although involving smaller variations in con-

crete strengths than the tests reported in this thesis, do how-

over substantiate the square root proportionality.

The square root proportionality had also been proposed by

the appropriate committee at the last CEB meeting in Lausanne

(103)
The present findings seem to substantiate this rather

than Leonhardt's linear dependency(65). However the square

root proportionality must not be interpreted as indicating any

relationship between the maximum resistance of a cracked web,

whose struts are in compression, and the tensile strength of

concrete whose proportionality to the square root is implied

(31)(64).
by some codes The resistance of the cracked web would

seem to be more related to the rigidity of the web concrete

which is proportional to the square root of the cylinder

strength(27)(123)(96)ý

6.3 Influence of shear reinforcement

Web crushing failures in the six beams with vertical stir-

rups occurred before yielding of the stirrups. This seems to

be the general case of beams with vertical stirrups failing in

web crushing as is substantiated by Robinson's results of

series BQ(103) and NR and NT(98) and Leonhardts beam T1(69)

The only exception to this pattern seems to be the case of

Leonhardt and Walthor's beam ET4(67) where web crushing occurred

at approximately the load at which yield strains were being

attained in the stirrups. This beam had one of the lowest

ratios of shear reinforcement for which web crushing was re-

ported (rq=1.0) and can thus be considered as a lower limit case.

309
- actually just one web crushing failure has been reported in

which the shear reinforcement ratio was lower than 1.0%; this

was the case of beam TA2 by Leonhardt and Walthor(68) . However

the concrete strength in that case was only 2200 lbf/in2 while

in beam ET4 the concrete strength was 3240 lbf/in2.

In view of the above and the fact that if the stirrups do

yield a shearing typo of failure would be expectod, it is con-

sidered more appropriate, when comparing web crushing failures

in terms of shear reinforcement present, to comparo geometrical

(rq) than (rqfyq)
rather mechanical percentages.

Only one beam with an increased ratio of shear reinforcement

was tested and a considerable increase in web crushing strength

was obtained as shown by tho table below.

TABLE 6.3_. 1 Influenco of the ratio of shear reinforcement

Beam Ucyl rq qu lt
1bf/in2 (%) 2
lbf/in

W1 4332 2.16 1405

W6 4920 2.16 1360

W7 4980 2.09 1417

W5 4970 3.25 1670

Those results of course can only indicate that there is a

tendency for the web crushing strength of a beam to increase

with increasing ratios of shear reinforcement. This tendency

is used as a guideline in chapter 7 whore all the existing

results of web crushing failures are studied in terms of in-

creasing values of this parameter.

310
6.4 Influence of stirrup spacing

Three-beams were tested in order to investigate the effect

of this parameter. The ratio of shear reinforcement was kept

constant by using stirrups of the same quality but having

different diameters. As a result, in order for the ratio of

shear reinforcement to remain constant, different stirrup

spacing was used in each case.

TABLE 6.4.1 Influence of stirrup spacing

Iu
Beam rq spacing qult
iibf/in2 (%) 2
inches Ilbf/in

W7 4980 2.09 3" 1417

W1 4332 2.16 4k" 1405

W6 4920 2.16 6.5" 1360

It can be seen from the above table that the web crushing

strength of the beams was not affected by the stirrup spacing

which in the case of beam W6 was more than twice that of beam

W7. The spacing in beam W6 was almost 0.65 dl which is very

close to the maximum spacing of 0.75 dl proposed by the Shear

Study Group for inclusion in the unified code(ll).

It can thus be reasonably concluded that in beams where no

excessive spacings are used, the web crushing strength is in-

dependent of this parameter. The reliance of the web crushing

(94)
strength on the stirrup spacing as postulated by Regan

seems to be a somewhat conservative approach.

6.5 Influence of the inclination of stirrups

Beam W4 was the only beam tested in connection with this

parameter. It had stirrups inclined at 450 to the axis of the

311
beam and a 4" spacing measured perpendicular to the direction

of the stirrups. The ratio of shear reinforcement was thus of

2.44% which is slightly higher than the 2.16 of comparable

beams with vertical stirrups. It sustained load up to 45 tons

which is far in excess of the 32 tons sustained by the beams

with vertical stirrups; It is even greater than the 38 tons

that produced web crushing of beam W5 having an rq of 2.25%.

(64)
This finding substantiates Leonhardts statement to the

effect that the oblique compressive stress causing web crushing

is largely dependent upon the direction of the web reinforce-

ment.

In the present test the, web did not crush at failure because

at 45 tons the stirrups broke at the point in the flange where

they were bent rather sharply in order to act as transverse

reinforcement of the flange.

(98)
In Robinson's IR and IT tests where 45° stirrups were

used, although high nominal shear stresses were obtained, web

/failed
crushing did not take place, and the beam in flexure.

Leonhardts beam T2 having stirrups inclined at 450 failed at

a much higher load than the corresponding beam Ti. Failure in

both T2 and Ti was by web crushing and the increase in load

in T2 was 45%. In the present tests the corresponding increase

in load at failure was 40% even though a rather premature

failure took place. It would thus appear that the recommendation

of the CEB commission on shear(103)

(max qult)45o = 1.5 (max qu1t)90o

is substantiated by the available test results.

312
 Failure in beam Ti took place as the stirrups reached yield

strains. Strains close to yield were also reported by Robinson

in the case of beams with 450 stirrups. In the present tests

as well the strains were quite close to yield when failure

took place. Although the yield value seems to play a more

i
determinant role in the web crushing of beams with inclined

stirrups the fact th; st only one case of actual web crushing

was reported and the observation that the greater part of the

increase in strength takes place while the stirrups have not

reached yield values would tend to justify even in this case

an approach based on the goowetricel percentage of shear rein-

forcement.

6,6 Strain measurements

6.6.1 Strains in the shear1reinforcoment

Typical load-strain curves are presented in fig. 12?, 126 tnd

129, corresponding to beams W1, WE, and W4 respectively. Beams

Wi and W5 differed in the ratio of shear reinforcement while

beam W4 had stirrups inclined at 450

.

It can be seen that in the case of vertical stirrups web

crushing occurred long before the stirrups reached yiald

strains. Furthermore, as expected, the rate of increase of

strains in beam W5 with a higher rq value is much slower than

in the case of beam Wl having an inferior r4 value. It can also

be noticed that when the ratio of shear reinforcement is in-

creased there is a tendency for the line representing the

stirrup strains to be closer to the theoretical line based on

the Morsch truss theory. This of course agrees qualitatively

1 313
 (failure end)

po=ition of strain gauges 179ý-

N
CY)

0
CIO

0
N

0
fý'

Fig. (127) Load strcin curves

strains in tth stirrups of becrn W1. rq : 216°! o
.
5 111
 ýP
wcb crushed on both ends

2- 1 1

position of strain gaugzs

I)

J
V

Fig. (128) Load strain curves

strains in the stirrups of beam W5 rq = 3-25"/o

315
 2s

positions of strain gauges

Fig. (129) Load strain curves

strains in the stirrups of beam W 4. r9 4ý1ý
q 3tß
with what has been said in the introduction concerning the

transformation of the shear resisting mechanism. However the

theoretical mörsch values are still somewhat higher than the

experimental values, even at failure. It would thus seem that

even in cases of very high ratios of shear reinforcement, a

proportion of the shear force is still being resisted by the

flange.

In the case of inclined stirrups the rate of increase of

strain is not too different from that of a beam with vertical

stirrups having an equivalent rq value. Since web crushing

does not take place at the same loads, the strains in the in-

clined stirrups continue to increase until near-yield strains

are reached. In the case of Leonhardt's beam T2(11) when the

,

web crushed the apparent stress in the inclined stirrups was

4400 kgf/cm2, their yield stress was 4270 to 4350 kgf/cm2.

while

Unfortunately as mentioned previously in beam W4 some stirrups

broke in a bend prematurely at 44 tons and the experiment was

not fully completed. The strains in the stirrups however had

reached almost yield values and it seems that yield of stirrups

precedes web crushing of beams with practical amounts of shear

reinforcement inclined at 450. Although in the majority of

cases a flexural failure will have occurred before this in-

creased web crushing strength is reached a method for its cal-

culation is developed in chapter 7. It shows a good correlation

with the limited number of test results on this subject.

6.6.2_ Strains in the main reinforcement

Strains were measured in the main reinforcement at several

sections along the span of beam W7. The resultant load strain

317
 failure side

position of strain - gouges

1500r
üit = 32.5 tons

5^f.

C
loco OJ
vs

e
in
? '-`t.
xý

C
lot
500 -D

U)

tons
. "P=5

Fig. (130) Load- strain curves

strain profiles in main reinforcement beam W7,
,

31 F>
4000

160t

3000 1'0 2.5rn 2.5m

ý1rný
P/2 P/2
120t

bc

2000 Ü,
"ý - loot
Lý üit

00,
1000
P: 40t
ýý
fy = 4740 kg/cm2
i

0
abc

Fig. (131) Strain in main reinforcement of berm T1 according to

profiles
measurements reported by Leonhardt and Walther54

tension in
lined
compression
stirrups
web concrete

change of force in
main rcinforcem; nt

Fig. (1-32) Equilibrium conditions at level of main reinforcement

ýý9
profiles are presented in fig. (130). The only other web

crushing failure where such readings were taken was that of

(64)
beam Ti by Leonhardt and Walther where strains in the

tensile reinforcement wore measured at three sections in the

span. These readings are plotted in figure (131) in a manner

similar to that of figure (130).

Examining these figures it can be seen that the behaviour

of the main steel is much more(flexural, than in a beam of the

same span with a lower degree of shear reinforcement failing

in the b&
shear, although some flattening of curve can still

noticed in the quarter span closer to the load. Leonhardt and

Walther this by stating that "The struts of the 'truss'

explain

thrust the tie rod already at some distance before the

against

bearings, part of the shear force being transmitted back into

the compression zone by stirrups". It should probably be

understood. that the portion of the shear force supported by

those struts is in this case than in beams failing in

greater

other modes. A closer illustration of this behaviour can be

obtained by considering equilibrium conditions below the shear

illustrated by fig. (132)1

cracks as

It can be seen from that figure that the change of force

in the main reinforcement will be more important at the region

where the stirrups are more stressed and the concrete more

compressed, in other words in the region of the quarter span

nearer the supports where the web actually crushes. In the

quarter span nearer the load there is a tendency for the con-

crete to be less stressed, witnessed by the fact that crushing

does not take place there. Consequently the change of force in

320
the main reinforcement is smaller in that case and the strain

profile tends to be somewhat flatter.

6.6.3 Strain measurements on concrete

The strains in the web were measured in all seven beams by

means of 4" demec gauges placed on the web. These measuring

points were placed in a direction at 45° with the axis of the

beams in the line of action of the expected compressive thrust

in the web struts. In beams W51 W69 and W7 measuring points

at 300 were also used, Figures 133 to 138 show typical results

of those measurements for beams differing in the ratio of

shear reinforcement and in the angle of inclination of the

stirrups. For comparison purposes analogous measurements on

Robinson's beam BQ12 are also included and appear in fig. 135.

These various diagrams will now be discussed individually:

BeamW6 (figs. 133 and 134)

The parts of the web in which measurements were taken aro indi-

cated in figures 133 and 134. Each measuring position consisted

of three pairs of demec points, two inclined at 450 and one at

300. Since the readings at these two inclinations did not-

differ significantly, only the strains at 450 are shown in this

case; a comparison between strains at 300 and 450 is made sub-

sequently.

The load-strain curves obtained from those measurements are

in figures (133) and (134) for measuring positions.

plotted all

It can be seed that the strains at the failure end (ABCD) were

much higher than those at the end that did not crush. This

difference did not become pronounced till after the 20 ton

load stage (65% of ultimate).

321
 U

;J

8
If)
N

of C
" 0

` ": ö:
t; %,.. in ._v CIO
Noa0C
ý. E 10 41
ö
O-'
E
E
;
o 70-
N

LU

Ch
r

8
to
E
V
.a

94-
O

10
9-
x
.C ti
r

N
0'1

ß Cý)
E
fr)
G7

U! PDOI 10104
SUOI
-
95 0NN if) O= it)
.- . -. 3 i'.:
 8
CV

' 8

0
0 o
Lo

N
Li\1 G

L
++ 0
N
O
U
O
"ý
O
C1 a
L L

E Cr.
U U

I-

O
Q Li.

+`+_.
suo} ui pooh Io; o} _

NN0 LO O LO
co cýj

62 :
 a
a
ýO
r
v

°
.äz, -
0
Ö_ ä

Läd
Nv

aÖ UI)
-ý, L.
ý\ NN
ß.

0
0
0
N
I-

E
U
to
N
I-

E
V
tf)
N cr) 61
W
N

<ý,
aCX.,

:3:3 C) 0
Cn 00
Ör)
Ul
NrN

'ii O
\©

`\\

fý (L

to

IN.

56i{OSOZ= I! un L) s4iun ui d pap1

(D M

Fig. (135) Strains in the web of beam Bß12 by Robinson

524
 There is also a tendency for the strains to be higher in the

region of low moments; this tendency is also more evident in

the failure side where it became manifest at earlier load

stages. The same thing can also be noticed in Robinson's beam

BQ12 whose load-strain curves appear in fig. (135) (results

-

obtained by the writer during his four month stay at the IRABA

Laboratories).

Beam W7 (fig. 136)

The arrangement of measuring points was similar to that for

W6. Each measuring point was composed of three pairs of demec

points. Two of these pairs were inclined at 45° while the third

pair, positioned between the other two, had a 300 inclination.

In an attempt to bring out any difference in behaviour due to

the inclination the load strain curves are plotted individually

for each measuring one line representing the average

position,

strains at 45 0 the 30°. Figure (136)

and the other strains at

shows these load-strain for the span in which web

curves shear

crushing took place. From that figure it can be seen that the

difference in strain these two directions is not significant.

at

The behaviour of this beam was similar to that of beam W6 and

the various observations made in connection with the latter

are also applicable in this case.

Beam W5 (137)

The positions of measuring points were similar to those

shown in figs. 133 gnd 134 for beam W6. Measuring points A and C

were constituted in the same manner as in the case of the pre-

vious beams but B and D were composed of three pairs of demec

points inclined at 450 and two pairs inclined at 30°. Average

325
 0 0
0 0
a
N N

0 0
O
0 0
0

4-1
N0 41 O 41
00
0 00O

0
0
O

O
0
0
U
O r-

Cli
)

Fig. (136) Compressive strains at different inclinations in the

web of beam W7. r_ 2-09% P : 32.5t
q ult
s2G
 0
S 0
Iný

0
0
to 0
\ oý

DLr)

O
0
LO . 0-1 t)
a
L

SUO; U( PDO( (DJOI 5U04 U! prate IDW-4

1Ir111IIrI I
&j 41
0N0
O :n0NpO
cf) Lr)
r- r'

Ö
a

öqt
"ý

CON

NI tn

ID 03 suol
CUO U! pool ui paoI Ie; off
n NQNO0
NNr, ý0

Fig. (137) Compr es5ivZ strain at different inclinations in

the vicb of beam W5. ra --3-25% Pult= 33tens

X27
 values for each measuring point and direction are plotted in

fig. 137 for the failure span of beam W5. It can again be seen

that no significant difference exists between the strains in the

directions measured. The similarity of strain is more pro-

nounced in the quarter span near the support where the actual

web crushing took place.

Keeping in mind that beam W5 had a higher ratio of shear

reinforcement than either W6 and W7 a comparison can be made

between the strains of these beams at the same load stage. It

can be seen that when a lower ratio of shear reinforcement is

used the web concrete is more severely stressed. The ratio of

shear reinforcement must thus be a determinant factor in the

web crushing strength of otherwise similar beams.

Beam W4 (fig. 138)

In this beam in the stirrups were inclined at 45° to

which
the horizontal, points along
strains were measured at various

the span using demec points inclined 450. Web crushing

also at

did not take place in this the were not

case and strains

significantly different throughout the shear span. In fig. 138

average values of the strains in the quarter span region are

plotted against the load. For the sake of comparison the

corresponding strains of beam W7 having vertical stirrups

and an almost equivalent ratio of shear reinforcement are

also plotted in that figure. It can be seen that with vertical

stirrups the web concrete is much more severely stressed than

it is with inclined stirrups. The difference in strain however

is not consistently in the ratio 2: 1 as would be implied b,,,.thetlorsch

truss theory provided some means existed for coruorting these

328
 Inclined (45) stirrups
W4 rq 2.44°1° Vertical stirrups
, -- "
W5 rq _ 3.25°I°
,
30 - '2
ro
C/-ra

" ý'ý Vertical stirrup

-
W ., rq= 2' 09°I°
0z el
20 .1e
Vý
wi

10
i
i
61 in
strains in in x 10

1000 2000

Fig. (138) Inclined compressive strains in the web of beams

with vertical and inclined (45°) ' stirrups
.

3L5
strains into stressas. As it is shown in chapter 7 it is not

possible to use the strains in a cylinder test for the purpose

of converting into stresses the strains measured in the web.

The reason for this is the biaxial state of strain that exists

in the web where the concrete struts are not only subject to

compression but also to tension through the stirrups.

To provide a further illustration of the favourable strain

conditions in the web concrete in the presence of inclined

stirrups the corresponding load-strain curve of beam W5 with°

a much higher rq value is also included in fig. 138.

330
 CHAPTER 7

SHEAR STREIGTH OF REINFORCE Q. CONCRETE BEANS S

_ý...... _ .. `. ý....... » .... ,_.......... 'f * '. _ ..._. ,. _....,, .,

The general purpose of this chapter is to present equations

for calculating the shear strengths of reinforced concrete beams.

Equations are proposed for the shear strengths of singly and

doubly reinforced rectangular beams, T-beams and thin-webbed

I-beams.

The theoretical considerations are assisted by the experi-

mental investigation, involving tests of 75 normal sized beams,

presented in the preceding chapters. Virtually all the beams

tested, having high yield longitudinal reinforcement with good

bond qualitiea, failed in shear in one of the following modes:

1. Diagonal tension

2. Shearing (Shearing of concrete above a shear crack)

3. Shear-compression

4. Web crushing

5. Shearing off of flange in T-beams

Consequently the proposed equations take account not only of

the failure loads but also reflect the actual behaviour and

modes of failure of the beams; furthermore the degree of in-

fluence of the various parameters is consistent with the ex-

perimental evaluation of their effects.

In sections 7.1 to 7.6 the derivations of the proposed

equations reflecting the close relationship between overall

behaviour, mode of failure and influent parameters, are pre-

sented for beams of normal proportions (ratio of shear span

to effective depth > about 2.5 to 3.0) for which the para-

331
 meters have been investigated over an extensive range. Section

7.7 is concerned with some detailing requirements necessary to

ensure the applicability of the proposed equations and any

design recommendation derived from them. The behaviour of short

beams is examined in section 7.8 in terms of a more general

discussion on moment-shear interaction.

In an attempt not to interrupt the logical sequence of pre-

sentation of the various equations, the detailed physical

interpretation of the proposed shearing equation, involving a

new failure criterion is presented separately in section 7.9.

Finally in section 7.10 a brief summary of the conclusions

relevant to this chapter is presented.

7* 1 Behaviour of beams failing in shear

It is generally agreed that shear. distress starts with the

formation of inclined These usually in

cracks, cracks appear

the middle of the shear independently by the

span, either or

inclination of an existing flexural If

crack. shear reinforce-

ment in the form of vertical stirrups is provided then its pre-

sence does not affect the shear cracking load. This is con-

firmed by the fact that only negligible strains exist in these

stirrups up to that load and the litte of concrete parallel to

their diroction remains constant. If inclined stirrups used

are

it is possible that some of them will cross the flex-

existing

ural cracks and become stressed before shear cracking. However

the widths of the flexural cracks prior to shear cracking are

quite small in the region in (< 0,00211)

question and the

strains in inclined stirrups although appearing at an earlier

332
load stage are very small till shear cracking occurs. It can

therefore be assumed that the beneficial effect of inclined

stirrups on shear cracking is quite small and can be neglected.

The question of calculating shear reinforcement thus arises

when a member is required to resist a collapse load greater

than its shear cracking load.

No general analysis valid for all load stages can be made

at present. Attention is therefore concentrated on two critical

load stages shear cracking and shear failure. The simpli-

-

fying assumption that can be made at these stages are separate,

but they have one important characteristic in common in that

both the cracking and the'failure occur in the concrete.

7.1.1 Shear crackina of reinforced concrete beams

Recent research reported in chapter 2 has shown that the

following components resist shear prior to shear cracking:

1. Shear in the uncracked zone above flexural cracks

2. Interlocking of aggregate across flexural cracks

3. Dowel action by the main reinforcement

4. Forces in any inclined stirrups

In most cases only the first three arise and the last one

can normally'be neglected in view of the discussion in the pre-

ceding paragraph. The other three are in some measure dependent

on the concrete strength and the ratio of main reinforcement.

A thorough analytical treatment of the three actions does not

seem to be possible at present, mainly because of the uncer-

tainty concerning the extent of interlock action and to a

lesser degree that of the dowel. The question of dowel action

has been the subject of many tests and research on it is dis-

333
cussed in chapter 2. Although no thorough theoretical analysis

of the problem exists, the shear force carried by dowel action

has been evaluated to be about 20 to 30% of the total shear

force. The question of aggregate interlock, also discussed in

chapter 2, has received very little attention. The small widths

of the flexural cracks at shear cracking loads must allow an

important transfer of forces to take place through aggregate

interlock. It would thus be quite erroneous to consider that

all the is resisted by the uncracked zone above flexural

shear

cracks. It remains however that it is in this region that the

inclined crack extends and causes a sudden diagonal tension

failure either the load stage at which it appeared, or for a

at

very slight increase in load, provided of course that no shear

reinforcement Defining shear cracking load as the load

exists.

at which a beam of normal proportions (shear span to effective

depth ratio 2.5 to 3) would collapse in diagonal

approx.

tension if it had no shear the shear cracking

reinforcement,

load Qcr can be expressed as

Qcr
h qcr (?. 1.1)
bd .............

on the assumption that the influence of concrete strength and

main steel ratio on all three actions are similar, since an in-

crease in steel increases the area of the compression zone, de-

creases the width of the flexural cracks and increases the

the dowel (and thereby the length

stiffness of of concrete

supporting the dowel),

axis/
In the above expression 'n' is the neutral factor at the

section in which the shear crack enters the uncracked concrete

334
zone and qcr the average shear stress at failure in that zone.

Because of the importance of the dowel and interlock actions

only an approximate attempt at establishing the degree of in-

fluence of the parameters can be undertaken. The solution must

then be compared with the dogrooc of influence derived from the

tests.

Since the difference in flexural neutral axis depths is not

very significant beyond an early load stage, n can be taken

approximately equal to no the neutral axis factor when the

its limiting compressive strain prior to

concrete reaches

yield of the main steel. The value of qcr must reflect a failure

criterion appropriate to this case where the direct stresses

in the part of the compressive zone concerned do not reach

very high values. Such a criterion is the one developed by

(93)
Cowan and verified by Reeves for direct stresses up to

0.5 Ucyl. According to this criterion, failure occurs when the

principle tensile stress reaches the limiting value ft =

= l. 02 U
cyl j...ýa,,12
1
2 L 2
where ft _, + , ý,

and cr = direct stress

Assuming a uniform distribution of shear stresses in the

compressive zone as was done by Walther(119) and is similar

in effect to Regan's approach in which the applied moment does

not affect the diagonal tension failure - as is the case in

these beams whose shear cracking load is independent of the

a/dl ratio - failure can be expected to occur when

2

qcr = 1.02 UcylU

The shear cracking loads of equivalent rectangular beams and

335
 T-beams of same web breadth were found to be the samet so in

the discussion that follows the case of a rectangular beam is

envisaged. The definition of no leads to the following deri-

vation of its value

1-n
&st =
no c

Fcc =b no d1 k1 U0 1
1-n
o Est rst b d1
Fst
-nF, c
0
Gc = 0.0035 (the
equating FSt and Fcc and taking generally

accepted value proposed by the CEB)

no2 0.0035Estrst

1-no kÜcl

n 2
The expression in is equal to 2.63 nos
approximately
0.0035 Est rst 0.4
n=
0 2.63k
1Ücyl

Qcr [0.0035Estrst ý-0.4 2

d x 1.02 Ucyl
283klU
-I --
cyl

=(ki0)0'4 (100rst)0.4 Ucyl0,27

4 or approximately

Qcr J370
100rSt Ucyll
bd
1k1
using the CEO proposed value ki = 0.67 which is a rather con-

servative value

(1QOrst)
är1
bQ
(Ucyl )s ............. (7.1.2)

Because of the many assumptions and simplifications in-

volved, this expression can only serve as a guide to the formu-

lation of an empirical equation. Comparison with test results

has shown that equation (7.1.2) gives good agreement with

336
experiments as it stands, both for rectangular and T-beams,

and provides a safety margin of 15 to 20%. It must be con-

sidered as a basically empirical equation and it would not be

appropriate to try to reduce this safety margin until more is

known about the exact contributions of dowel and interlock

forces. Furthermore the sudden nature of the failure in this

case warrants safety margins greater than those associated with

more gradual failures.

It can also be noticed that the contribution of the concrete

in the final expression is in terms of Ucy14. This dependence

of the shear cracking strength on the cube root of the cylinder

strength with relevant test results reported in chapter

agrees well

chapter 4. A comparison between tensile strengths of concrete

obtained in uniaxial tensile tests and those calculated using

function of cylinder crushing strengths in terms of Ucyli and

Ucyl is presented in fig. (139) where it can be seen that al-

though z
Ucylý is a better lower bound fit to the experimental

points the proposed function of UcYlý would take better account

of the favourable distribution of stresses due to the greater

ductility of weak concretes and of the unfavourable distribu-

tions caused by the brittleness of very strong concretes.

As has been mentioned previously no significant difference

was found between the shear cracking load of equivalent rectang-

ular and T-boams so in the case of T-beams the percentage of

main steel appearing in equation (7.1.2) should be taken in

terms of the web breadth only. Because of the fact that greater

amounts of steel may be used in T-beams and particularly in I-

beams, than in a rectangular beam with the same web breadth

337
 [
, ...
r, r 1LI F) I 1
F) - - - -t---

- - -, - - -- L --- -- -
f
L ± J !
---

__ _
C,
) I

400

vi,
r.
r
:: J0

c_
4J

C
k2

200

10

ýý" ý1 6 v000 7000

1''ýý'ý
ýý ':eoc)) 5000
CcrT"Qr335 IV3 ý trenc'th
J
;rp. s. i.

F Iý. (1.ßr;; i la tion bct\veOfl Uli tensile n,-', -Z Z-:3ive

d d, Cif rl"na-A
y ý J-ýýiJ
ý.,
l1ýi.:
the value of this ratio of main steel based on the web can

become very high in these cases. The increased effect of the

main steel on shear cracking in these cases is very small or

nonexistent beyond a certain point so an upper limit should be

placed in equation (7.1.2). Based on the similarity of nominal

shear stresses at shear cracking loads between T-beams with

high bsd values but whose shear cracking load is still pre-
r1
dicted in manner by equation (7.1.2) and I-
a satisfactory
A
beams with much higher - values, this upper limit has been
r1
found to be 12.5 (Ucyl) The final form of equation (7.1.2)
.
for all beams is therefore.
loo a
(Ucyl st lbrdl (Ucyi)* (7.1.2b)
8 12.5 brdl ...
. b--- .G
L rl J

The approximate equivalent expression in s. i. units is

Rcr 100 A
-
JUc
0.3 stl 0.45 (Ucyl3
brdl 1d
Yr1

where and U are expressed in N/mm2

b cy1
rd1

7.1.2 Shear failure

If the shear cracking load is to be exceeded shear reinforce-

ment is required for the beam to reach its full flexural capa-

city. If the shear reinforcement and web sections available

are insufficient for the full flexural capacity, but suffi-

cient to prevent a diagonal tension failure, then a failure

due to shear will occur in one of the following modes:

1. Shear failure of the concrete in the compressive zone

2. Compression failure of the compressive zone, the area

of which is reduced by shear cracking

339
 3. Web distress

The exclusion of other minor types of failure (dowel, bond

etc. ) none of which occurred in any of the tests, assumes the

provision of minimum shear reinforcement and satisfactory de-

tailing of the main steel. What these requirements actually are

is explained in section 7 of this chapter. The question of

the shearing off of flanges in T-beams is a very particular

problem and is therefore examined separately.

The first two modes of failure are discussed in terms of

equilibrium at an inclined plane in the sections immediately

following. The third mode is treated separately in a subsequent

section.

7.2 Equilibrium Conditions at an inclined plane

7.2.1 Behaviour of beams following shear cracking

Shear cracking in a beam is manifested by the appearance of

inclined cracks. When these inclined cracks are formed in a

region of high moments they appear first in the tension zone

and then propagate towards the compressive zone. If they are

formed in a region of low moment then they usually appear near

the neutral axis and propagate in both directions. In both

cases these cracks reach the level of the main steel almost

upon formation and if no shear reinforcement is present failure

by diagonal tension occurs at this or a slightly increased

load. When the presence of shear reinforcement prevents a

diagonal tension failure then these cracks continue to propa-

gate and to widen, their width exceeding more or less quickly

that of the flexural cracks.

340
 One of these cracks usually becomes predominant at some

load stage and. close to failure the condition of the beam is

characterised by the presence of a principal inclined crack

that divides the beam into two parts.

Observations made in the course of the tests indicate that

failure of the beam can occur in two different ways

By rotation of the two sides of the beam around the com-

a)

pressed zone whose area becomes reduced and which crushes

in a manner similar to flexural compressions producing

horizontal cracks (Shear-compression)

small

b) When a rotation as previously described does not

such

take of the two sides of the crack

place(the movement

being failure occurs by the

more of a vertical nature)

the zone in an inclined direction

shearing of compressed

from the inclined to the top surface of the

shear crack

beam. (Shearing)

These two modes of to the rotation

failure and their relation

or lack of rotation at the principle crack have also been

)
observed by Borishanski(1? in the case of both rectangular and

T-beams and in shear spans similar to the present ones. In the

case of T-beams which do not normally fail in shear compression

(35)
Evans has also observed that diagonal cracks produced an

vertical movement. A numerical example of the

appreciable

movement at an inclined crack in a T-beam is provided by Leon-

hardt and Waltheris(68) beam TAll whore it can be seen that the

displacements are much more important than the hori-

vertical

zontal ones see fig.

(140)
-. A close up of Regan's beam 66 in
-

fig. (141) shows how vertical the displacement was at the crack,

341
 rP
8
3
Q
N
3
s
fi

s'
J

0
0
3
El

100 60 20 R1

Horizontal movement in 1/ 1JO mm

Fig. (140) Prcgressive, opening of shear crack in' aT- beam with
vertical stirrups fcilin3 in shear. Beam TA11 by Leonhardt and WaltherGB
Failure load P: 68.4 tonnes.

i
i
 ý:
:uxýC; r.;ý

`J

8 fp

5
z. i

NIP

z -.
,"1

'1.

Firj. 141: Ileitical ýnoVu,(-riL ,t Shý2ýtr Crack

Beam B6 by ýcogan(94)

343
the stirrup having remained virtually vertical inspite of the

large crack opening.

In order to determine tho conditions under which failure

occurs it is necessary to study the conditions existing at the

critical inclined crack on which failure takes place.

7.2.2 Equilibrium conditions at an inclined plane

Referring to fig. (142), the forces acting on an inclined

crack plane and its projection to'the compressed surface are:

1. Fcc the longitudinal force in the concrote compressive

zone

2. Qc the transverse force in the concrete compressive

zone

3. F. the interlock force inclined at an angle to the

horizontal

49 F forces in the shear reinforcement crossing the

aq
crack plane and inclined at an angle CN' to the

horizontal

5. Fat the horizontal force in the main tensile reinforce-

ment

6. Fsc the horizontal force in the compression roinforce-

ment

7, Fd the dowel force in tension reinforcement

The effect of the dowel action can be assessed from the

following considerations.

In beams containing shear reinforcement the effect of the

dowel action in the main steel takes the form of a transfer of

forces to the stirrups situated outside the lower end of the

344
 5C

ICc
voc

cc I?
sc

crack tergth 'c

Fig. (142) Equilibrium conditions at an inclined plane

ý_ý
principal inclined crack. Usually only the stirrup close to the

crack becomes significantly stressed (see sections 2.9 and

4.9.1) although in some tests (section 5.10.1) more than one

,

stirrup can be affected by the dowel action. This transfer

of forces to the stirrups by the dowel action of the main

(l2)(9)
steel is confirmed by dowel tests. These dowel tests

however have also shown that very large vertical displacements

(larger than those occurring in normal beams) are required in

for force be
order a dowel of any appreciable magnitude=to

developed, with the exception of the force that can be trans-

ferred to one very close to the start of the crack. A

stirrup

to the effect of the dowel action

reasonable. approximation

thus be achieved by taking into consideration the effect

can

of the force in that stirrup. This force is, in effect-automat-

ically taken into account if the number of vertical stirrups

crossing a crack is taken as c/s where c is the horizontal

projection of the crack length and s is the spacing of the

stirrups.

The equilibrium conditions at the inclined plane are thus:

F+ Fcc Fi cos = Fst + Fsl

cos b1 (7.2.1)
'` . o...
Sc
q.. qc+F1 sin Esq sin o< (7.2.2)
............

FM Ewhoro+ +SF z+ Fiz. ........ (7.2.3)

sc Sc cccc sq sq
tho 'z's ar 9 the lever arms of the respective forces

about the centroid of the main steel in the vertical section

through the crack head, and m is the external moment about the

same pvi't.

Equations (7.2.2) and (7.2.3) represent the two modes of

failure described in the preceding paragraph. Failure criteri

346
can be formulated for the modes of failure implied by equations

(7.2.2) and (7.2.3) and expressions for the ultimate load in

each case can then be established. If equation (7.2.2) governs,

the failure will be of the shearing type, while a shear-com-

pression failure will occur if equation (7.2.3) is the governing

One.

7.3 Shearing mode of failure

This is the type of failure dealt with by equation 7.2.2

and by section 7.2.1 (b) and involves primarily transverse

deformations. If the (9 is large any transverse

crack angle

displacement across the crack brings a considerable interlock

force into action and this causes a flatter crack to appear at

increased load stages see fig. 143, fig. 68 of section 5.1

-

and section 2.9.

Progressively flatter cracks may be formed until a stage is

reached at which the interlock is insufficient to produce

further cracks, Due to the large opening and flat angle of the

last crack produced the interlock across it is smell and can be

neglected in calculations.

Strain measurements on stirrups during these tests have

shown that shear reinforcement crossing the critical crack

yields prior to this type of failure.

In view of the above, equation (7.2.2) becomes

Qult Qc +Z ßq fyq sin pc ...........

where: Aq is the cross sectional area of one member of the shear

reinforcement

fyq is the yield or proof stress of the shear reinforcement

347
 latcr crack Original
I1 crack

Fig. (143) Progressive flattening of crocks

011 simple beam

____ __
I

continuous beam

Fig. (144) Support of shear by inclined thrust

i '/

Fig. (145) Crack pattern in a beam with significant compression

reinforcement and light stirrups

!54!.
 During tests of beams failing in this mode it was observed

that in rectangular beams the critical crack generally rose

higher than in cases of shear compression failure, while in T-

beams the crack did not enter the flange until failure was

imminent. In view of these observations, for the purpose of

determining the amount of shear reinforcement crossing the

crack, the height of the crack can be taken as 0.75 d* in

rectangular beams and as (d*-t) in T-beams: the symbol d* re-

presents the depth from the compressed surface to the anchorage

of the reinforcement in the tensile zone - i. e. to the

shear

lowest layer of main steel when stirrups are used and to the

layer from which bars are bent ups when bent up bars are used

as shear reinforcement.

The difference in behaviour between beams having vertical

and inclined shear reinforcement is reflected in the different

inclination of the failure crack and consequently these two

cases will be examined separately.

a) vertical stirrups

From the test results of various types of members failing in

this mode it has been established in chapters 4 and 5 that the

final crack angle in all cases is approximately equal to arc

tan 0.50.

When the shear reinforcement consists of vertical stirrups

equation (7.3.1) becomes

Quit Qc + rgfyq brc (7.3.2)

' ...........
Aq
where rq is the shear reinforcement ratio
brs sin '<

c is the'horizontal projection of the crack length

349
and s is the spacing of shear reinforcement measured horizontally

Expressing the horizontal projection in terms of the height

and inclination of the crack, equation (7.3.2) becomes

Quit ° Qc + 1.5 br d* rgfyq (7.3.3)

...........

for rectangular beams

Quit Qa +2 (d*-t) br rq I' (7.3.4)

' ........

for T-beams

where t is the thickness of the flange (in compression)

0 Inclined stirrups

If we examine rectangular beams with shear reinforcement

inclined at 450v the inclination most commonly used in practice,

the number of stirrups crossing the failure crack is approxi-

mately

c+0.75d*
s
the force F in one member of the shear reinforcement is
sq
FS9 Aq fyq
=

and its vertical projection Fsgv

Fsgv = Aq fyq sin 45°

The total vertical force vq on the stirrups crossing the failure

crack is thus

(c + 75d*
0s.
vq _ Aq fyq sin 45°
*) b
v (cam Af (sin 45°)2 r
qs sin 450 q yq b
r

but A
yq
,r
bssin450
r

vq = 0.5 (c + 0.75d*) br rgfyq

the tests 0
In all of beams with stirrups inclined at 45

failing in this modo the inclination of the crack was found to

350
be approximately the same and equal to arc tan 0.33 - see section

5.6 - which in rectangular beams gives a crack length of about

2.25d*

vq = 0,5 (2.25d* + 0.75 d*) br rgfyq

1.5 brd* rqf

This is similar to the stirrup contribution term of equation

(7.3.3)

In the case of T-beams the expression for vq becomes

vq = 0.5 cc + (d*-t)J br rgfyq

crack length c= = 3d* - 3t'.

0*33

v 0.5 (3d* 3t" + d*_t) br rgfyq

-

2(d*-t)bri'I'

which is also similar to the stirrup

contribution term in equation (7.3.4)

Although no tests were carried out with inclinations other

than 45o it is reasonable to assume that similar results would

be obtained for intermediate Equation (7,3,3) and (7.3.4)

cases.

can therefore be considered applicable for any stirrup in-

alination botween 900 at 450 provided rq is taken as

A

brs sin c

The case of bent up bars which are normally used only to-

gather with vertical stirrups, ie of course much more complex

as the inclinations and crack lengths are affected differently

by those two types of shear reinforcement.

A reasonable approximation to an exact solution, which must

be based on crack lengths and inclinations having character-

istics intermediate between those intrinsic to each type of

reinforcement' can be obtained by considering separately the

351
action of the two types of shear reinforcement in the
- which

case of bent up bars involves taking the number of bent bars

actually crossing the crack - and adding their individual con-

tributions to the shear resistance. Use of this method has

produced adequate agreement with the test results of reference

58. (Applied to the 19 T-beams up bars

with stirrups and bent

reported in that reference this approach and the completed

formulae for T-beams developed later in this section give a

mean
Qe-a-1= value of 0.95 with a 0.12 coefficient of variation)
test

The further development of these equations (7.3.3 and 7.3.4)

which must take into account the failure criteria for concrete,

and the eventual effect of compression reinforcement will be

undertaken separately for the different types of member con-

cerned, As mentioned in the introduction to this chapter the

detailed physical interpretation of the shearing equations in

rectangular and T-boams in terms of the proposed failure cri-

terion is presented in section 9 so as to allow the uninter-

rupted presentation of the various equations in an orderly

sequence.

7.3.1___5ingly_reinforcod rectangular beams

The shear resistance of the concrete Qc depends on the area

of concrete resisting shear and its ultimate unit resistance

to shear.

Adopting Lichtenberg's(73) variant of a Coulomb type failure

criterion - see section 7.9.1 -' applicable to a type of failure

that is of the "plastic" shear mode rather than of the brittle

tensile model Qc can be expressed in terms of the main steel

352
 100A
ratio (rstý st ), the effective depth to shear span ratio
- bdl
1)
( and the properties of the concrete - see section 7.9.2.
a
Expressing Qc in terms of the relevant parameters with the

degree of influence arrived at in that section

d
Qc =k (rst Ucyl)s bdl (7.3.5)
. äl . ...........

The value of k has been determined by analysis of experi-

mental results and can be taken as 12.5 when units of ibf and

inches are used.

The final equation is thus

dl
(rst Uc, bdl (7.3.6)
Qult = 1.5 bd* rqfyq + 12.5 .a' ..

the expression in s. i. units is

approximate equivalent
d ý
+ 0.45 (r Uoyl) bd1
Quit ' 1.5 bd* rgfyq st " äl "

It be pointed out that when the above value of Qc

should

is used in a neutral axis factor of 0.25 to 0.3

connection with

high shear stresses result and an explanation must be given

concerning them. The way in which transverse forces in the con-

creto are supported in a rectangular beam is primarily by the

main compressive force sloping towards the main steel in

regions of decreasing moment - see fig. 144 and also section

7.9.3. The slope of the main compressive force is more pro-

nounced when light shear reinforcement is present - see Appen-

dix A. 2 - which is also the case in which this mode of failure

is likely to occur. The resulting tensile stresses are thus

QC
than the apparent unit shear stress ---
much smaller ' where
bnd
linevitably
n is the neutral axis factor; furthermore there is

some transfer of forces by aggregate interlock in the upper part

of the failure crack - see section 7.9.1 Thus high values of the

unit shear stress are supportable.

353
7.3.2 T-beams

In T-beams only the area of flange adjacent to the web can

be considered to b_. operative in resisting shear. Tests of

T-beams with varying geometry have shown this area to be a

function of the flange thickness and web breadth. As the exact

distribution of stresses in the flange is not known it

shear

is desirable to establish the value of this effective area

so as to take account of this distribution. Assuming an average

in the of the flange, equal to

shear stross' effective area

the tensile of the concrete expressed in terms of

strength

the the the area involved in

cube root of cylinder strength,

the total ultimate shear force, minus the shear

supporting

force taken by the stirrups crossing the failure crack and

stressed at yield values, was found to be approximately equal

to t(b +6) in all the T-beams tost©d, A constant value should

r
of course be expected in view of the constant inclination of

the failure the fact that it dons not enter the

shear crack,

flange until failure is imminemm t, and the preceding dis-

cussions on dowel and aggregate interlock actions.

The effective area of the flange resisting shear, which

takes account of the distribution of shear in that flange is

thus (br+6) (7.3.7)

of =t ...........

Measurements of concrete strains in T-beams have shown that

the entire width of the flange carries compression right up to

the limit state of collapse. Furthermore these measurements

have that the loss of compression at increased load

also shown

in sections progressively closer to the load appli-

stages,

is pronounced in the region above the

cation point, much more

354
web-than in the rest of the flange. The compression in the area

t(br) is therefore a relatively small part of the total and it

is only this part of the compression that can slope down into

the web. This question is dealt with in more detail in section

t
7.9.3 but it can be appreciated thatevertical component of the

force over the web is a smaller percentage of the shear force

than the equivalent force in a rectangular beam. The remainder

of the compressive force slopes downward only within the flange;

this has also been verified by strain reading at symmetrical

points in the underflange and top surface of the flange - see

figs, (118) and (119), section 5.10.3. It follows that the

tensile stresses involved in carrying the force Qc are more

significant in a T-beam than in a rectangular one. In all

practical cases it is these tensile stresses that are res-

ponsible for failures of T-beams (the great majority of

shear

singly reinforced rectangular beams fail in shear-compression,

if they fail in shear at all).

Shear cracks reaching the web/flange junction do not rise

into the flange until failure. It is important in the dis-

cussion that follows to keep in mind the fact that failure

occurs as soon as these cracks enter the flange:

Prior to failure longitudinal stresses at the level of the

underside of the flango in the quattor span nearer the load are

low, so an almost pure shear condition exists there. The ex-

pression for Af takes into account the distribution of shear

stresses within the flange, and the principal tensile stress

can be approximated to the nominal shear stress, in which case

the critical stress at the web/flange junction should be ex-

355
 pected to be equal to the tensile strength of the concrete;

this is consistent with the derivation of the effective area

Af so long as it is remembered that the area actually stressed

is greater than Af and the true average stress in it is less

than Ut.

Local failure will thus occur when

QC
25 (Uoyý)ý (7.3.8)
tb = Ut = approx. .......
+6
r
Again here the cube root form takes account of the more

even distribution of stressas obtained in weaker concretes

due to their ductility and the unfavourable distribu-

greater

tion by the brittleness of very strong concretes. Its

caused

use reflects quite adequately the influence of concrete strength

in the tests where this parame-!: c:i was varied.

The failure equation for a T.-beam thus becomes

pult (d*_t) br rq fyq 25 (Ucyl)* t (br+6)..: (7.3.9)

-2 +
in ibf and inches

the approximato equivalent exproosion in s. i. units is

Quit 2(d*-t) brrgfyq + 0.9 (Uoyl)* t(br+l50mm)

`

Applied to the test results of the 30 relevant T-beams

where a groat number of parameters was varied, this equation

Qcalc
Qcal
gives an average value of 0.91 with a 7.5% coefficient
Rtest

of variation. The extent of the safety margin thus obtained

will be discussed in more detail in the conclusions to this

chapter.

7.3,3 Beams with compression reinforcement Rectan ular

In a manner similar to the flanges of T-beams1compression

reinforcement could be expected to take up a part of the

356
longitudinal compression without allowing it to slope, and thus

not allowing it to make any positive contribution to shear

resistance. Although this may be true at earlier load stages

it does not seem to be a sufficient explanation of the behaviour

at collapse.

The conditions at the head of a shear crack, where com-

pression steel is present in appreciable quantities, are com-

The shear crack seems to rise higher than in an otherwise

plex.

similar beam without compression`reinforcemont, and vary

locally the probably carries shear by a sort of dowel

steel

action. However the length over which bars can support shear

forces be put
in this way is very limited, and the must soon

back the So far as overall behaviour is con-

onto concreto.
being to the
cerned the concrete can be viewed as subjected

same shear as in a beam with no compression steel.

If the dowel force is transferred as described above and

if the continued to carry longitud-

compression reinforcement
y
inal forces, analogy with T-beams, the shear resistance of

the be expected to be lowered due to the re-

concrete could

duction the on it. However at loads close to

of compression

failure, the presence of high bond foces9 indicated by the

appearance of bond cracks at the level of the compression

(145) that the

steel - see figure - suggests most of com-

pression is transferred back to the concrete just behind the

head of .the crack. In chapter 4 section 4.7 it can be seen

that release of bond in these cases has also boon reported by

Wilby(122) and that in the test of beam R16, representative of

the typo of beam with heavy compression reinforcement failing

357
in shearing, the strains in the compression steel show a ten-

dency to decrease at loads close to failure. This produces

inclined thrust conditions approximately similar to those

in singly reinforced beams.

In view of the above, equation (7.3.6) derived for singly

reinforced beams applies equally to beams with compression

reinforcement,

7.4 Shear-compression mode of failure

This is the type of failure dealt with by equation 7.2.3 and

by section 7.2.1(a). Since in this type of failure, as described

in section 2.1 and indicated by fig. (146), the movement is one

of rotation at the cock, not involving important displacements

parallel to the crack, the interlock force may be neglected

and equation 7.2.3 can be simplified to

(7.4.1)
m= Fsczsc + Fcczcc +ZFsgzSq ...:.....

Strain measurements on stirrups in tests of beams failing

in this mode have shown that all stirrups crossing the failure

crack yield prior to collapse, thus simplifying somewhat the

third term of equation 7.4.1 although not completely as its

contribution to the shear resistance depends also on the crack

length which in this case is variable. Evaluation of the other

terms and of the relevant crack lengths that enter into equation

7.4.1 is made separately for the various types of beams con-

sidered.

7.4.1 Singly Roinforced Rectag9ular Beams

In this case equation (7.4.1) can be written as

M= hnd1 k1 Ucyl (1-k2n) dl (7.4.2)

+j_FSq zSq .......

358
 P (86 1,
/* o ultimate)

Iiý! 4i t
t Beam
I, Iitt R-17
t1II1I i13

(67°l0 of ultirnate) P

1Iº1 Beam
º111 1II
10,72 R- 22
1I1 1200 Iýl
20
10 20
ºIII_, II

(92°/a of ultimate) P

I Beam
R-10
1iIf

IIIIII

((T
Fig. (146) Crock widths (in 0.001") showing the rotational tendency of
inclined crack opening in shear-compression failures

ZJ5w
.
where k1Ucy1 is the average stress in the compressive zone

with a fully developed stress block

and k2nd1 is the distance from the extrem© compressed fibre

to the centroid of the compressive force in the

concrete

The first term on the right hand side represents the moment

of the concrete force balancing the force in the main steel at

the bottom of the crack. When inclined shear reinforcement is

present, the total concrete force is increased by the horizontal

component of the force in the shear reinforcement. The moment

of this force can be accounted for in the second term by adding

to it a moment%(Fsgcos c<) 0.9d1.

Various methods have been proposed for the solution of shear-

compression failures, but as discussed in the introduction to

chapter 2 whore those theories are appraised, the one proposed

by Regan seems to satisfy not only the actual mode of failure

at the head of the critical crack (crushing of the concrete) as

do some of the other theories but also reflects the general be-

haviour of the beam as expressed by the variable length of the

critical crack. A variable crack length his. also been proposed

by Borishanski and by Hillerborg but their concept of the

criterion of shear failure is one of transverse force while

in the cases of shear-compression failures it is the longi-

tudinal forces, witnessed by the very high compressive strains

measured in the relevant tests of chapter 4y that bring about

the final failure of the compressive zone.

In Regan's analytical method for determining the ultimate

neutral axis depth in shear-compression failures the value of

360
 In' is established in terms of the percentage of main steel,

the length of the shear crack, the concrete strength and the

shear span to effective depth ratio. In chapter 4 it was seen

that the tests reflect the dependence of the critical crack

lengths on the various parameters in a manner consistent with

the tendencies implied in the theory. In the light however of

some differences between test results and that theory it was

decided to modify the latter appropriately. A detailed sub-

stantiation of this theory and the proposed modification ie

presented in appendix A; briefly the three points on which this

theory was modified or extended are the following

1. The basic value of the coefficient k1 reflecting the

average compressive stress in the compression zone was increased

in view of the test rosults and the consideration that there

is no absolute need for the longitudinal stress to be zero at

the crack, and different values are suggested for different

concrete strengths - see Appendix Ael.

2. The contribution of shear reinforcement in tho compati-

bility equation was taken into account by considering the actual

varying slope of the compressive thrust.

1) slope of compressive thrust

used by Regan Q
-ý ----ý F
cc
ýý,
2) varying slopo of compressive
ýý thrust
1 q-(c-x)q fY9b
-'f
F
:ý cc
ýý
w .._ ._....... ý. ___.. see Appendix A. 2

361
 3. The analysis was extended to inclined stirrups. An

approximate equation that takes account of their effect was also

developed in terms of the degree of shear reinforcement, the

Qdl
concrete strength and the ratio see Appendix A. 3.

Taking (dl-k2nd1) ti 0.9d1 a numerical simplification of

(95).
that theory can be made Including the proposed modification,

the resulting expression for the moment F z can be given in

cc cc
the form:

rii (7.4.3)
Fcc-cc - qd 'c. mSbd12 .......
1

where tho ratio is of the external actions across the

QId-
1
crack line and its projection to the compressed surface, and
i
mS 27 (rst Ucyl) (7.4.3a)
. .......

in lbf and inch units

The shear compression failure equation becomes:

d
ý. (7.4.5)
i1u1t ` Qd ms bd12 +1Fsgzsq .......
1
This equation is a general one and applies to all types of

shear reinforcement. It can be applied to any section of a

beam and to all possible cracks with heads at the section con-

cerned. In specific cases it can be simplified as follows.

7.4.1. a Closely spaced vortical stirrups

Close spacing is dofined as being such that the number of

stirrups crossing the critical crack may be taken as c/s. This

treatment of the stirrup force is apparently non-conservative

but if a stirrup lies just beyond the tail of a crack it is

brought into play by the dowel action of the main steel, If the

strength along a plane between stirrups is less than the strength

calculated assuming close stirrups then this lesser strength

be presumed to be critical (see 7.4.1. d below).

must

362
 Assuming close spacing
2
F zsq =cb rq1' rgfY9 b2....... (7.4.6)
2=

Substituting this value into equation (7.4.5)

d2
2+ c2
f1ult b (7.4.7)
- Td 0ý " msbdl rgfyq ...
1"

The value of c corresponding to a minimum moment of resist-

ante at a given crack head can be found from the condition

Ü
ult
= 0. This condition defines the value of lot corre-
ac (minimum
spending to the plane of minimum strength Multi.

Performing this differentiation and resolving the resulting

equation the following expression for the critical crack length

is obtained.

1-7, ms'
dI...... (7.4.8)
f Qd r..
l Yq
Substituting this value into equation (7.4.5) and expressing

the results in terms of Qult°

)* 1ý
mss (r (4d
b1=2x qf yq Pil

finally

n
1t lý
(Ucyl (rgfyq) (7.4.9)
bd = 13.5 . rst3)s ...
1

The approximate equivalent expression in s"i. units is:

Qult s adl
115 (Ucyl rst (rgfyq) (r)
bd y
1
7.4.1. b Closely spaced inclined stirrups

In this case the

equivalent of equation (7.4.6) is
2 rgfY4 sin o1
Fsgzsq = bdl (äoc + 0.8 cotes )(d + cots-ýC)
2
1
(7.4.10)
.......
In the same way as for vertical stirrups, differentiation of

363
Mult with respect to 'c' gives a solution for c/dl and thence

for Qult' The form of the solution is complicated and both

analytical - see appendix A/3 - and experimental findings show

that the difference between the ultimate loads of otherwise

similar (same rqfyq) beams with on the one hand vertical and

the other 450 stirrups is a modest one. The gain in ultimate

load obtained by using inclined stirrups increases with in-

creasing amount of stirrups and could theoretically raise the

ultimate load by up to 15% in practical cases - see appendix

A/3. This order of magnitude of the possible difference in-

volved is also confirmed by tests by Wilby(122) in which the

greatest increase in strength due to the use of stirrups in-

°
clined at 45 was 17/ol.

In appendix A/3 an approximation to the exact solution

gives the following expression for this increase

00 0- rf Qd
45 90 90 1 24 10910 (167 q yq ( 7.4.11)
Qult = Qult + Qult L T55 U'7 C11)] "'
Y Qd
1
with an upper limit of 0,3 for the value of m to be used in

this equation. Negative increases need not be considered.

7.4.1. c Bent-up bars

,
Treatment of this type of reinforcement, which generally

involves rather wide spacing of shear reinforcement, should be

on the basis of equation (7.4.5) taking into account only such

bars as actually cross any given crack. Dowel action cannot

be assumed to have an evening-out effect. The critical crack

will thus always have its tail so positioned as to just miss

the'bend of a bar. The area of main steel used in calculating

rst must be that at the tail of the crack and must not include

any steel bent up along the length of the crack considered.

364
7.4.1. d Widely spaced shear reinforcement

If the shear reinforcement has a wide spacing, equation

7.4.3 gives the strength of the longest possible plane not

crossed by shear reinforcement as

d
m1 bdl
utt
^ Q-dl .7,, . ms

Actually a value of mS 0.94ms should be used in this case,

the factor 0.94 representing approximately the lack of the

beneficial effect implied in the value of ms, as used hitherto,

of taking into account the contribution of the stirrups in the

compatibility condition.

Qult ms bdl (7.4.12)

whence = C? " .........

where c* is the horizontal projection of the longest possible

crack not crossed by shear reinforcement.

If this value of Qult is loss than that given by equation

(7.4.9) in the case of vertical stirrups, then the unreinforced

plane must be assumed to be critical. This is unlikely to

occur when s sin cwt ý- 0,75d1.

7.4.2 Beams with Compression Reinforcement

When a section contains compression reinforcement, this rein-

forcement contributes to the moment resistance. It is not

possible to define its contribution precisely, since' even if

the ultimate neutral axis factor at the head of the crack can be

calculated, the factor varies from section to section so that

the conventional steel stress equation has no clear meaning.

The effect of the compression reinforcement can be treated

by a multiplying factor'of the form

cs Ucyl fscrsc
ncs
me (7 A 13)
Ucylno
11lco

365
where the subscripts tcol and to, respectively refer to the

presence and absence of compression reinforcement and the

moment n1c = Fcczcc + FsozSo"

In this equation the factors ncs and n0 both refer to ulti-

mate conditions. To simplify the approach we may use elastic

theory values. Any overestimate of the tn' values so obtained

gives a safe underestimate of the effect of compression steel.

The error caused should not be great, since, if the compression

steel fulfills its function, ultimate conditions are near to

flexural ones and the discrepancies in '' values should be

small. In other cases, oven if the error is somewhat larger,

it is not important as the shearing conditions of equation

(7.3.6) will govern.

Equation (7.4.13) thus becomes

s
Mcc tics +k rso
(7.4.14)
'o
01°
c

where k is approximately equal to the modular ratio m and can

10000
be taken as 5+ --'"" Both neutral axis factors are cal-
U .
cyl
culated using this modular ratio.
J(mrst)2
no +2 mrst - mrst
Jm(rt+r0)
2+
ncs 2m(rst+rsc ^ m(rst+rsc)
dl)

d2 is the depth from the compressed surface to the centre

of the compression steel.

Equation (7.4.8) can then be used to predict shear com-

pression strengths with the term ms increased by the factor

1Cs/fio obtained from equation (7.4.14) with the provision that

the value of fsc, when taken as fsc Esc E

^ c(ncsdi-d2)/ncsdl

366
should not exceed the yield or proof stress of the compression

Fc may be taken-as the normal flexural value of 0.35%.

steel.

7.4.3 T-beams

The shear compression equation for a singly reinforced

beam can be applied to T-beams by substituting the effective

flange breadth for flexure as 'b', so long as the neutral axis

lies in the flange. In other cases the relevant area of con-

in compression should in principle be substituted for

crete

bndl. "

In fact, because of the large compressive resistance of

the flange, conditions are not critical in

shear-compression

T-beams, and the strength of T-beams is governed

practical

by equation (7.3.9) for shear or by flexural consider-

either

ations. This. has been the case throughout the extensive series

of T-beam tests reported in this thesis, in which all the

parameters affecting shear strength were varied. The

major

fact that T-beams did not fail in shear compression was, also

noticed. by Ramakrishnan in his extensive series of tests on T-

beams with or without web reinforcement; the reduced dimension

of these beams should not be expected to invalidate this

qualitative observation.

Actually if the area of the flange is small then a T-beam

be calculated as a rectangular having a breadth equal to

could

the web. An anlysis of results and pertinent equations has shown

that in design, when the area of the flange is less than

§(br+6)t, a T-beam could be considered as a rectangle of breadth

br. If the in the flange 3(br+6)t but is

available area exceeds

less than (br+6)t then equation (7.3.9) can be used with the

367
actual area of the flange considered effective. When the

available area exceeds (br+6)t then equation (7.3.9) should be

used as it stands.

7.5 Web crushing

In the previous sections three modes of failure have been

described i. e, diagonal tension, shear compression and

shearing, and appropriate equations proposed in terms of the

stirrups crossing the critical crack and the concrete zone

above it. Consideration is also paid to dowel and aggregate

interlock actions. Neglecting doweling and interlock, the

overall behaviour of the beam, in respect to shear, can be

viewed in terms of the following actions

1. Inclined compression in the concrete above the shear

cracks

2. Tension in the main reinforcement, decreasing towards

the regions of low moment but at a rate slower than

that obtained by flexural theory.

3. Tension in the shear reinforcement

4. Inclined compression in the web struts between and below

the shear cracks.

The methods developed in the preceding sections do not

enable the exact magnitude of the last three actions to be cal-

culated. This can be illustrated by the fact that forces in the

stirrups not crossed by the critical crack are indeterminate

since they cannot always be assumed to yield as is the case of

the stirrups crossing the failure crack. This is confirmed not

only by these tests but also by the tests of a number of other

368
investigators(87)(23)(71). This in turn implies that the slope

of the main compressive thrust (see Appendix A/2), if required,

cannot be determined in sections throughout the whole length

of the beam. Any failure caused by these indeterminate forces

must however be prevonted, or some approximate method for its

evaluation must be developed. In the case of the action of the

main steel, good bond qualities and detailing in accordance

with section 7.7 should prevent any premature failures. In the

case of the main inclined thrust the critical section is at

the head of the principal shear crack and this is taken into

account by the proposed methods. In other parts of the beam

more concrete is available to resist it as far as the rectang-

ular beams are concerned. In T-beams, even though most of the

compressive force not directly above the web stays and slopes

inside the flange, some of it will inevitably slope down to

the web towards tho end zones. However the inclined thrust by

itself is unlikely to produce critical conditions in other

sections because of the greater area of web involved, and the

fact that its slope is less pronounced in beams with higher

degrees reinforcement (web crushing occurs normally

of shear

only in beams with very high shear reinforcement) - see

appendix A. 2. A critical condition due in a greater degree to

thisforce may in fact be possible in the support region, but

in practice this is avoided by a local increase of the web

breadth or restriction imposed by permissible bearing stresses.

It has already been discussed that after shear cracking shear

in a beam is ru, ci, -: tied by dowel and aggregate interlock forces

and through a combination of an arching action, by the main com-

369
pressive thrust above the shear cracks, and a truss action, by

the shear reinforcement and the concrete below the cracks.

Equilibrium conditions for this latter action imply compressive

stresses in the concrete of the web. If the web is sufficiently

thin those stresses can cause failure by crushing of the con-

crete between and below the shear cracks. This mode of failure

is generally referred to as web crushing.

Formulation of Equations for the web crushing strength of re-

inforced concrete beams

7.5.1 Equilibrium conditions

In the presence of shear reinforcement, it is impossible to

determine the ratio of the shear force carried by the web to

the total shear force applied, or the height Ohl of the cracked

zone at every section of the beam. In view of the fact that

in the toots the inclined cracks separating the web concrete

into struts generally reached the bottom of the flange Ohs can

be taken as (d* depth of flange).

-

The condition of equilibrium of a web crossed by shear

cracks can be written to give the following equation for the

force that crushing of the web - see fig. (147).

shear would cause

Qh=
fcw br h sin P cos Cý + qw h sink cos'<

whore 0wc = shear force carried by web at web crushing stage

= total force - shear force carried by flange

b= wob breadth
r
fcw = critical average compressive stress' for web

concrete

Cýý = anglo of inclination of stirrup reinforcement

370
 i,
-- --- fcw \ `'
ý t.
\
h C? -
ý:. ý .'

., f Ifi
0

\ ;i `r`/
t

-j --_ýýi -

5re^es Jti"'ti_°325
in cone ti.<<ý in stirr, i 5

Fig. (147) cc C!, C`'15 1^ Y, '" h

;'
 Q= angle of inclination of compression in web

= approx. angle of shear cracks

q= the force in the shear reinforcement per unit length

of the beam, measured along the line of the mrtn rein-

forcement.

qw can be evaluated by considering the vertical equilibrium at

the lower ends of the stirrups

fcwbr sing ýý qw singt f

Substituting this value into the preceding equation,

Qwc ý-) 1
e tan
brh =f sin cos (7 i+
tan
7.5.1)
cw _

A reasonable and, with the exception of extreme end zones,

slightly conservative approximation is obtained by taking

qwc h
q

Equation (7.5.1) then becomes

q tan
b d'
r
fcw sin Ü cos J1+
F tan t_-< """""
(7.5,2)

In cases where web crushing occurs the angle of inclination

of the web compression is in the range of 30 to 450. The

variation of the compressive strain with G was measured

during the tests and found to be very small; furthermore the

hand side of equation (7.5.2) does not vary much in this

right

range so that one can take 37.5 as a fixed value for G giving

approximately

Q 0.75 (7.5.3)
0.5 fcw_1 + tan
b d* .........
r

7.5.2 Considerations on the stress-strain relationship of

the web concrete in web crushing failures

Because of the biaxial state of strain within the wob

372
1
 Pi P2
b_

'f Pi

_._F,.

rý

n
P2
i r. ..

CYO A

Eý

Fi

P1

I P1

ý..._.. I CY, B, CY C

ý1
tI

P1 f1

i, 2; tc.
 N

L
C.t. C)
O

C7

i0 0

C)
C)

0
80

Fig. (i'O) Typiccl rýý"u!; of RobinEon5 test in 1:ccýin, ý ihr differl, nt

StreSS 5r'cin Clir'ý/'i ýüunc4r ciffe_rer.,... Ic; idin goi cr i': io c

/i
 `ý
ý,

ps M-- W,

Af:ýý.
--700
so .M

NJ . +

ý_ ýý Mob

Fig. 150 : Biaxial tests by Robinson : test arrangement

375
 (concrete in compression in the between but
struts shear cracks

subject to extension on account of the stirrups) it is not

possible to arrive at a fictional value of IfcwI by comparing

measured web strains with uniaxial stress-strain curves for

concrete. Tests at the I. R, A. B. A. Laboratories reported by

(103)
Robinson have shown that concrete struts under biaxial

strain of the type occurring in web crushing failures, have a

greater deformability and a lower resistance than any uni-

axial test specimen.

As a result of a four months stay at that Research Institute

the writer has been able to obtain some of the results of

these tests. The loading arrangements are shown on fig. (148);

the results of a typical series are shown in fig. (149) which

is selfexplanatory. The testing arrangement is shown on fig.

(150). The series on which results are available corresponds

more to the case of inclined stirrups and tests are presently

under way to investigate the case of vertical stirrups.

[email protected] Formulation of Equations fir the web crushing th

streu

of reinforced concrete beams

In view of the consideration presented on the preceding para-

graph it is not yet possible to arrive at rational values of

IfcwI by comparing measured web strains with uniaxial stress-

strain curves for concrete. On the other hand, since it can

fairly be assumed that fcw depends on the quality of the con-

crete and the shear reinforcement present, the equations to be

produced should reflect these facts.

The analysis of the test results carried out in chapter 6

of this thesis has indicated that IfcwI is proportional to the

376
square root of the cylinder strength of the concrete and is

influenced by the ratio of shear reinforcement rq rather than

its spacing or the degree of shear reinforcement expressed as

rgfyq. The reason for the non-dependence on the rgfyq value is

that we are concerned here with a type of failure not generally

involving yielding of the shear reinforcement. Yielding of

stirrups may occur if the stirrups are inclined but the in-

crease of web crushing strength intrinsic to inclined stirrups

is very important and will normally lead to a web crushing

in excess of the flexural or shearing strength. In all

strength

reported tests only one web crushing or rather web disintegration

was reported when stirrups inclined at 450 were used. It may

be assumed, for the purpose of calculations, that if the web

did not crush prior to the yielding of the stirrups, it is

unlikely that it will crush subsequently since, theoretically,

there can then be no increase in the shear force carried by

the web.

taken /1 k and rq
With 'fcwI as k a relationship between

by plotting {< = Qult/0.5 brd* (1+0.75/tan <)

was obtained cyl

against rq for the available results of beams with vertical

known to have failed by web crushing (table 7.5.1).

stirrups

The relationship between 'k' and rq is shown in fig. 150 from

which an expression for k can be derived as

k= 25 +5 (rq%) (7.5.4)
.......

whence bud*
[12.5 (1+--7=) (7.5.5)
° + 2.5(rq%)'( U ..
r
the equivalent expression for s. i. units appears in section
-

8.7 where this formula is compared with the provisions of the

377
70

Qult f br d4

05
Lýanäj

Equation (7 5.4) : value of K-

KEY
V Robinson NR, NT
A Robinson BQ
O Leonhardt
0 Plcccs

-.
ýý

ratio of shear reinforcement LJ °lo

Fig. (151) Relationship between wcb crushing strength and

geometrical percentcge of sheer reinforcement

J/8
 KEY
V Robinson a_ 67ý-3Ö
web crushed
ri Leonhardt --=45°
web crushed
o Placas cc _ 45°
web not crushed
A Robinson cc : 45°
web not crushed

ratio of shear reintorcement rqj WooI

Fig. (152) Influence of stirrup inclination (oc) on the web crushing strength

37 1
Draft Unified Code -.

The correctness of the last torrn as applied to beams with

inclined is demonstrated by fig. (151). It can be seen

stirrups

that the points plotted for beams where web crushing took placo

are all situated above the relevant lines. Any points below the

relevant lines relate to beams which although having high web

stresses, did not fail in web crushing.

TABLE 7.5.1 Beams with high web stresses

Beam Ncyl rq% qult iFailure Type Reference

oe

Uc
ll
W1 90° 4332 2.16 20.04 web crushed PLACAS A.

W2 If 1627 2.16 17.5 Chapter 6

W3 to 6605 2.16 20.5

W5 if 4970 3.25 22.5

W6 " 4920 2.16 18.5

W7 4980 2.09 19.0 'º "

BQ6 90° 4900 2.75 18.3 web crushedl Robinson J. R.

BQ7 If 4660 2.75 18.4 Rapport Explicatif

BQ12 " 25601 2.75 18.3 Commission V Effort

Trenchant CEB session

pleniere de Lausanne

April 1968 (see note 1bp ow)

NR8 90 2070 2.8 19.2 web crushed Robinson J. R.

NT8/ " 2760 2.8 28 Essais a 1'Effort Tranchant

NT8/1 2820 2.8 de poutre a ame mince

. 128.6
NR10 " 3300 3,3 127.7 an baton arme
NT10 " 2350 3.3 24.1 Annales des Ponts at
Chaussees, March-April 1961

90 0 3390 7_.84 25.3 Leonhardt F& Walther R

Tl 11 web crushed
ET4 " 3240 1.03 15 if to Deutcher Ausschuss fur

TAl 'º 2205 1,33 16.7 if it Stahlbeton

TA2 " 2205 0.9 15.0 Heft 152 (Tl), 151 (ET4)
,
{ 156 (TAl, TA2)
and

360
 W
Z

ý\

< ;, \ý J
ý: .1..
i"ý
ý\(I\

,
\I

ýS
ý

rte-..

F. cý.

hJ4.

Fig. Web of a beam having stirrups inclined at 670- 30'

153 : crushing
Beam B04 by Robinson

let
 TABLE 7.5.1 Continued

Beam I Ocyl fr% quit Type Reference

L\ ý IFailure
Ucyl

W4 45 0 4985 2.44 26.7 l PLACAS A.

web not crushed
Chapter 6
T2 45 1 3060 '2.84 38.0 Leonhardt F
web crushed
& Walther R
as above Heft 152
IRO 45 3715 2.95 35.2 web not crushed Robinson J. R.
ITS 3825 2.95 34.2 """ as above
IR10 " 2860 3.1 44.5 " it " Series NR & NT
IT10 3610 3.1 39.4

BQ5/DT 40-140A 67°30z! 5260 2.55 29.3 web crushed Robinson J. R.

BQ4/DT 40-1408 " 5120 2.87 29.0 (see note 1 below;
8Q3/DT 40-140 " X4801 2.36 26.3 if "`

Note 1 Full details of those tests are not yet published al-

though partial reference to them is made in ref. (103) and (31)..

/Thoy were carried out in the I. R. A. B. A. Laboratories, Paris.

A photo of beam 9Q4 is presented in fig, (153) showing the web

crushing of a beam with stirrups inclined at 67°30'. The direction

of the stirrups can be seen in the damaged sections.

7.5.4 Validit of the web crushinequations

Fig. 154 servos as an illustration of the validity of the

web crushing equation (7.5.5) and its relationship the

with

equations presented in the preceding sections.

In that figure the web crushing strength in terms of rq% is

presented for a beam having a concrete cylinder strength of

4000 psi. Also drawn are lines representing the shear-com-

pression and shearing equations for fyq = 90000 lbf/in2 and

382
 Shcaring sfy x`0000 lbf/in2
Equation (73.9. )

300 ý-

Shear compre: 3ion: ý9cOOO Ibf/ in`

O00
Equation (74.9. )

2500 Flexural failure br 2.5"

on --

NC
1
A

'O
.
L

II
web crushing
15pp-/
Equation (7.5.4. )

- Flexural failure on br = 611

,000 ýshearin
shearing ; fya 4000

:f yQ . 42500

500

0 0.5 1 1.5 2 2.5 3.5

ýrq)°! °

Fig. (154) Relationship between web crushing and otf ý?r shear
strength equations

3p=
br = 2.5 in and the shearing equation for fyq = 40000 lbf/in2

and 42500 with br =6. All beams are considered to have the same

(8116mm) and dl
amount of main steel = 10,2"p a/d = 3.5, t=3",

and b= 24".

Considering first the case of fyq = 90000 lbf/in2' it can be

seen that the shear-compression equation is not the governing

at any point. At low rq values the shearing equation governs

while after an rq value of 0.5%o the web crushing strength will

govern, and flexural capacity will not be attained. The flex-

ural capacity is presented for this case in terms of a br =

= 2,59 the web breadth used when high yield steel was used.

It can be seen that, as in the tests, failure takes place long

before the flexural capacity is reached.

When stirrups having fyq = 40000 lbf/in2 are used the rele-

vant shearing equation intercepts the web crushing equation

at about rq = 1.2%. In the tests of beams with fyq = 40000

the highest rq value used was 0.82% (T6, T32, T37) and shearing

failures were obtained as would be expected. It can also be

seen that it would be impossible in that case to obtain web

crushing failure as the flexural strength expressed for the

web breadth of 611 used in that test, intercepts the shearing

equation at rq=0.95% - before the latter intercepts the web

crushing equation. A fortiori a web crushing failure could not

be obtainedwhen inclined stirrups were used as in the case of

beam T-14 which also failed by shearing.

If somewhat higher yield stirrups are used - i. e. 42500,

the intercept between the shearing and web crushing equations

is at a lower rq% value, in this case 1.1%, which is logical

384
in view of the low intercept for very high yield stirrups. It

is not surprising therefore that b3am ET4 (rq=1.03) with not

too different geometrical characteristics from the typical

case reviewed here but having a somewhat higher flexural capacity

in terms of the smaller breadth (4.7") though a less in-

-

creased shearing strength in terms of that breadth since the

term involving br in the shearing equation is smaller just

-

managed to fail in web crushing since it had stirrups having a

yield strength of 42500 lbf/in2 and also a lower quality con-

crete (Ucyl = 3240 lbf/in2)0

7.6 Horizontal Shear and Transverse Reinforcement in T-beam

Flanges

The object of this section is the study of the factors

affecting longitudinal shear cracking and failure at the

vertical junction between the projecting parts of a flange

and the central part of breadth lbr1.

The various equations developed are more particularly rele-

vant to the type of beams tested - simply supported and loaded

by central point loads over the web breadth only. The basis

of the approach however is quite general and could thus be

extended to other types of loading.

7.6.1 Beams without transverse flange reinforcement

If the part (b-br) of a flange is to cooperate in the

support of longitudinal compression - and strain readings

across the flange show that it does -a compressive force must

be introduced into it along the vertical planes at the junction

of the web and flange. The resulting horizontal shear on these

385
planes may be instrumental, in the absence of any transverse

reinforcement in the flange, in detaching the outlying portion

of the flange from the web - see fig. 155. In the absence of

flange reinforcement the flange is detached upon the appear-

ance of longitudinal cracking along the flange. This of course

can occur only if the ultimate load of the beam exceeds the

load causing this longitudinal cracking. Beams T21 and T24

described in section 7 of chapter 5 provide an illustration

of this phenomenon, see fig. 155.

In principle the criterion of cracking of the type in

question should be the attainment of a principal tensile stress

equal to the tensile strength of the concrete. The major

stresses acting at the junction are:

1) The shear stress required to transmit forces to the

outstanding parts of the flange ( Z)

2) The longitudinal compressive stress in the flange ((S'c)

3) A transverse direct stress required to maintain equili-

brium of horizontal moments for the outer part of the

flange (ß-t).

The principal tensile stress due to these three stresses

can be calculated as

C"I =2 'Zý sin e cos C) + CTt cos29- Cj'c singe .

(7.6.1)

2 'r
where tan 2 G) =-ý-cc+
'6t

A perhaps somowhat nominal value of ", ' can be calculated by,

ordinary beams theory as

b-b
L ()......... (7.6.2)
2Q

(actually a value of d1 -2 could have been used to ropresent the

386
 AU, -:
w.ý ,.

Fig 155 - Detachment of flange in T- beams due to

horizontal shear beam T 24
-

-587
 lever arm-; this would be a justifiable approximation unless

the flange is very thick, in which case this phenomenon is not

likely to occur. )

The average value of a-c for the whole depth oftho flenge

is
11
0=b....... (7.6.3)

The calculation of the transverse direct stress is more

complicated. Considering in the first instance only cases

where the loading is applied within the web breadth, referring

to fig. (156), it can be seen that if rotational equilibrium

is to be maintained, the total moment due to the transverse

stresses in a halfspan must be equal to:

Fec
(b- 2
fit = (7.6.4)
.............

Viewed as a horizontal beam the projecting part the

of

flange is an extremely deep cantilever and it is difficult

to know what distribution of transverse stresses be

should

assumed, but the is that

simplest assumption of a linear

variation of stress from a maximum compression (-(Tt) at the

support to
a maximum tension (6t) at midspan. This gives
Cý-ta2t
flit _ ---- 6 (7.6.5)

Equating the values of_01t from equation 7.6.4 and 7.6.5

0.75 Fcc (b_br)2
(7.6.6)
tba2t . '......
so 0
ba2t

in point loaded. beams this givos

0.75 Q (b-br)2

batz

Values of the transverse stresses at points between the

supports and centre span can easily be obtained from the

linear distribution assumed.

388
 Stresses determined from equations 7.6.2,31 and 6 can be

substituted into equation 7.6.1 to find maximum principal

tensile stresses at any load stage.

For centrally point loaded beams it is apparent that the

maximum principal tensile stress must occur at midspan (this

is also supported by the fact that during the tests the first

longitudinal crack in the flange appeared always at midspan;

furthermore measurements of strains in the transverse flange

reinforcement have shown that the reinforcing bars closer

to the load are stressed first).

At that section:
b-b
9(b r)
L=0.5

Qa 2a
_ b-b
c z-tb
r
1.5(b-br)

In the standard T-beams of the present series of tests

b-b
ärý0.5whence (Sc=4V' Gt= 0.75i

These values give O=0.954 i and longitudinal cracking

be expected once V=1.05 Ut. For practical pur-

-should all

poses longitudinal cracking can be considered to occur when

OrI =L= Ut in this case. Equation 7.6.2 can then be used to

determine the longitudinal shear cracking load Qor*. If no

transverse flange reinforcement is present the overhanging

portion of the flange will become detached from the web when

this load is applied. This is exactly what happened in the case

of beams T21 and T24 of the present series of tests and beam
(94)
D3 by Regan .

Table 6.1.1 presents a comparison between the actual load at

389
which the flanges of these beams became detached - upon appear-

ance of the longitudinal shear crack in the flange - and the

prediction of the foregoing analysis.

TABLE 7.6.1 Dotachment of flange in T-beams without transý, li

verse flange reinforcement}

Beam r f b lb t ýdl , Dt calc

o< ßctest Qcr* ßcc
g yq r r* awl c
2 Qt©st
lbf/in in in in in lbf/in2 lbs lbs

T21 1326 24-6 3 10 0° 420 33600 30300 0.90

1
T24 326 2 3 10 5° 435 38000 31370 0.84

D3 137 116114 1.5 7.5 90° 418 '11760 11300' 0.96

1
i
In calculating the above z is taken as 0.9 di and Ut=25

(UcY1)*. The term Ind11 which may be more relevant in some

cases than the flange depth t- though this does not occur in

the above - may be calculated by linear elastis theory with

10000
the modular ratio taken as 5+ The neutral axis depth
cyl
so obtained is to be viewed as an approximation to the average

value along the span.

The slightly improved strength of beam T24 as opposed to

T21 may be attributed to the fact that inclined stirrups tend

to transfer some of the shear away from the transversely

tensioned central region back towards the transversely com-

pressed support zones.

7.6.2 Transverse reinforcement in flan es

_
This reinforcement is necessary when there is a danger that

horizontal shear cracking of the type described above might

occur in the flange of a T-beam.

The present CEB recommendations on transverse flange rein-

forcement are based on the traditional Märsch theory which im-

390
plies cracks at 450 and axial compression in the concrete bet-

ween the cracks. In tests of T-beams without transverse flange

reinforcement. shear cracking of the flange takes the form of

a single crack parallel to the axis of the beam, and it is

difficult to visualize any inclined compressed struts of con-

crete in a flange containing such a crack. Even in tests where

large amounts of transverse reinforcement are present and

there is some tendency for a pattern of inclined cracks to

form, a continuous longitudinal crack still extends along a

considerable part of the span. Assuming that longitudinal

rather than inclined cracking is characteristic of the be-

haviour of flanges a more rational understanding of the func-

tion of transverse steel can be obtained by considering shear-

friction theory.

Shear-friction theory

A crack is assumed to have occurred along a shear plane

crossed by reinforcement having a total area Aqf and yield

stress fy. The reinforcement is assumed to be perpendicular

to the shear plane. The faces of the crack are rough and ir-

regular so that, when slip occurs along the crack, the pieces

of concrete on either side of it are separated slightly. The

separation stresses the steel in tension and this in turn

creates compression in the concrete across the crack. The com-

pression provides resistance to slip by virtue of friction

between the irregular crack faces. It is assumed that the sepa-

ration is sufficient to stress the reinforcement to its yield

point (the tests described in section 5.7 confirm this - see

figures (86) and (87) so that if tong is the

-) co-efficient

391
of friction between the faces of the crack' the shear-friction

resistance along the crack is given as: -

Q= Atf fy tan 0 (7.6.7)

.........

or dividing by the relevant area of the crack plane

qu =p fy tan I (7.6.8)
..............

The relationship between 'qu' and IpfyI has boen studied

experimentally in push-off tests by Hofbeck9 Ibrahim' and

(124)
Matto ck(44) and Zia who studied the case of an initially

uncracked section such as that with which we are concerned

here. Zia arrived at a failure criterion in the form of a

transformed Mohr envelope with sliding failures for low 'pf

y

and shear failures when large amounts of reinforcement are

used (pfy > 900 lbf/in2).

In T-beams we are concerned with tpfyI values for transverse

flange reinforcement that are well below 900 lbf/in2, and

'failura'involves the yielding of the transverse steel, widen-

ing of the crack, loss of friction between its faces, and the

transformation of the T-beam into a rectangular section. (If

the shear friction is a sufficiently plastic action the flanges

may of course continue to carry the longitudinal force they

held when yielding of the transverse reinforcement occurred. )

For normal concrete Ziats failure criterion approximates to: -

qu = 1.4 ut + pfy tan I (7.6.9)

...........

In tests on specimens with preformed (rough) cracks tang

lies between 0.8 and 1.4 and the slope of the straight line

part of figure 157, reproduced from reference (44), is prac-

tically the same as that obtained in tests without preformed

392
 ý. -ý
ýýreb flcn-+e juction
_-^, ---- -
b-I

rý

cer, ýre
ne Ii ný

r j. (15:') Forces in aT- bcurri flano 2

1400
111-Y
1 'L:, A initidly crcc',, _:d
on sh ýcr p cn

cr
Qu
in ccc
Ibf) it j
GC :.

x. 00

0 2.10 '. CO NCO ý?CC 1000 1'%C")1';C0

F,y in Ibf/i:

Sh r 3trc v

accordin jj to P-?
attcc1,4-'
cracks in the range of pfy values normally encountered in T-

beams. Taking an average value of tans = 1.0 the equation be-

comes: -

qu = 1.4 Ut + pfy (7.6.10)

...........

The increase of load with full T-beam action preserved by

the presence of transverse flange reinforcement, over and above

the flange cracking load may be expressed by: -

Quf qu pf
= 1.4 +y........... (7.6.11)
Quu
cr tt
In the tests of three T-beams -- described in section 5.7 -

differing only in the amount of transverse flange reinforce-

ment present (p=0.56%, 0.28% and 0.185% of plain round mild

steel), only the beam with the least reinforcement showed signs

of being near to failure by flange separation. Pertinent

results for this beam are: -

1) flange fully co-operative, ordinary shear

failure calculated by equation (7.3.9) q= 44300 lbf

2) separation of flange calculated by

equation (7.6.11) Q= 47200 lbf

3) actual ultimate load obtained in test Q= 47200 lbf

4) ultimate load obtained in test of an

otherwise similar rectangular beam Q= 40300 lbf

It seems that the T-beam was just able to reach its full T-

beam strength. Due to the low safety margin against flange sepa-

ration, signs of distress in the flange were evident at load

stages close to failure, and a greater safety margin would appear

to be advisable. It would seem that the minimum ratio of rein-

forcement for slabs permitted by the present British code of

394
 (21)
practice which is 0.15% is not always sufficient to ensure

the satisfactory behaviour of the flanges of T-beams. (It is

understood by the writer however that the situation will be

remedied in the new Unified Code by the adoption of a minimum

value of p=0.3% in the case of 1-beam flanges - section 6.2.2.2

of the Draft Unified Code for Structural Concrete. )

7.7 Detailing requirements

7.7.1 Effect of shear on forces in main reinforcement

7.7.1. a Beams without shear reinforcement

When shear cracking occurs a redistribution of internal

forces takes place and the stress in the main steel does not

follow the moment diagram any more. In the absence of shear

reinforcement strains in the steel at the supports become

almost as large as in midspan - see section 4.1 and reference

(91) the longitudinal bars

- and any curtailment of should

result in an important reduction of the shear strength. This

conclusion is also arrived at, through tests, by Leonharöt(63)ý

In the case of beams failing upon shear cracking, the shear

cracking strength of a beam as expressed by equations 7.1.2

is dependent upon the ratio of main reinforcement which to-

gather with the concrete strength determines the value of the

vortical axis factor no which has been used as an approximation

to the actual neutral axis factor n. Consequently if at some

section, the area of reinforcement is Ast and the correspond-

ing neutral axis factor is nö and at some other section the

area of steel is reduced to Ast' the neutral axis factor is re-

duced to nö and the shear cracking strength may be considered

395
 2
n
reduced proportionately to the ratio Conditions within the
n
anchorage lengths of curtailed bars ca not be analysed since

the anchorage length is dependent on the magnitude of the

applied load, and present knowledge of bond and anchorage

behaviour does not permit a full investigation of this de-

pendence. In design it should be satisfactory to assume that

the shear cracking strength varies linearly along the anchorage

length required by the code of practice in the ultimate load

condition.

7.7.1. b Beams with shear reinforcement

When shear reinforcement is present the strains in the steel

have a lesser tendency to become uniform throughout the whole

shear span than in the case of beams without shear reinforce-

ment. Although there is an increase of strain near the sup-

ports., this increase is not as rapid as that in beams without

wob reinforcement. However the strains in the steel along some

distance from the load tend to be almost uniform. This situation

produces higher stresses in the longitudinal steel than would

flexural conditions and this must be reflected in its detail-

ing. The traditional way of taking account of this increase

of stress in the longitudinal steel at sections in the span is

by designing the steel on the basis of a moment diagram result-

ing from a lateral translation of the flexural moment diagram

by a quantity dd.
(63)
Leonhardt proposed a value of dd=d1 based on results of

beams with rather high degrees of shear reinforcement. Rusch

on the other hand proposes formulae for different inclinations

of. the stirrups based on either the degree of shear reinforce-

396
ment or the nominal shear stress. Leonhardt's proposal appears

in fig. 158 while Rusch's equations as given in reference (63),

are presented in a graphical form in fig. (159) where

is the ratio between the amount of shear reinforcement

provided and the minimum amount which would be necessary,

according to the förseh truss analogy, to produce

flexural failure.

Lb
brZ

Fig. 160 provides some examples illustrating the application

of the above proposal to actual tests.

In the case of the tests reported in this thesis the lateral

translations implied in the above formulae were found not to

be substantiated except in the case of web crushing failures

where the very high degree of shear reinforcement produced

more flexural behaviour in the main steel; in this case Leon-

hardt's proposal of dd^dl seems to be adoquate. In the case of

normal shear failures however values in excess of even the

maximum predicted by these formulae - corresponding to the

case of minimum shear reinforcement - were noted - see chapter

4 section 4.9.2 and chapter 5 section 5.10.2 Consequently

-. a

new value for dd is proposed on the basis of the following

considerations:

Considering a strip of concrete below a shear crack - see

fig. (161) the equilibrium conditions the
- relating compress-

ion in the concrete to the tension in the stirrups and the

bond of the main steel give

Fb = rgfyq b cot Q .,.....

(7.7.1)

where Fb is the bond force per unit length.

397
 p iý
p
jih

force in main steel Fst

Fig. (158) Lateral translation of moment diagram

proposed by Leonhardt

1"

1.0
dd_O3* O9
\ ddNh =1'2-O9
h 35'ßb yl
0.8

d_O3*354-c
0.6
dd /h 1.2-1.21' U
b cyl .
ý\
04

0.2

0'ßr 0.6 0.8 1-00 0-02 0.04 0.00 0.08 0'ý0

0 0.2
'Cb/ Ucyi ý--,.
KEY
vertical stirrups
---- stirrups or bent up bars inclincd at 45° to 60°

Fig. (159) Calculation of the value of the lateral

translation dd according to Rösch

arc
 Fig. Examp1 s iilustraiting Leonhardts and Rüsch's
proposals. Tests reported by Leonhardt

P/2 P/2
II i ý'
ý{
jam' ,

i i ý:a
"(17

I %I
d05h for 'n : 10 It/

dd = (1.2-0.911 h=0.86 h
for 'Y)=0.38 M/z

I ýý týt
LL V
1-00
I.
Fig. (160a) Variation of the force (St) in the main reinforcement of
b6;
T- beams with stirrups. 1) 0.38 and 1.0
P12.1P12

I,

43
LLN

Fig. (150b) Variation to the force (Ft) in the main reinforcement of

rectangular beams with stirrups - -q : 1.0

Jaýý
 '}q- fyq
,q

Fig. (161) Sei L,ss cor, di ýions ct r; ain St ýI

I'° I

rt-;ý',mcri S

fo: cc in--`ý.
main

Fig. (62) Extension of t-,? force

II
 The equivalent expression for stirrups inclined at an angle

C< is:

Fb = rgfyq b sin 0< (sin oS cot s+ cos c<) (7.7.2)

....

If the equations presented above are used in design, then

rqfyq < Q/bz (the value used in the morsch truss analogy). It

follows that the bond force of the main steel is reduced. The

simplest safe idealization is to assume that there is no bond

along the length of any shear crack. This assumption is sup-

ported by measurements of strains in the main steel - see

4.9.2 and 5.10.2 -. Provided that at least nominal stirrups

are used, the maximum length of crack obtained by the proposed

method is about 1.5d*. Longer cracks occur when inclined stir-

rups are used, but the provision of a unique and even smaller

value for dd is also proposed by Leonhardt. Furthermore the

assumption of lack of bond along the length of any shear crack

is a safe one and the reduction in bond force due to the fact

that rgfyq <b is more pronounced in beams with vertical

z
stirrups - since the forces in inclined stirrups have a--

horizontal component.

From the above it follows that if a bar is required to

carry a force FAQ calculated by normal flexural theory, at

section A, then it must be extended beyond that section for a

distance equal to "l. 5d* + development length for force FA"

see fig. (162).

It is interesting to note that experimental values of dd up

(63)
to 1.2d1 have been reported by Leonhardt Actually Pro-
.

fessor Leonhardt in a conference given at Imperial College in

may 1969 mentioned that in the light of recent unpublished

401
 test results the value of dd should probably be increased from

its present level.

7.7.2 minimum requirements for shear reinforcement

Although in the tests reported in this thesis no specific

attempt was made to establish the minimum effective degree of

shear reinforcement, observations made during the tests and

study of other test results lead to the following consider-

ations concerning the minimum requirements that the shear re-

inforcement must satisfy in order to be considered effective

in design.

1. The amount of shear reinforcement must of be

course

sufficient to raise the ultimate load of the member in which

it is used the cracking load. The value

above shear recommended

by, the Shear Study Group(107) of the Institution of Structural

Engineers i. e. 0.87 rqfyq 60 psi or in other words r

4f Yq' .e
69 psi was found to be quite adequate. The CEB(26) require-

ment of rq = 100 when mild steel is used would lead to a

minimum rqfyq value of 80 psi when fyq=40000. This is un-

necessary high as very good results were obtained in these

tests from beams having rgfyq values of 55.7 and fyq=40000.

(The
A. C. I3ý requirement of rq- would lead to a minimum
100
of 60 psi in this case.

A minimum effective degree of shear reinforcement the

of

order of rqfyq 60 psi thus seems adequate.

2. To be effective shear reinforcement must be spaced

sufficiently closely that critical shear cracks cannot form in

the unreinforced spacings. Referring to measured

spacings along

the member the CEB recommend a maximum spacing (s) dl

of

402
 d.
87
actually dl - and the ACI code a spacing of 21 for vertical

stirrups and 0.75d1 for stirrups inclined at 450; the S. S. G.

recommendation is 0.75d1.

The ACI recommendation seems to be somewhat severe in view

of the test results but the CEO maximum allowable spacing is

rather questionable in view of the smalld values intrinsic

1
to shear compression failures of rectangular beams with a high

degree of shear reinforcement In the commentary to that CEO

recommendation it is however mentioned that in the present

circumstances it would be advisable to use maximum spacings

of about 12 inches for low values of the nominal shear stress

(q <220 psi for vertical stirrups and bent up bars and q <280

psi for inclined stirrups, after removal of the CEO safety

X
factor and 6 inches if the value of q reaches the
s=1.4)
maximum allowed by that code (456 psi and 560 psi) with linear

interpollation for intermediate values of q and o( . This may

produce quite stringent conditions depending on the geometry

of the beam. The S. S. G. recommendation of 0.75d1 seems quite

adequate in all cases and it is substantiated from a number

of other test results as fig. 11 of reference (11).

This value of 0.75d1 seems adequate not only because of

the considerations regarding shear compression failures of

s/
beams with high degree of shear reinforcement and the satis-

factory behaviour of even larger spacing used in these tests

but also because there seems to be a tendency for cracks to be

larger than that. Furthermore it is

wider at spacings related

to the geometry of the beam which is not the case of the CEB

commentary. If larger spacings than this are used and only the

403
minimum degree of shear reinforcement is provided, a dowel

failure may become critical (see 4 below) especially if heavy

main reinforcement is also present. This comment is based on

the case of beam T1Gy which had a very long span (6 feet)

minimum shear reinforcement (rgfyq: heavy main steel

=55.7)

(4.16%) and a stirrup spacing of 0.9d1. Although the calculated

shear strength of that beam was reached, when failure occurred,

the failure was followed by a separation of the concrete along

tho level of the main steel all the way back to the support.

3. The shear reinforcement should be adequately anchored so

as to develop its full stress at the positions where it is

crossed by shear cracks. Stirrups should be anchored both in

their upper and lower ends along a minimum length, Thus' even

in the case of deformed steel, hooks or bends should be pro-

vided. Hooks or bends should be made around a longitudinal

bar and facing inwards so that the stresses developed should

result in a compression directed towards the web of the beams.

s/
4. Dowel action consideration also demand that stirrups

enclose all the tensile reinforcement. In this manner the

dowel force is transferred into the stirrups and dowel failure

is avoided. Such failure occurs when a shear crack opens and

the tensile reinforcement acts as a dowel and tears away its

concrete cover.

5. Bent-up bars cannot be considered as being attached by

themselves to the tension zone. Stirrups must be used together

with bent-up*bars if the shear reinforcement is to be con-

sidored effective. Tests on beams having only bent-up bars as

shear reinforcement have produced premature failures duo to

404
the separation of the tension zone from the web(58)(92). It is

interesting to note that the T-beam BX1 of reference 92 that

failed in this manner had no stirrups at all as was the case

of the pertinent beams of reference (6); On the other hand

beam BX3 which failed in a similar manner, although at a later

load stage, had shear reinforcement consisting of bent-up bars

and stirrups. The main reinforcement consisted of 12 bars TT25

disposed in four vertical columns of three bars each. Six of

those bars were bent upwards in three sections of the shear

span in pairs. The stirrup arrangement at a given section

consisted of two TT10 bars with hooks at both ends which

surrounded only the two outer columns of TT25 bars. The four

bottom bars in those two rows ware not bent but extended

throughout the whole length of the beam. This example is a

good illustration of the need for the requirements implied

in this as well as in the preceding paragraph 4.

(58)
On the other hand tests by Khan have shown that the

CEB recommendation that stirrups carry at least 40% of the

total shear, when used in conjunction with bent-up bars can

be quite excessive. It would appear from Khan's tests that

an rqfyq value of 60 to 70 lbf/in2 provided by stirrups is

sufficient.

7.8 Short and Doep ©oams: - Moment-Shear interaction

_
It has been seen in chapters 4 and 5 that when the shear

span to effective depth ratio is reduced below a certain value

situated between 2.0 and 3.0, the resulting increase in shear

strength tends to be in excess of that intrinsic to the

405
pattern established at higher a/dl values. In the sections of

these two chapters concerned with the influence of this para-

meter it was also explained that no single shear theory can

represent adequately the behaviour of the beams at low and

high a/dl values. The dividing point in beam behaviour-which

can also be construed as a boundary between the preponderances

of beam and arch actions - is not fixed but stoma to depend on

a number of parameters that have not yet been fully investi-

gated. One of these is illustrated by fig. (163).

The study of the effect of a/d1 in beams subject to con-

contuated loads is in fact oquivalent to the study of the

influence of flexure on shear resistance. The discussion that

follows examines this interaction of moment and shear in general

as only in this manner can the problem inherent to short beams

be situated in its proper context.

7.8.1 Beams without shear reinforcement

In examining first rectangular beams without stirrups it

can be seen that in sufficiently large shear spans (say

a/dl )6 to 8) failure occurs by flexure without any shear

cracks forming. In somewhat shortor spans (say 3< a/dI < 6)

a shear crack forms from the head of an earlier flexural crack

in the shear span. With no appreciable increase in load this

crack breaks through the compressive zone, simultaneously

tearing the concrete at the level of the main steel back to-

wards the support and the beam collapses. The entire failure

is of a tensile type and is seemingly initiated by the over-

coming of the concrete's tensile resistance by a principal

tensile stress in the neighbourhood of the neutral axis.

406
 PP
°ý nc a

ýb J. As

(b)

14-0 p. S. i. b.d
psi (n.
d. =1-0)
oo;
rc
ý Ü.: i1

ti. 6.0 in.

500
C'ý 10-7
c.. Z50in
400

00

2 DO .-o

100 -P 10

p 05093 r)=o-`o°;
o
`/0 rj1
1234567_. `.

F19. (1S3) J Ci! S'il°: al fodur- : ''"'rsu ci/d

7esl: s by 5
Kam -`'
 At a section with a flexural crack, shear can be supported

in three ways

1) by shear stresses in the uncracked zone

2) by interlocking of aggregate across the crack

3) by the dowel action of the main reinforcement.

At present-it is not possible to calculate either the dis-

tribution of shear between those actions or the distribution

of stresses within each. Tests on`aowel" beams have shown

that the dowel action can be taken as a fraction of the criti-

cal load in equivalent normal beams, From this fact and the

knowledge that the shear stress must be zero at the extreme

compressed fibre, and probably rises through tho compressive

zone, and that the interlock "stress" probably dins away to-

wards the wider part of the flexural crack, the distribution

of shear across the section may be expected to be more or less

(164). lie between the limiting

as shown in fig. It must two

cases shown in fig. (165).

For both the extreme case assumptions of fig. (165) it

can be shown that the magnitude of the bending moment, and

thence the (a/dl) ratio should have little effect on the shear

cracking load. In the probably more realistic model of fig.

(164) some slight influence might be expected as widening of

the floxural crack should reduce the interlocking' and thereby

increase the maximum shear stress. Tests of normal beams show

that the effect of (a/dl) on the critical shear force is

negligible. This last remark cannot necessarily be said to be

valid once the main steel yields at the critical section for

shear cracking, but this is something that does not arise in

408
normal simple beams failing in shear.

As mentioned above longer spans without shear reinforcement

fail immediately a shear crack forms, so the problem of struct-

ural behaviour in the presence of shear cracks does not arise.

This is not the case with shorter shear spans. Figures (166)

and (167) provide an illustration of the influence of the

(a/dl) ratio on ultimate shear strength. The interaction shown

in fig. (166) is that which is predicted by most shear com-

pression theories (and plane of minimum strength theories,

when the length of the crack is limited by the shear span a)

and seems to be obtained in current tests at Imperial College

by E. Oib(32). The scheme of fig. (167) is duo to Kaniý53)

and is supported by his own tosts and by some of tho tests by

Leonhardt and Walther.

It is not obvious why tests should give such conflicting

evidence but one reason may be that superior bond character-

istic of the stool used at Imperial College (and in tests by,

Clark(25) which produced similar results). Another possible

explanation could lie in differences in the degree of freedom

actually achieved at supports - any longitudinal restraint

could load to misleadingly high strengths being obtained in

tests of short beams. For the time being it would seem prudent

to use the more conservative version fig. (166) for any

- -

design purposes.

The excessive scatter in the results of short beams con-

stitute one of the major difficulties in arriving at a rational

conclusion simply by studying them. Do Cossio and Loera(29)

(54)
in a discussion of Kani's article observe that two identical

409
iLýýrlirl C:: i--r

Regan Mörsch R: ttcr

-

F i9. (ß 1-)F- ooc'ülc cistribu'i,,,; i Fig. (165) Extrern. - assumptions of

o` shear Stres _:ý at a of shear stress distributions
c; ock' d c_ction at a crac; tcd section

,ýýi ,- ý; -, _..
. -ý_.-

P''uI"1: suit

I
7/
/ +_...
I1 _1... __ý
12345670
Old

....._+
I I- I....... ý
. ý.
11I1I1!

12345G78

Fig. C37)Morncnt-skicar interaction

in rzctcr, ý. lc.r barns v; ith: out in rectcn:; ular beams without
Ctc) stirrups (Kani etc)

prob: °
-: 1in,:., o`

ý Iý vfl

3) Su-r Drt in Fick.. (1ä9) Forces in

Fi(' of E;i: ar and stresses
a short bm a short beam
 beams, tested in the same laboratory with a/d1 ratios of two

had failure loads differing by 41%. An, albeit much loss im-

portant, difference in failure load between two similar beams,

was also obtained in the testa reported in this thesis. Beams

R4 and R6 with an a/dl ratio of 2.24 failed at 17 tons and

14 tons respectively.

Moreover the exact loading and support conditions must play

an important role in the shear strength of short beams as

demonstrated in tests reported by Ferguson(38) and by R.

(115),
Taylor where beams with small a/d1 ratios were tested

with the load and the reactions not directly applied to the

faces of the beam, but through lateral stubs. These beams

failed at loads of the order of the inclined cracking load,

since the confining action of the bearing plates was not

present. This consideration is of course closely linked with:

1) Walther's vertical load transmission stresses, opposing the

destruction of the concrete in the compression zone; these

stresses, according to Walther, are significant only when the

i. d in the
ratio - e. case under discussion - is less
Qýd
11
than, 3.0 (see section 2.4).

2) Increased dowel action in short beams (see section 2.9 and

reference 9) as will be discussed subsequently.

It may be also interesting to note that while the influence of

the a/dl ratio on the shearing force at failure is consider-

able, Ohno and Arakawa(84) believe that the load at which the

first diagonal tension crack occurs is practically independent

of the value of a/dl.

The scatter of results existing at low (a/dl) ratios is also

411
demonstrated by the fact that a statistical analysis of exist-

ing results by Zsutty(12J enabled adequate formulae to be

developed reflecting the effects of the major parameters

affecting shear failure, for beams whose a/d 1 ratio exceeded

2.5. According to Zsutty's own admission, the analysis of

available test data for short beams, with a/d 1(2.5, did not

provide a satisfactory prediction equation.

Anticipating somewhat on the conclusions of this section

it is obvious from the foregoing that a comprehensive series

of tests, analogous to the ones carried out in this thesis,

is necessary at varying a/dl values lower than 3.0. The follow-

ing other parameters should also be varied or studied:

1) percentage of main steel reinforcement

2) percentage of compression reinforcement

3) concrete quality

4) amount of shear reinforcement (and inclination)

5) effect of load and support conditions

Such a series of tests on rectangular beams is presently

underway at Imperial College by E. Dib. The boundary value

of 3.0 seems to be adequate not only in view of the results

becaus7 it
of the test reported in chapters 4 and 5 but is also the value

suggosted by a number of other authors - Walther(119)1 Jager(50)0

The present discussion oil short beams therefore is only an

attempt to situate the problems and suggest the type of solu-

tion that could be envisaged. Although, due to their increased

flexural capacity, T-beams can fail in shear at larger (a/dl)

ratios than rectangular beams, their behaviour at short spans

in terms of load carrying capacity is analogous to that of

412
rectangular beams. Because of the fact that such beams are

rarely used in practicer the author agrees with Ramakrishnan's

opinion that separate treatment of short T-beams is not

warranted.

Returning to the two interaction diagrams it can be seen

that either scheme involves four zones of behaviour. Zone 1

is purely flexural and needs no further consideration. Zone 2

is the zone of diagonal tension failure and has already been

discussed. The improvement in shear strength in zone 3 is of

immediate concern here.

Zone 3 The greater shear strength arises when such spans

as fall in this zone do not collapse immediately upon formation

of shear cracksi The structural system to be studied is thus

that acting in the presence of shear cracks as illustrated by

figure (168). Its noteable features are:

1) The shear crack extends virtually the whole length of

the shear span

2) The head of the shear crack can be very close to the top

surface of the beam

It appears convenient"to-treat the shear resistance of such

structures as boing the sum of two parts

a) The zone of concrete above the crack and the part of the

zone below the crack from which forces are transmitted

across the not very open part of the crack.

b) The main stool acting as a dowel, now very possibly sup-

ported by direct compression radiating from the support

platten.

If the major part of the load is taken by the concrete and

413
failure is initiated in the concrete then the beam's behaviour

can be likened to that of a tied arch. The failure at the crown

of the arch is due to "compression" or "shear + compression"

as shown in fig. (169).

Looking first at the shear + compression idsa, a Coulomb

type'friction equation seems to be a reasonable failure cri-

terion and can be written as

q=c- nlr

where the sign convention is as in fig. 169 and the other

symbols have been defined in the previous section dealing with

the shearing type of failure.

So long as it is free to do soy the concrete fails along

the plane at an angle 6< that for the (f:

, such given ratio v)

a minimum load is obtained. With the reasonable assumption of

(V _ 0.9v the angle t`C is in the range arc tan 2 to j i. e.

d
approximately arc tan (ä) at the start of the strength in-

crease.

The ultimate shear strength predicted by this shearing

approach is

Quit vcrit x Acc

where vcrit is of the order of 0.4 to 0.5 Ucyl'

The area of concrete involved in the failure "Acc" is very

difficult to determine. Reference to the area of the uncracked

zone above the head of the shear crack is not very helpful

since in many instances the crack reaches virtually to the top

surface distinctly before failure. This could be interpreted

as indicating that it is not arch but dowel action that pre-

dominates. Such an explanation seems however improbable because

414
 although the dowel force is no doubt increased in short beams,

a very considerable area of concrete is damaged in the failure.

One possible assumption is that the area A., is approximately

equal to that of the flexural elastic compressive zone, on the

basis that it is only below this zone that the crack widens

sufficiently to prevent interlocking if the failure movement

involves a displacement with an appreciable component parallel

to the crack.

It is the author's opinion that the shearing approach as

developed in this thesis is a valuable one in some types of

beams with shear reinforcement. It seems less applicable in

the case considered here because as presented it inevitably

predicts a failure criterion for the beam that is closer to

constant shear rather than constant moment as (a/dl) varies;

it should however form the basis of an approach provided it is

expressed in the form of a fixed angle of failure for the

concrete, in which case higher loads will inevitably be pre-

dicted as increased shear strengths are given by Coulomb type

(42)
theories when the plane of rupture is imposed. Furthermore

when the beams become progressively shorter the failure crack

may not attain the 1: 2 inclination at which interlock may be

neglected and the residual interlock in these cases must be

taken into account.

Looking at the "compressioh'idea, it is well known that the

interaction line of zone 3- fig. (166) can be predicted by

-

shear compression theories giving

mutt 'k Ucyl b d12 (1-k2

where n is the ultimate neutral axis factor (or more use-

415
 fully bndI in the area of the compressive zone by
as reduced

shear cracking) and kUcyl is the ultimate average compressive

stress in the concrete.

The logical defence of this approach lies in the consider-

ation that there is as such no need for shear stresses to

exist near the load point. If the critical element is taken

in the way shown in fig. (170) it is to

subject only a primary

compression and such secondary restraints as are provided by

the surrounding concrete. What is difficult to be certain of

is how a failure corresponding to ß" = approx. Ucyl should

appear. One would expect some signs of the 'columnar' cracking

common in cylinder crushing tests, and this type of cracking

is presnt in many short members. It cannot however of itself

give the freedom of movement necessary for the collapse of the

beam; it seems inevitable that the final coming apart must

occur on an inclined plane' since very little of the top sur-

face is highly compressed and a rotational failure cannot

occur unless the concrete under the load platten or the stub

column crushes. This in turn is prevented by the presence of

biaxial restraints. On this basis the cause of failure is

shear compression (0 '= approx. Ucyl) but once the concrete

is coming apart internally (increasing C. for constant Q" ) the

final motion of collapse is a shear movement.

This explanation agrees with actual zone 3 behaviour and

could even support some increase in Mult with decreasing (a/d1)

since the restraining effect of the surrounding concrete should

be greater in shorter beams - see fig. (171).

416
 ---2"

Fig. '7/U) S=riss state r cr a

toad in a short born

r---1_
ý
ý, `ý
// //

/r /7

ý/

Relatively IittI. restrc nt Concid rcb! 2 rostrcint

(arge port of critical zon . criciccl zone surrcý. incLd
is at top surfccc by lic1ntly stress-, d
ccncrLtz

Fib. (171) Influsnce of sh: ar sp n on rz--trc: int of or-P.iccl a.onc

Inas

Fig-(772) Split cy! iric cr ana! o(j:y Fig. (173) SxLit cri tip c ;ti,:
for very : dort b: c!rns in

/ cn of conci-c', c
c1 _?ro cc in
,//

L-U

FiJ, I, ý-", ccti": );1 of cFiJ. (7%5) Si°ýar pr 0; 3r

"); .ý , UmýU
vr_ry cl ,-!:?p cor"L ýl fciitw
 It can thus be agreed that the choice of method depends on

the proximity of the head of the failure crack to the point of

load application since the fraction of the vertical component

of the inclined thrust taken up by the load platten must depend

on this distance. The limited test evidence at only one value

of a/dl < 3.0 reported in this thesis tends to favour the shear

compression approach.

Zogno 4 This typo of shear span (zono 4 of figs. 166 and 167)

occurs in what may be called deep beams. Various attempts have

been made to deal with shear in such members.

One advocated by Brock(22) amongst others is made by analogy

to the split cylinder test - see fig. (172). Physically this

analogy seems wrong as there is no evidence of short beams.

splitting, and the cracks similar to that shown in the figure

occur well before the ultimate limit state.

Another approach is that adopted by Regan and others. It

assumes that subject to specified limitations the main steel

will yield and the structure may be treated as a simple strut

and tie, as illustrated in the case of a column bracket in

fig. 173. Regan's(94) main limitation to ensure yield of the

steel is that the neutral axis depth nd1 as shown in the figure

should not exceed 0,5d1. This approach works quite well and is

probably true enough for lower percentages of main steel. With

higher reinforcement ratios in very short shear spans it has

been shown mathematically by Regan that the greatest calculable

shear force can be obtained if only a part of the total depth

of the member is assumed to be operative see fig. (174).

-

Except for very (unrealistically) high percentages of rein-

418
 forcement this gives good agreement between calculated and ex-

perimental ultimate loads, but physically it seems rather

dubious.

Physically what appears to happen in very short shear spans

is that there is an almost total destruction of the concrete

along the line of the crack shown in fig. (175). When the

span is short enough and the crack angle is deep, it seems

reasonable to suppose that:

1) The maximum supportable interlock stress is increased

as the displacement of the crack (predominantly vertical) is

more nearly parallel to the crack line.

2) The crack opens so littlo that interlock is well pre-

served throughout the tensile zone.

This description agrees reasonably well with LaupaTs(61)

concept of tshear proper'. Unfortunately Laupa's equation is

so little developed that it is of no real use. It may well be

that the best approach here will be the development of equations

dealing with aggregate interlock expressed in terms of aggregate

characteristics, and the angle between the crack and the vert-
d
ical, noting that the angle is arc tan (äl) if a is taken as

the distance between the edges of bearing of the load and

support.

7.8.2 Beams with shear reinforcement

The influence of flexure on the shear resistance is genor-

ally taken into account by both the methods proposed in earlier

sections of this chapter. To illustrate this point a`briof

review of the two methods and the intrinsic mode of failure in

longer beams is necessary. '

419
a) Shearing failure

If equilibrium is reestablished after shear cracking, it

seems that the sequence of events on further loading to failure

is as follows. Initially slightly inclined cracks still trans-

mit shear by interlock in the tensile zone and thus break down

to progressively flatter cracks. An angle with the horizontal

of CC = arc tan z seems to be the limit of cracking produced

in this way when vertical stirrups are used, and it may be

presumed that with an angle this flat, the interlock trans-

mitted along the main part of the crack is negligible.

At the top of the crack there may still be no great dis-

continuity across the crack - evidence for this lies in the

fact that particularly when significant compression reinforce-

ment is present cracks can reach Virtually to the top surface

without causing failure. The concept of a residual effective

area of concrete of size dependent on the same parameters as

the flexural elastic compressive zone thus seems more useful

than the definition of a crack height.

The stress conditions at the head of the cracks are as des-

cribed above in connection with zone 3 beams without shear re-

inforcoment, with the important proviso that in this case the

shearing failure of fig. 169 can be significant as shear stresses

must exist at crack heads if (as is generally the case) there

is no loading platten present to take up the inclined thrust

see fig. (176).

It has been shown that in this case the shear carried by the

concrete can be approximated to by

(bcyl
bd . rat
1-k

420
 The horizontal projection of the open part of the crack can

be taken as about 1.5d1. Assuming shear roinforcement along

this length yields prior to shear failure

Qc
k(U rst dl/a)* + 1.5 rgfy
bdlt - cyl ' .
11q
This type of failure is likely in beams with light (defective)

shear reinforcement and in beams having sufficient compression

reinforcement to prevent shear-compression failure. Again in

this second instance it is basically a question of defective

shear reinforcement since unless the compression reinforcement

is excessive (see sections 4.1 and 4.7) shear-compression

failure will generally occur if the stirrups are only moder-

ately underdesignod.

Only if the main reinforcement Jrstfyst* is very small is

it likely that with increasing shear reinforcement one could

go straight from shearing to flexural failure without paP51ng

a shear compression phase.

In the above equation (dl/a) is a measure of the ratios3of

stresses due to shear and stresses due to flexure and could

Qdl
well be replaced by (m) to. give the equation a more general

applicability.

b)-Shear-compression failure

This is the typo of failure previously discussed in many

papers and need not be discussed too much here. Since the

failure involves something like a rotation of the critical

plane (somewhat restrained at the level of the main steel)

it seems likely that interlock effects are negligible. The

area of concrete involved in a compression failure is then the

area above the crack head and should be determined by compati-

421
 bility conditions involving the slope of the thrust in the

concrete (influenced by the degree of shear reinforcement) and

admitting all values of the crack length c as possible. These

effects are implied in the shear compression equation of

section 7.4 based on the modifications (see Appendix A) to

Regan's theory.

In this case the shear-moment interaction enters into the

equations in two ways. First it comes because the failure is

by crushing end the relevant equilibrium condition is one in-

volving moments (see sections 7.2 and 7.4). Secondly it comes

again in the compatibility equation in the form of the slope

of the thrust in the concrete - see Appendix A/2. Provided

that the effects of any loads applied along the length of the

crack are ignored in the compatibility condition, the inter-

4dl
action effects are both covered by the single term and
ný,
once again using this term the equation should be applicable

to all loading cases, at least for simple beams.

In beams with stirrups since the diagonal tension mode is

supressed the sharp division between zones 3 and 2 disappears.

The shortening shear spans of zone 3 can still have more effect
Qdl
however than is predicted by the terms discussed above,

for the following reasons.

1) Because the above discussion implicitly assumes that there

is sufficient space available for the concrete to fail at its

minimum strength - i. e, in shearing the final failure angle

in the compressive zone is assumed to be that giving a minimum

422
 shearing strength; while in shear-compression it is asswmed

that the crushing zone is not abnormally rostrained.

2) Because of an increased dowel action in short beams,

(g)
as discussed in section 2.9 and substantiated by Arroyo
Because of the
3)/ possibility of increased residual interlock action in

short beams - see discussion on the shearing approach in the

previous section.

The overall interaction effect obtained by the above methods

is illustrated by the full lines of fig. (177).

As the shear span becomes short conditions must none the

less change. One reason for this is that the maximum possible

length of the shear crack is limited which in turn limits the

possible contribution of the shear reinforcement, even if the

shear reinforcement that crosses the crack is still assumed

to yield.

QW= rgfyq be

M=rfb -C?
wq Yqý2

this effect leads to the broken lines of fig. 177.

Finally in very short shear spans a new problem arises as

to whether the shoar reinforcement can really be at all effective

- whethor it yields and whether it makes much difference if

it does. It would seem that it does not have much effect in a

shear-proper failure, other than as a sort of dowelling across

the critical plane. In this sense vertical stirrups would be

very ineffectual as they are in the wrong direction. For want

of further evidence it may be assumed that the upper limit line

drawn for zone 1 beams without stirrups can be extended upwards

423
 ý40

ri 1. (176)

C. Gi
C)
ýýv C,

g
Pý'iýiý:: ý`' mil
iýý

I
/ S'
/
I ,/

,s
frj
'`, /
ý! ii: /

a 1d.1

_L,. _
23456

FiJj. 077) tnor1' nt - si%. ci crc: cc:,On 1.1)

rectangular beams
to deal with the present case the dotted line of fig. (177).
-

Discussion

Because of the important scatter of existing results in the

case of short and deep beams and the fact that in this series

of tests only a few beams at what is basically one value of

a/d 1<3.0 were tested no attempt has been made to establish

numerical formulae. The proposed equations for longer beams

could be transformed to apply to the few tests-mentioned but

such a procedure would be too restricted to be of any general

s/
application. Instead the author's opinion on this subject

were expounded, the problem was situated in its proper context

and the additional parameters present in the regions of low

(a/dl) values in load region

- confinement and support -

brought forward. A test programme is suggested and the type

of solutions envisaged explained. Upon availability of con-

sistent test evidence the suggested approaches can be used

to establish the necessary equations.

7.9 Physical Interpretation of Shearing Equations

7.9.1 Failure criterion for concrete

In order to predict the strength of any concrete structure

a failure criterion for concrete must be postulated. The com-

pressive zone of a beam is subject to longitudinal compression

and transverse shear while all other stresses are negligible.

It follows that one principal stress is compressive and the

other tensile, and consequently a two dimensional analysis

should be adequate. Various methods of'approach to this problem

(10)
have been proposed, some theoretical - Baker
1 Anson(s) -

425
 and some empirical - McHenry(76)1 aresler and Pister(18)(19)

and the problem is by no means solved yet. In order however

to formulate the shearing equations a failure criterion was

necessary and it was decided to adopt a failure criterion pre-

dicting shearing failures in the approximate manner that was

observed during the tests, i. e. a plastic shear mode rather

than a brittle tensile one.

If two adjacent stones are considered, failure at the junc-

tion can take place in two ways: 1) The stones will separate

and move in the direction of the tensile stress when the ex-

ternal tensile stress overcomes the tensile strength of the

joint; 2) The other mode of failure is by transverse sliding

and occurs when

q>c +ýn

where: q= external shear stress

^ co-efficient of internal friction; (it is sometimes

referred to in terms of an angle of internal friction

and in that case it is that angle whose tangent re-

lates the resistance offered to the sliding of two

planes to the normal stress between them)

n= external compressive stress

c= cohesive strength of the material; It is basically

the same phenomenon as that referred to as "tensile

strength" but may have quite different values duo

to the circumstances under which direct measure-

ments of the latter are usually made.

The failure envelope representing these two ways in which

failure can take place consists thus of a part of a circle

426
 passing through the point (Ut-0) with its centre on the hori-

zontal axis, and a line q=c+t? n.

If concrete were a homogeneous material with c=Ut, the

failure envelope would consist of an inclined straight line

of slope t1 intercepting the two axes (Ut, O) and (O, Ut).

at

Concrete however is a heterogeneous material and the strengths

of the joints betweon the aggregates vary. In"that case the two

ways of failure can be interpreted in the following manner:

a) The tensile stress at the weakest joint reaches its

limiting value Ut and a brittle tensile failure occurs in the

mortar; the load supported by that joint is passed on to other

joints and a chain reaction is then initiated which CaUsos

failure of the concrete without any increase of the external

load above the level at which the tensile strength of the

weakest joint was reached. In this case the cohesion is a

minimum and equal to Ut.

b) The chain reaction described in a) does not take place

and an increase in load is necessary to produce failure in the

next joint. The plastic nature of the transverse sliding type

of failure implies that the frictional force continues to act

while the displacement at the weak joints increases and causes

neighbouring joints to take up the excess load till the load

reaches a value approximately equal to the number of joints

times the average joint strength. Consequently the value of the

cohesion c in the equation q=c+UU n is an average value, greater

than Ut.

To construct the failure envelope Ut can be established by

tests. If the value of LIJ is known then a line tangent to the

427
 circle representing uniaxial compression (established by tests)

can be drawn having a slope equal to p. An arc of a circle

tangent to this line and intercepting the normal stress axis

at the uniaxial tensile strength may then be drawn to complete

the failure envelope, This type of failure criterion' based on

Coulomb+s and Couard's work on soil mechanics, has been dove-

(73)
loped by Lichtenberg who suggests a value of qJ 009. Its
LEY-'
construction is illustrated by fig. (178) for Ut =
10
In view of the preceding interpretation of the manner in

which failure occurs, this graphical representation of the

failure criterion seems reasonable and it gives the following

equations in terms of the stresses on the failure planes

n= Ut (7.9.1)
.......
qc- UU n ....... (7.9.2)

where n= normal stress On failure plane (tension positive)

q= shear stress on failure plane

tj)= frictional co-efficient for concrete

c= cohesive strength of concrete

7.962 Rectangular beams

In a rectangular beam local failure, corresponding to

equation (7.9. L) occurs when shear cracking Either

commences.

this causes a sudden diagonal tension failure, or else equili-

brium is reestablished within a changed structural system. Dia-

gonal tension failures are thus dependent on the tensile strength

of the concrete, and this is the parameter used in equation

(7.1.2). If diagonal tension failure does not occur and equili-

brium is reestablishedt the part of the shear force carried by

the concrete is supported by an inclination of the main com-

428
pressive force. When the failure load is eventually attained

in this modified system, the forces on an element of compressive

zone bounded by a possible failure plane, the compressed surface,

and a vertical plane are as shown in figure (179)

,
In this representation no regard is paid to the distribution

of stresses along the surfaces an which they act. This is

justified by the facts that the failure gods through to the

compressed surface and thus involves all the material con-

sidered, and that the type of failure is the "plastic" shear

(73)
mode described by Lichtenberg
1 and not
the brittle tensile

mode. The stresses all along the failure plane are probably

the same, and equal to the limiting values for the concrete.

The equilibrium conditions give:

q=u (sing + fý5 sin cos (7.9.3)

.....
(sin n- ý3 sin2 (7.9.4)
n=w cos .....

where the stresses Ifs and 'v t are as defined by figuro (179)

and P=f/v.

The failure criterion in this case, which is the one referred

to in the main text as shearing, is equation (7.9.2).

Thus: -

(sin 2+p oC+v(f Ü cost)

v sin cos sing - sin tt

v-i.,... (? 1
or C (I- %Vp, sin 46 + ý... """' '9'S)
qi+ sin S cosh

Failure occurs on the plane for which the equation gives the

minimum value for v-i. e. the plane for which the denominator

of the right hand side has a maximum value. Setting the differ-

Yj
ential of the denominator with respect to equal to zero

gives the critical plane as that for which: -

429
 ýý,
ýý
rý
ý ýý
/
io
riJý,.
I_ ý"
C, ý
tý
-`'
I//

t!
1
Ü1 Ucy
l

Fia. f 173) ra, Iurc critcrion fcr concrete

For cs Wir; rs
.., ..

Fig. (172) in rh2aring failure

i. C) - ------

r Tc;
t; an H fcýri
(1. ). J1

Vi C

o.'. º-____---

0-,

1 2.3 6 5
f /v

^^\ 1 1'1,
( OO) %; ,: UI' : )' G1''.::;. '.'::, cl- £+ i.: Ci ii,, 2
1, il-lii-1: ý1 - is
 tan 2t_ (7.9.6)
kPP 1 .::....

Assuming a value of 0.9 - proposed by Lichtenberg and

Johansen which is not too different from the value of approx.

0.8 to'. 9 suggested by Harris(42) in a similar theory for con-

cretes of varying composition and strength equations (7.9.5)

-

and (7.9.6) can be used to calculate (v /c) as a function of

(p)t as plotted in figure (180). Figure (180) that the

shows

influence of (fý)= f/v ) is not very large, and that it causes

no appreciable error in the calculation of the term Q. in

equation (7.3.3. -) if (f/v ) is taken as being proportional to

(a/dl) rather than to the more exact (M-mw)/z(Q-Qw), particul-

arly since the shearing mode of failure occurs when the amount

of shear carried by the stirrups is small compared to that

carried by the concrete.

Taking (f/v )= K1(a/dl), it can be soon from figure (180)

*,
that (v /c) = approx. K2( v /f) or that
d
v= approx. K3 `c (äl) (7.9.7)
.......

whore K11 K2t and K3 are constants.

The area of concrete involved in this type of failure cannot

be viewed as an area above the head of'a shear crack, since if

a high percentage of compression stool is used (a typical con-

dition for this type of failure) the crack reaches the top

surface before failure occurs. Furthermore the stresses in the

concrete above and below the crack show no marked discontinuity

at the region near the head of the wrack. It seems that only the

lower part of the crack, in the flexural tensile zone is so

open as to interrupt the flow of stresses. The area involved in

431
 the failure may thus be taken as being to the
proportional

flexural neutral axis depth prior to yielding the

of steel -
'Ind An expression for In' can be derived
l". as follows: -
l-nnIE l
_'
"nc
st

Fcc = bnd1 kl Ucyl

1-n 1
FIc
Fst Estrstbd1
- ýn .

where e1 is the limiting compressive strain for concrete.

Equating Fst and Fcc

1
f. Estrst
n2 c
1-n k1Ucy1

Noting that 1
n2/(1-n) = approx. k2n2'5 and that F, depends

on the cylinder strength of the concrete and can be expressed

c1=k
as %-- (U )3/8 see reference (94)
c3 cyl - tha
-, expression

reduces to

k4 (rst/U (7.9.8)
cyl11/8)0.4 .......

where k1l k29 kit k4 are constants.

The shear carried by the concrete at the limit state of

collapse is equal to the area involved in the failure multi-

plied by the critical average stress: -

QG = nbd1 v

0.4 [d'13
st
bd = k5 5c .......
(7.9.9)
1 a
ýOcyl

Since 'c' is the concrete strength in a state intermediate

between uni-axial tension and uni-axial compression, it can be

taken to be proportional to a parameter lying between Ut and

Ucyl, but much nearer to Ucyl since cannot be expected to

432
 to 0.9
vary greatly. Taking c proportional to Ucy10.8

*
0 4 (r) )
d1- k6rst. Ucy10.25 to 0.35 (7.9.10)
.......

For all practical purposes this can be taken as

QC dl
Ucyl (7.9.11)
bd = Kir . .a.......
st

7.9.3 T-beams

Shear cracks roaching the web/flange junction do not rise

into the flange until failure. Prior to failure longitudinal

stresses at the level of the junction are low, so an almost

pure shear condition exists and local failure will occur -

figure (181) This implies:

see - when n=v=U. -

Qc =Utbst....... (7.9.12)

where tw thickness of flange

b8 = effective breadth of flange resisting shear. It is a

function of the flange thickness and web breadth,

taking account of the distribution of shear in the

flange, and can only be established experimentally.

Having established the failure criterion we must also justify

the way in which it is applied, and show why it is equation

(7.3.9) involving this criterion that is critical for all

practical T-beams.

It is commonly assumed that the transverse forces in the com-

pressive zone at the head of a shear crack can be supported

by an inclination of the main compression. In fact the extent

to which this is possible in T-beams is very limited as will

be shown by the following considerations.

Figure (182) shows the forces acting on a simply supported

433
 /
i ý. l

flirn-'iiI(! ilrIiTIiillr
LU

Sh @r stres. ', s Direct str-; 5: cs

Fitj. 0B1) Stress conditions near

near soffit of f! cinc:

stale at

bottom of
flcrLv

F1

a C.

cl

-
h: 'L: t !iiT-
 beam element bounded by a shear crack and its projection to

the compressed surface. The main compression is divided into

two forces. The sloping one has a horizontal component F1 and

a transverse component Qc. The other has a horizontal component

F2 and no transverse component.

At any section 'x1 alono the length lot, the condition of

transverse equilibrium for the part of the element to the left

of the section is: -

dx
Fl + xrgfyqb. = Q

whence by integration
2
Y`-F (Qx+A-rgfyqb2 )
1

whore A is a constant of int egration.

QaFi
When x=c bc2
y=g-so A Qc + rgf
Fl +F yq 2
2 2
QaF1 rf
1 Lý"-Q(c-x)
and y=F + -ýý b(c2-x2)
112

When x=01 QaF1

bc2
FF F%f -Qc+rgfyq 2
112

The maximum possible slope along the length (a-c) is thus

QaFl
d1 bc2
-X
dx = Fl a-c
----r-
Fl+F2 - Qc +rf q yq 2

This can support a transverse force

QaF1
1 bc2
Q `r _---- .. Qc + rgfyq 2
a-c F1+F2
F
l
Writing -N and multiplying both sides by (a-c)
FF
12
rf bc2
bc2
Qa Qa + rgfyq -2- i. e.
2a(1

Now summing the shear resistances along the crack and its

projection to the surface, we have

435
 q rgfyqbc '+ Q0

Combining this equation with the previous limitation on Q

[ (7.9.13)
qý L rqV be -- 1-N - 13
.......

For p= 1p the geometrical conditions treated here thus im-

pose no limit on Qc1 and the transverso force supported by the

slope of the force in tho concrete at the head of the crack is

limited only by the strength criteria for that concrete. For a

T-beam in which the effective flange breadth for flexure is bf,

br/bf, (7.9.13) imposes a limitation

V= and equation on the

value of Qc which can be obtained by a slope of the force F1

within the web. The resistance achieved by the slope F1 is

significant only if Qc as limited by equation (7.9.13) is

greater than the shear on the concrete required to break down

the system in which QC is carried as a "shear°' rather than a

sloping compression. As previously discussed the breakdown

shear has a value

Qý UtbSt

The sloping action must of course always exist along the

crack, but in view of the above it will exist at the crack head

only if

Utbst rgfyqbc

This expression can be re-arranged to give

dbb2
2a ltst
bf dl +d2
cc rgfYq
1

The second term on the right hand side is very small, so v

has to be approximately equal to 1 (i. e. the beam has to

behave virtually as a rectangular member) if the slope of the

436
 force at the head of the crack is to be of significance. Ih

other cases (i. e. in all practical T-beams) the to

resistance

shear given by equation (7.9.2) is greater than the resistance

that can be offered by the sloping action. In such cases the

beam fails in shear if the shear force reaches Q= rgfyqbc +

+ Utbst.

7.10 Discussion

The use of two equations, reflecting two modes of failure,

put forward in this thesis, not only allows the calculation

of the ultimate load as the lesser of the two calculated

values, but also gives a good indication of the manner in

which a beam will fail (by comparing the two calculated values

obtained).

Furthermore, for the first time, the influence or lack of

influence of compression steel is given a rational explanation.

This has been a controversial point among researchers since

both influence and lack of influence have been observed in

tests. The explanation is that when a high degree of shear re-

inforcement is present the shear-compression mode of failure

is usually critical and in this case compression reinforcement

is beneficial. On the other hand when light stirrups are used,

particularly if the beam is not too long, it is the shearing

mode of failure that occurs, and compression steel has no

appreciable effect.

The basic assumption concerning the stirrups implied-in

these equations, namely that stirrups crossing the failure

crack yield prior to shear failure, is supported not only by

437
 measurements during these tests but also by the findings of other
(23)(71)(87)0
investigators

Applied to the tests reported in chapter 4f 5y and 6 the

the c
relevant equations presented give same average ratio
Qcm
test
of 0.91 in the case of rectangular beams with shear reinforce-

mentf T-beams with shear reinforcement and web crushing fail-

ures, with co-officients of variation of 0.064,0.075p, and

0.056 respectively. Applied to the beams without web reinforce-

ment the corresponding statistical results are Qýalc _ 0. '

test
and a co-efficient of variation of 0.05.

The 9% safety margin obtained in the case of beams with

stirrups is probably due to a dowel action in excess of that

transferred to the stirrup nearest to the critical crack, and

to residual interlock action. It would be ill advised to

attempt to remove this margin without knowing more about the

parameters that may affect these actions.

In the case of beams without shear reinforcement the dowel

action cannot be properly taken into consideration in the

formulation of equations, as is done in some measure for beams

with shear reinforcement, nor can the interlock action be

rationally explained in terms of the beam behaviour as has

been attempted for beams with shear reinforcement. In fact a

considerable part of the total shear force must be expected

to be carried by interlock action in beams without shear rein-

forcement failing in diagonal tension. Much research is still

required on these two actions before a truely theoretical

method of calculation can be produced. Consequently, because

of the relative importance of these actions, the safety margin

438
provided by the proposed equations for the rather sudden dia-

gonal tension failures has been kept greater than that pro-

vided for the more gradual failure of beams with shear rein-

forcement, whose behaviour constitutes the main subject of

this thesis.

The application of the various equations proposed in this

chapter to-the tests carried out is illustrated by fig. 183

where the comparison of calculated ultimate strengths with

test results is presented in a graphical form.

439
 /

150C

,.
pp
Q/brd1 (test)
in ICI in2

l00C

6A
60
KEY
o web crushing
AM equation (7.5.5)

a shear compression
A equations (7.4.9. ) etc.
op
A shearing T-beams
500 (7.3.9. )
equation

CA v shearing rect. ' beams

13
equation (7.3.6. )

V diagonal tension
0
0 (7.1.2)
d equation
0
C8
c8

0
500 1000 1: 00
01b d1 (calculated) in Ibs 1 ir?

Fig. (183) Comparison of calculated ultimate strengths with test results

.44
 CHAPTER 8

COMPARISONS WITH TEST RESULTS

The purpose of this chapter is to assess the prediction

capability of the proposed equations and to compare it with

that of other theories and Codes of Practice.

In the opening section the test results discussed in

chapters 4 to 6 are presented in condensed form and the ulti-

mate shear strengths obtained during the tests are compared

with the shear resistances of the beams as predicted by the

equations developed in chapter 7.

The relevant recommendations of three basic Codes of Practice

Shear Study Group Report, ACI/ASCE, CEO - are presented in

-

the next section. Equations for ultimate shear strengths are

derived, where necessary, and the predicted values compared to

the actual test results.

The test results are also compared with the prediction of

three shear theories, selected as being representative of the

various types of approach to the problem of shear. Regan's

method is used to typify a theoretical shear-compression app-

while Hillerborg's development of Borishanski's solution

roach,

is typical the plane of minimum strength methods. Laupa's

of

theory is also included in

semi-empirical shear-compression

the not only as an illustration of the American

comparisons

to the problem, but also because it includes specific

approach

for the treatment of T-beams distinct from those

provisions

applicable to rectangular beams.

The various comparisons are discussed in section 4. It is

441
 particularly interesting to note that only two methods make

any distinction between the shear strength of rectangular and

T-beams; all the others consider the shear strength of a T-beam

as being equal to that of the rectangular beam formed by its

web and the latter's projection to the top surface. The two

methods that consider the effect of the flanges are those of

Regan and Laupa. Unfortunately both approaches predict failure

in terms of shear-compression and consequently do not give

a correct interpretation of the physical behaviour of the

beams.

One of the reasons for developing an equation specifically

for the case of T-beams was the increase in strength shown by

T-beams when compared to equivalent rectangualar beams of the

same web breadth. This is illustrated in section 5 by a com-

parison of the equivalent beams tested. The increase was found

to be of the order of 20%. When the separate methods proposed

in this thesis for the case of rectangular and T-sections were

applied to these comparative results the same average value of

Qcalc
= 0.90 was obtained for boththe co-efficients of vari-
Qtest

ation being quite low in both cases. This can be construed as

an indication that the increase in load observed in T-beams

is adequately predicted by the proposed equations.

To verify further the validity of the proposed T-beam equ-

ation a comparison is made in section 6 between the predicted

and actual shear strengths of a number of T-beams with shear

reinforcement tested by other workers and reported to have

failed in shear.

Finally the proposed equation for web crushing failures is

442
compared to the more recent provisions on shear of the Draft

Unified Code.

8.1 Comparisons with th© proposed thou

_
The results are presented and comparisons made in three

different tables for rectangular, T- and I-beams. Only the

pertinent geometric and reinforcement details are presented

for each beam in these tables since full details have already

been given in chapter 3. The statistical analysis of these

comparisons is also given in a table form. No comparisons'are

made for beams of a/d1 < 3.0 since they are beyond the in-

tended range of applicability of the equation.

443
 TABLE-801.1 Rectangular Beams

Beam U, rst rsc rqfyq a/d1 Qn Re Qcalc

1 %it
lkips I nsc
I
No lbf/in % 1bf/in Qtest Qtest

R1 3800 0.975 3.36 10.1 0180

.- -
R2 3800 1.46 34 3.36 10.6 0.85
. -
R3 3600 1.46 - - 3.36 10.1 0188
R4 4950 1.46 34 - 2.24 19.0
.
R5 4950 0.975 - - 2.24 10.7
R6 5000 1.46 - - 2,24 15.7
R7 4070 1.46 34 3.36 12,25 -, 0.76
. - ----- --
_L
H8 3870 1.46 34 83.5 3.36 17.9 96 97 0.96
. . .
R9 4290 1.46 34 167 3.36 23.45 1.08 95 0.95
. . ;
R10 4295 0.975 34 83.5 3.36 16.9 96 95 0.95
. . .
R11 3800 1.95 34 83.5 3.36 20.1 90 91 0.90
. . .
R12 4920 4.16 37 83.5 3.6 24.6 88 87 0.87
. . ,
R13 4680 4.16 37 167 3.6 33.6 87 80 0.80
. . .
R14 4210 1.46 34 55.7 3.36 20.10 76 77 0.76
. . .
R15 4330 4.16 1g49: 167 I 3.6 31.4 92 91 0.91
, .
R16 4580 4.16 2.98? 167 3.6 31.4 93 1.02 0.93
.
R17 1850 1.46 34 83.5 3.36 15.7 97 94 0.94
. . .
R18(1) 4540 1.46 34 83.5 3.36 19.05 93 1.00 0.93
. .
R19(1) 4390 1.46 34 167 3.36 26.9 95 92 0.92
. . .
R20 6230 1.46 34 83.5 3.36 20.2 93 95 0.93
. . .
R21 6980 4.16 37 167 3.6 33.6 93 87 0.87
. . .
R22 4280 1.46 34 83.5 4.5 17.9 93 90 0.90
. . .
R23 4370 1.46 34 83.5 2.24 22.2
.
R24 4480 4.16 2.61 83.5 5.05 20.7 93 1.05 0.93
.
R25 4470 4.16 2,61 83.5 3.6 23.5 93 1.06 0.93
.
R26 5300 4.16 2.61 83.5 2.4 33.6
R27 1980 4.16 37 167 3.6 21.3 1.20 1.03 1.03
.
R28 4580 4.16 2.61 326 3.6 40.3 1.09 91 0.91
.
D2 4400 1.46 - -j 3.36 11.8 0.81
,
NOTES: (1) 45° stirrups

In the Qegn/Qtest tables"s"denotes the and

shearing
"sc"the shear-compression equations.

444
 TABLE 8.1.2 T-Beams

Beam Cross U 100A rf a/d 1 q Q Remarks

°yl q yq ult cýalc
section ý--brill Qtest
No ýt b lbf/in lbf/in kips
in in %
T-1 3 24 4050 1.25 83.5 3.36 24.75 0.89
T-2 3 24 4070 1.46 - 3.36 12.3 0.75 DT
T-3 3 24 3990 1,46 83,5 3.36 23.5 0.93
1-4 3 24 4710 1.95 83.5 3.36 24.6 0.92
T-5 3 24 4890 1.46 167 3,36 31.35 0.97
.
T-6 3 24 3740 4.16 326 3,6 46.0 0,94
T-7 3 24 3970 3.0 83.5 3,46 24, x' 0.69
T-8 3 24 4530 4.16 83.5 3.60 28.0 0.80
T-9 3' 24 2930 4.16 167 3.60 34.7 0.81
T-10 3 24 4090 1.46 55.7 3.36 19.5 1.00
T-11 3 24 5360 416 167 3.60 36.0 0.86 45° stirrups
T-12 3 24 4450 4.16 83.5 3.60 32.5 0.70 45° stirrups
T-13 13 24 1850 1.46 83.5 3.36 20.2 0.92
1-14 "3 24 4840 4,16 326 3.60 49.3 0.90 45° stirrups
T-15 3 24 4810 4,16 83,5 7.20 23.5 0.97
T-16 3 24 4740 4.16 55.7 7.20 20.8 0.97
T-17 3 24 4790 4,16 167 7.20 30.1 1.00
T-18 3 24 4120 4.16 - 3.60 16,8 0.74 DT
T-19 3 24 4340 4.16 83.5 5.40 25.5 0.88
T-20 3 24 4655 4.16 167 5.40 34.6 0.87
T-22 3 24 4980 1.46 83.5 3.36 24.6 0.93 45° stirrups
T-23 3 24 4460 4.16 167 2.40 43.7 short beams
T-25 3 24 7840 1.46 83.5 3.36 25.8 0.99

T-26 3 24 8260A 4.16 167 3.60 40.3 0,83

T-27 3 24 1740 4.16 167 3.60 29.7 0.87

T-28 3 24 4620 4.16 - 2.40 26.85 short beam
T-29 3 24 5350 4.16 83.5 2.40 34.7 short beam
T-30 3 24 4160 1.46 83.5 2.24 24.6 short beam

T-31 3 24 4495 1.46 83.5 3.36 21.3 1.05

T-32 3 24 4000 4.16 326 3.60 48.6 0.90
T-33 3 24 5340 1.46 167 4.50 24.6 1.25 Flexure
T-341 3 12 4920 4.16 83.5 5.40 25.2 0.91

445
 Table 8.1.2 Continued

Beam Cross u 100A rf a/d 1 Q ýQ Remarks

t ion cy1 st q yq ult calc
No sec 2 bd 2 est
lbf/in . lbf/in Q
r1 kips
in in
T-35 3 42 4880 4.16 83.5 5.40 25.8 0.89
T-36 6 24 3500 4.16 167 3.60 40.3 0.90
T-37 3 24 4615 4.16 326 3.60 47.1 0.94
T-381 6, 24 4380 4.16 326 3.60 53.8 0.87
:

TABLE 8.1.3 Thin webbed I-beams

Beam Ucyl Stirrups Qult Qcalc Failure Type

si l r Qtest
No lbf/in a( %q kips
m nch
mm
W1 4330 10 4.5 90° 2.17 36.0 0.86 web crushed

1830 10 4.5 0 0.96

W2 90 2.17 20.2 web crushed

W3 6605 10 4.5 90° 2.17 44.8 0.85 web crushed

W4 4985 10 4 45° 2,44 50.4 1.23 web not crushed

W5 4970 10 3 90° 3.25 42.6 0.90 web crushed

W6 4920 12 6.5 90° 2.17 34.8 0.95 web crushed

W7 4980 8i 3 90°f 2.08 36.4 0.90 web crushed

1
Table 8.1.1 Rectangular beams

The shear strength of beams without shear reinforcement is

calculated according to equation (7.1.2). In the case of beams

with shear reinforcement two series of calculated values, one

based on the shearing equation (7.3.6) and the other on the

shear-compression equation (7.4.9 etc. ). The smaller of the two

values for any given beam is its predicted shear strength and

in the alc
appears value*
Q
test

446
 It can be seen that there is a tendency for the two values

to be similar in the case of beams with light stirrups and

light compression reinforcement; this indicates that the fail-

ure is of an intermediate mode not accurately defined by either

mode, or that either of failures could occur. This is borne

out by the indeterminate physical aspect of the relevant beams.

When light or intermediate stirrups are used and heavy com-

pression reinforcement is present, a more clear cut shearing

failure is predicted, as would be expected. Conversely when

intermediate shear reinforcement is used or the shear span is

longer and only light compression steel used, there is a defi-

nite tendency for shear compression failures to be predicted

and this also agrees with the test results. This is even truer

when heavy shear reinforcement is present. Beam R28 serves as

a good example since its'shear-compression failure is predicted

in the table inspite of the fact that heavy compression steel

was present. In this manner the use of the two equations enables

the varying effect of the compression steel to be taken into

consideration.

Only one value of Qcm in excess of 1.0 was obtained; it

test
is the case of beam R27 with very low strength concrete. A defi-

nite shear-compression failure is predicted, the difference

between the two predicted values being quite important; this

indicates a more flexural behaviour of the beam and it is not

surprising that a final of 1.03 is obtained. It is not

Qcalc
test
an unsafe result since in practice such results would be eli-

minated through flexural considerations. The physical aspect

of the beam actually indicated a probable flexural compression

447
 failure, although it was a shear crack that opened wide giving

the ultimate rotation.

A statistical analysis of the results of beams with shear

reinforcement alb
shows an average value of Q. 91 with 0.064
pc
test.
coefficient of variation.

It can thus be seen that not only adequate and consistent

predictions are made but a comparison of the two values ob-

tained by the proposed equations, for a given case' can provide

a good indication as to the actual behaviour of a beam and the

manner in which it will fail.

Table 8.1.2 T-beams

All the T-beams are calculated by equation (7.3.9) except

those that have no shear reinforcement in which case equation

(7,1,2) was applied.

Comments similar to those made in the preceding table can

be applied here for the case of beams exhibiting values

Qýcalý
test
equal or greater than 1.0. Beam T-33 is not included in the

statistical analysis as it was an'intentional flexural failure.

Statistical analysis of the results of beams with shear re-

inforcement alc 0.91

shows an average value of with 0.077
Qc
test
coefficient of variation.

Table 8.1.3 Web crushing of thin-webbed I-beams

The calculated web crushing strengths appearing on this are

based on equation (7.5.5), and only the beams where web crush-

ing occurred are taken into account in the statistical analysis

cal o
which gives an average . value of 0.91 and a 0.056 coeffi-
test
cient of variation.

448
 Summary of the statistical analyses reflecting_ the performance

of the proposed equations:

Type of beam Number of average Standard Coeffi-

Qc cient of
beams tested alc derivation
Qtest variation

Rectangular beams

with shear reinforcement 19 0.91 0.058 0.064

T-beams with shear

reinforcement 29 0191. 0.070 0.077

Web crushing failures

of thin-webbed I-beams 6 0.91 0.051 0.056

Beams without shear

reinforcement 7 0.80 0.053 0.066

8.2 Comparisons-with Codes of practice

In order to evaluate the true prediction capabilities of the

various Codes of Practice it is necessary to remove the safety

factors contained in their recommendations. This presents no

particular difficulty in the case of the ACI/ASCE and S. S. G.

proposals but a problem does arise in the case of the CEB re-

commendations which are formulated only in terms of permissible

stresses. In order to render the predictions of all the three

codes comparable, boththe safety and load factors have been

removed from the CEB equation for permissible stresses and an

equation in terms of the resulting ultimate stress derived.

8.2.1 Ultimate Shear Strength Equations

a) ACI/ASCE

The equation appearing in the ACI standards are based on the

work of the ACI-ASCE Committee 326 "Shear and Diagonal Tension"

(4)
1 which examined a great number of test results and proposed

449
 empirical equations for calculating the shear strength of rein-

forced concrete beams. The work of that committee was prompted

by the collapse of a warehouse because of shear failure in a

beam without shear reinforcement.

The proposed equation for beams without shear reinforcement

is based on more or less rational considerations and empirical

evaluations of the constants involved (based on 440 tests). The

maximum load of these beams is considered to be that at which

the first shear crack appears, caused by diagonal tensile

stresses. A basic'formula for the principal tensile stress is

adopted and assumptions for the effects of various parameters

are made.

It is first assumed that the load-carrying capacity is

reached when the principle tensile stress reaches the tensile

strength of the concrete which is considered proportional to

ý-Ucyl. tensile
The principal stress is given by the following

equation - by considering. the stresses at a certain point below

the neutral axis in a homogeneous beam.

2+I(2'2+
ft `2

where ft = principal tensile stress

f longitudinal tensile stress

shear stress

Although the exact distributions of flexural and shear

stresses in a cross section are not known it is assumed that f

E
varies as E--ý times the tensile stress in the reinforcement com-
s
puted by the cracked section theory, that 'r varies as the aver-

age shear stress, that E0 is proportional to Ucyl, and that

at cracking ft klIUcyi. Then letting Q and m denote the shear

450
 force and bending moment at the section considered:
Ecfs EcfYi in u1 riif Ucy1 i uCyl
f oC 10CE ý--- C) aC k2
E E E, r bd 2
ss1 st sd1A st s st 1 Esrstbd1

3 bd

Substituting these three quantities in the principal stress

equation
[ fi1fUoYl ml 22
k1Ji k2 + k2 ýq + k3
1 bdl Er dl E r
sst sstnl

then
/(72-
k
[22 l11U k2 mýUcy1 22
cy ý
bd _ kl EsrstQnd1 + k3
i" Es1)
cyl

In that last equation the variables are observed to be

m-1
Q 1-Er.
--- and d1
It may be noted that these two vari-
bd1JUcyl sstQý;
ables are non-dimensional quantities, because is con-
l
2
sidered to be in lbf/in. In the statistical study, the ulti-

mate shear force was defined as that causing the critical dia-

gonal tension crack and m as the corresponding moment at the

location where the diagonal crack begins. On the basis if 440

tests the relationship between these two variables was obtained

as follows:

4cr rS---t 1
= 1.9 + 2500 3.5
TU
-Cy
bd1 Ucyl m 1

where

Qcr = shear force causing diagonal tension (or shear)

cracking

r= ratio of main reinforcement (ast)

st bd1
Qý shear force at eection considered

M= bending moment at section considered

451
 The final form of the equation is then

Qcr Qdl
JU ' /
bd = 1.9 + 2500 rst fll
3.5'/ Uoyl
1 cyl

There is a certain intrinsic oddity in the appearance of

this final equation which must be considered basically empirical.

This oddity lies in the fact that a term involving the concrete

strength is added to a term involving the ratio of main rein-

forcement and the shear/moment ratio. Since aggregate interlock

and dowel action are neglected, the shear strength of a beam

without reinforcement must be considered to be the shear re-

sistance of the concrete; this implies that all terms in the

expression for shear strength should be functions of the con-

crete strength.

When web reinforcemont is used an empirical term repre-

senting its contribution is added to the above expression for

Qer and the value of Qult is obtained. The stirrup contribution

derived from the study of 166 test results can be construed to

be based on the yielding of all stirrups crossing a critical

crack of a fixed length dl.

The total shear strength is given by the following equation

4bß
= bdr + sin Cl (sin i>< + cos u>l) rf L8 to 10 U
1.1 q yq cy1

where Quit = ultimate shear force

A
rq = ratio of shear reinforcement ( sint>< )
bs s uý

fyq = yield or proof stress of shear reinforcement

s= spacing of shear reinforcement along main rein-

forcement

C-ý = angle of inclination of shear reinforcement (taken

452
 in respect to the direction of the main reinforcement)

The various limitations on these equations are not discussed

here but are taken into consideration in the evaluation of

test results.

b C. E. 6.

Contrary to the ACI/ASCE Committee, the CEB does not consider

beams without shear reinforcement. The latest CEB recommendat-

ions are those appearing in the "Conclusion techniques de la

12e session pleniQre - Lausanne 22/27 avril 1966"(26), The

basis for the calculation of shear strength is given in section

R. 4.220 of reference (26) as follows.

"The resistance of reinforced concrete beams to shear is

assumed to be due to the formation at the cracked state, of

complex triangular structures, of which the classical Ritter-

fI1rsch truss constitutes a simplified model. This last model,

corrected empirically to take account of experimental results#

remains the basis for shear strength calculations. "

The basic formula proposed by the CEB is that the porcentage

of shear reinforcement in the web should satisfy the inequality

Wto TIW

where wto = geometrical percentage of shear reinforcement

wtm = the geometrical percentage of shear reinforcement

corresponding to the application of the fi1örsch

truss theory.

=a reduction co-efficient (defined below)

11
Using the CEO notation(27), wtl and rL are defined as follows
1.15 T*

wt wtc, CTät(sina+cos to sin u;

453
 01 ]0 1.15 ö
wtm = 1.10 - 0,045
sin o<+cos( < sin «
o at

Transforming this equation to comply with the notation used

throughout this thesis, noting that in the now CEB recommen-

`e= bh bZ
dation where h=da, and not as previously, and

taking into consideration the various safety factors implied

in the CEB notations (concrete 1.5, steel 1.15, load 1.4) the

following equation is derived in terms of ultimate values.

Qlt
< 0.75 xrxfx (sin LX + cos(>! )sino! + 0.030 U
bdl q yq cyl

Although not too different in form from the ACS equation

the CEB proposal is generally more conservative.

(107)
c Shear Study Group Report (S. S" Gam,
--ý--- -- . w. - tho/
This is the most recent British proposal concerning shear

strength of reinforced concrete beams. This report was accepted

by the Council of the Institution of Structural Engineers in

October 1968 and it is intended to include its findings in

the new Unified Code. A comprehensive explanatory note on this

report appeared in "The Structural Engineer" in July 1969(5.

It is a semi-empirical method whose theoretical basis is re-

portedly similar to Regan's theory described in section 2.

The maximum design shear strength allowed by this code is

qG0.87 rqfyq + 0.8 k qc 4 qc

Since this equation includes already a partial safety factor

for materials equal to 1.25, its removal gives the following

ultimate load equation.

quit k qc + 1.1 rgfyq 5 qc

where k=1.0 except for short beams (a <2.0)

dl

454
 4

and qc is given by the following table in terms-of the ratio of

main steel and the cube strength of the concrete.

S. S. G. Table for yalbes of q3 in2

Percentage of Cube Strength of Concrete

tensile rein-
forcement at UW lbf/in2
sec ti on
10001\ 30001 4000 5000 6000 & over
st
b d
r 1

0.15 40 50 55 60

0.25 50 60 65 70

0.5 65 00 90 100

1.0 05 100 110 120

1.5 100 115 130 140

2.0 110 125 140 150

2.5 115 130 145 160

3.0 120 135 150 165

3.5 & over 125 140 1155 1 170

1

8.2.2 Comparisons with test results

These comparisons presented in tables (8.2.1) and (8.2.2)

are

for rectangular and T-beams respectively. In calculating the

ultimate shear strength by the ACI/ASCE method a section d1

from the point of application of the load was considered

according to the proscription of that code. In the case of the

S. S. G. method the cube strength of the concrete was used as

obtained from the tests of relevant specimens (see chapter 3).

The statistical analysis of the comparisons is presented in

table 8.2.3 for two general cases. One case includes the same

type and number of beams that appear in the preceding comparison

with the proposed theory; the other case includes all the beams

455
tested and it can be seen that the application of a single

'to' both long 'and short beams leads'' o ä' defiiiitä

equaticn

increase in the coefficient of'variatibn.

7AýCE 'b. 2. i Rectänqül'är Beams - Comparisons with Codes of 'Pr c fite

Beam Ucyl rst rso rgfyq a/dl Qult I SSG
QACI Q QCEB

% % test
Q
No lbf/in 1bf/i2 n kips Qt Qtest

R1 3800 0.975 3.36 10.1 0.81 0,67 -

- -
R2 3800 1.46 0.34 - 3.36 10.6 0.80 0.75 -
R3 3600 1.46 - - 3.36 10.1 0.83 0.77 -

R4 4950 1.46 0.34 2.24 19.0 0.55 0.47 -

-
R5 4950 0.975 2.24 10.7 0.92 0.72 -
- -
R6 5000 1.46 - 2.24 15.7 0.67 0.57 -
-
R7 4070 1.46 0.34 - 3.36 12.25 0.71 0.68 -
D2 4400 1.46 3.36 11.8 0.77 0.73 -
- -
R8 3870 1.46 0,34 83,5 3.36 17.9 0.78 0.78 0.65

R9 4290 1,46 0.34 167 3.36 23.45 0.84 0.81 0.70

R10 4295 0.975 0.34 83.5 16.9 0.83 0.71 0.73

13.36
Rll 3800 1095 0.34 83.5 3.36 20.1 0.70 0.76 0.58

R12 4920 4.16 0.37 83.5 3.60 24.6 0.63 0.64 0.52

R13 4680 4.16 0,37 167 3.60 33.6 0.60 0.63 0.48

R14 4210 1.46 0,34 55.7 3.36 20.1 0.62 0.62 0.54

R15 4330 4.16 1.49 167 3.60 31.4 0.64 0.67 0.49

R16 4580 4.16 2.98 167 3.60 31.4 0.64 0.68 0.51

R17 1850 1.46 0.34 83.5 3.36 15.7 0.74 0.78 0.49

R18* 4540 1.46 0.34 83.5 3.36 19.0 0.77 0.77 0.68

R19* 4390 1.46 0.34 167 3.36 26.9 0.74 0.76 0.62
R20 6230 1.46 0.34 03.5 3.36 20.2 0.79 0.73 0.80

R21 6980 4.16 0.37 167 3.60 33.6 0.65 0.63 0.60

R22 4280 1.46 0.34 83.5 4.50 17.9 0.80 0.79 0.69
R23 4370 1.46 0.34 83.5 2.24 22.2 0.69 0.66 0.56

R24 4480 4.16 2.61 83.5 5.05 20.7 0.68 0.73 0.58

R25 4470 4.16 2.61 83.5 3.60 23.5 0.64 0.67 0.51

R26 5300 4.16 2.61 83.5 2.40 33.6 0.53 0.47 0.40
R27 1980 4.16 0.37 167 3.60 21.3 0.82 0.87 0.53

R28 4580 4.16 2.61 326 3.60 40.3 0.74 0.78 0.58

450 stirrups

456
TABLE 0.2.2 T-beams Comparisons with Codes of Practice

Beam u 100A rqf a/d 1 Qult ACI SSG CEB

cyl st yq Q Q Q
No calc c_ lc calc
lbf/in b rd1 lbf/in kips Qt Qt Qt
es t ©s t ©s t
T-2 4070 1.46 3.36 12.3' 0.71 0.68
-
T-18 4120 4.16 3.60 16.8. 0.58 0.59
-
T-28 4620 4.16 2.40 26.85 0.46 0.38
-
T-1 4050 1.25 83.5 3.36 24.75 0.56 0.56 0.48
T-3 4990 1.46 83.5 3.36 23.5 0.60 0.61 0.50
T-4 4710 1.95 83.5 3.36 24.6 0.61 0.63 0.54
T-5 4890 1.46 167 3.36 31.35 0.65 0.65 0.56
T-6 3740 4.16 326 3.60 46.0 0.63 0.67 0.47
T-7 3970 3.00 83.5 3.46 24.6 0.59 0.62 0.46
1-8 4530 4.16 83.5 3.60 28.0 0.54 0.54 0.43
T-9 2930 4.16 167 3.60 34.7 0.54 0.61 0.37
110 4090 1.46 55.7 3.36 19.5 0.64 0.64 0.55
T-111 5360 4.16 167 3.60 36.0 0.58 0.58 0.48
T-121 4450 4.16 83.5 3.60 32.5 0.46 0.46 0.37
T-13 1850 1.46 83.5 3.36 20.2 0.58 0.61 0.39
T-141 4840 4.16 326 3.60 49.3 0.61 0.63 0.48
T-15 4810 4.16 83.5 7.20 23.5 0.59 0.64 0.54
T-16 4740 4.16 55.7 7.20 20.8 0.61 0.66 0.55
T-17 4790 4.16 167 7.20 30.1 0,63 0.68 0.54
T-19 4340 4.16 83.5 5.40 25.5 0.55 0.58 0.46
T-20 4655 4.16 167 5.40 34.6 0.56 0.61 0.47
T-221 4980 1.46 83.5 3.36 24.6 0.61 0.60 0.56

T-23 4460 4.16 167 2.40 43.7 0.51 0.46 0.36

T-25 7840 1.46 83.5 3.36 25.8 0.67 0.57 0.75

T-26 8260 4.16 167 3.60 40.3 0.56 0.53 0.56

T-27 1740 4.16 167 3.60 29.7 0.58 0,63 0.37

T-29 5350 4.16 83.5 2.40 34.7 0.51 0.45 0.39
T-30 4160 1.46 83.5 2.24 24.6 0.61 0.58 0.49
T-31 4495 1.46 83.5 3.36 21.3 0.67 0.68 0.60

T-32 4000 4.16 326 3.60 48.6 0.60 0.64 0.45

T-34** 4920 4.16 83.5 5.40 25.2 0.57 0.61 0.51
T-35 4880 4.16 83.5 5.40 25.8 0.56 0.60 0.49
T-36+ 1 3500 4.16 167 3.601 40.3 0.48 0.48 0.35
X

457
TABLE 8.2.2 Continued
II
Beam U 100A rf a/d 1 Q ACI SSG CEB
cyl st q yq ult Q Qcalc Qcalc
ca 1c
brill Qte t
No lbf/i2 lbf/in kips Qtest Qet

T-37 4615 4.16 326 3.60 47.1 0.63 0.66 0.49

10.55
T-38* 4380 4.16 326 3.60 1,53.6 0.58 0.42

1 450 stirrups

*t= 6" (t = 3" in all other beams)

** b= 12" in T-34, b= 42" in T-35 (b = 2411 in all other

beams)

TABLE 8.2.3 Statistical Analysis of Comparisons with Codes

of Practice

Type of Number Code average Standard Coefficient

beams tested of Q deviation of
beams calc
--- variation
Q
tested tes t

Rectangular ACI 0.71(0.72) 0.086(0.080) 0.121(0.110)

beams with 0.124(0.097)
21 SSG 0.71(0.73) 0.088(0.071)
shear rein-
forcement (19) CEB 0.5857.60) 0.097(0.093) 0.166(0.155)

T-beams with ACI 0.58(0.59) 0.055(0.050) 0.095(0.083)

shear rein- 32 SSG 0.60(0.61) 0.063(0.054) 0,105(0.089)
forcement
(29) CEB 0.48(0.49) 0.084(0.077) 0.175(0.158)

Beams without 11 ACI 0.71(0.75) 0.137(0.087) 0.193(0.116)

shear rein- (7) SSG 0.64(0.70) 0.127(0.061) 0.198(0.087)
forcement ,

Note Figures in brackets are relevant to the statistical ana-

lysis of tests on boams with a/dl > 2.5.

B. 3 Comparisons with other shear theories

The three theories used in these comparisons have already

been described and appraised in chapter 2.

In the case of beams without reinforcement Hillerborg'a

theory was applied only in the case of the shorter beams

458
(a L 2.5) and Petorsson's diagonal tension equation was used
dl
for the longer spans. This was done because Hillerborg's method

as presented in chapter 2 does not apply to beams failing

immediately upon shear cracking. Petersson's method was chosen

because of the similarity of the two shear strength equations.

Furthermore the cube strength of concrete as measured in tests

was used in connection with Hillerborg's method. In the case

of Regan's theory both diagonal tension and shear compression

values were calculated for the'longer beams and the governing

value used in the comparisons; actually predictions of dia-

gonal tension failure resulted in all cases except for beam

D-2 and T-18"where both values'were approximately the same. In

the case of short beams only the shear compression prediction

was calculated. Laupa's shear compression theory, although not

strictly applicable to the diagonal tension failures obtained,

is nonetheless used in order to illustrate, through its un-

satisfactory performance, the need for any given theory to be

based on the actual behaviour and mode of failure.

The results of these comparisons appear in tables (8.3.1)

and (8.3.2) for rectangular and T-beams respectively. Statist-

ical analysis is prosented in table (8.3.3) in a manner similar

to that described in the preceding section' and for the same

reasons.

459
TABLE 8.3.1 Rectangular Beams- Comparison with Shear Theories

Beam Ucyl rst rsc I rgf a/dl Qult Regan Laupa Hillerborg
Q. Q1 QW
2 % % 2
No lbf/in lbf/in kips
Qtesti Qtest
test
R1 3800 0.975 3,36 10.1 0.71 0.91 0.82
-" - X
R2 3800 1.46 0.34 - 3.36 10.6 0.76 1.05 0.93
R3 3600 1.46 - -. 3.36 10.1 0.81 - 1.04 0.97

R4 4950 1.46 0.34 2.24 19.0 0.71 0.97 0.82

-
R5 4950 0.975 - - 2.24 10.7 1.04 1.39 1019
R6 5000 1.46 - - 2.24 15.7 0.90 1.13 1.00
R7 4070 1.46 0.34 3.36 12.25 0.67 0.93 0.84
-
02 4400 1.46 - 3.36 11.8 0.77 0.97 0.69
- .
RB 3870 1.46 0.34 83.5 3.36 17.9 0.89 0.73 1.04

R9 4290 1.46 0.34 167 3.36 23.45 0.80 0.67 1.24

R10 4295 0.975 0.34 83.5 3.36 16.9 0.85 0.68 1.02

Rll 3800 1.95 0.34 83.5 3.36 20.1 0.85 0.72 1.00
R12 4920 4.16 0.37 83.5 3.60 24.6 0.86 0.73 1.00
R13 4680 4.16 0.37 167 3.60 33.6 0.69 0.60 0.99
13.36
R14 4210 1.46 0.34 55.7 20.1 0.75 0.65 0080

R15 4330 4.16 1.49 167 3.60 31.4 0.82 0.69 1.04

R16 4580 4.16 2.98 167 3.60 31.4 1.03 0.74 1.06

R17 1850 1.46 0.34 83.5 3.36 15.7 0.72 0.58 1.06

R18* 4540 1.46 0.34 83.5 3.36 19.05 0.89 0.74 1.01

R19* 4390 1.46 0.34 167 3.36 26.9 0.71 0.59 1001

R20 6230 1.46 0.34 83.5 3.36 20.2 0.92 0.73 1.01
R21 6980 4.16 0.37 167 3.60 33.6 0.86 0.62 1.05
R22 4280 1.46 0.34 83.5 4.50 17.9 0.79 0.56 1.05

R23 4370 1.46 0.34 03.5 2.24 22.2 0.95 0.94 0.96

R24 4480 4.16 2.61 B3.5 5.05 20.7 0.95 0.68 1.11

R25 4470 4.16 2.61 83.5 3.60 23.5 1.20 0.90 1.02
R26 5300 4.16 2.61 83.5 2.40 33.6 0.98 0.98 0.87

R27 1980 4.16 0.37 167 3.60, 21.3- 0.62 0.60 1.35
140.3
R28 4580 4.16 2.61 326 3.60 0.79 0.74 1.13
.
* 450 stirrups

460
TABLE 8.3.2 T-Beams - Comparison with other theories
Ü
Beam 100 A r f - a/d Q Regan Laupa Hillerborg
yl s t g yq l ult
2 bd Qcalc Qcalc Qcalc
lbf/in 2
No r1 lbf/i n ki ps - -- . Qtest-,
Qtest Qtest

T-2 4070 1.46 3.36 12.3' 10.78 1.12 0.83

-
T-18 4120 4.16 3.60 16.8 0.85 1.28 0.99
-
T-28 4620 4.16 - 2.40 26.85 0.71 1.26 0.85
T-1 4050 1.25 83.5 3.36 24.75 0.63 0.64 0.74
T-3 4990 1.46 83.5 3.36 23.5 0.80 0.69 0.81
T-4 4710 1.95 83.5 3.36 24.6 0.77 0.84 0.86
T-5 4890 1.46 167 3.36 31.35 0.73 0.65 0.86
T-6 3740 4.16 326 3.60 46.0 0.96 0.72 0.98
T-7 3970 3,00 83.5 3.46 24.6 0.77 0.96 0.90
T-8 4530 4.16 83.5 3.60 28.0 0.73 0.92 0.86
T-9 2930 4.16 167 3.60 34.7 0.86 0.70 0089
T-10 4090 1.46 55.7 3.36 19.5 0.74 0.80 0.82
1-111 5360 4.16 167 3.60 36.0 0.83 0.88 0.93
1-121 4450 4.16 83.5 3.60 32.5 0.65 0.76 0.77
T-13 1850 1.46 83.5 3.36 20,2 0.72 0.57 0.82
T-141 4840 4.16 326 3.60 49.3 0.93 0176 0.94
T-15 4810 4.16 83.5 7.20 23.5 0.63 0.57 1.00
T-16 4740 4.16 55.7 7.20 20.8 0.64 0.62 0.96
T-17 4790 4.16 167 7.20 30.1 0.70 0.50 1.11
T-19 4340 4.16 83.5 5.40 25.5 0.63 0.67 0.92
T-20 4655 4.16 167 5.40 34.6 0.69 0.58 0.96
T-221 4980 1.46 83.5 3.36 24.6 0.71 0.72 0.79
T-23 4460 4.16 167 2.40 43.7 0.87 1001 0,79
T-25 7840 1.46 83.5 3.36 25.8 0.78 0.53 0.82
T-26 8260 4.16 167 3.60 40.3 0.70 0.51 0.89
T-27 1740 4.16 167 3.60 29.7 0.92 0.58 0.96
T-29 5350 4.16 83.5 2.40 34.7 0.82 1.19 0.84
T-30 4160 1.46 83.5 2.24 24.6 0.84 0.93 0.84
T-31 4495 1.46 83.5 3.36 21.3 0.68 0.56 0.89
T-32 4000 4.16 326 3.60 48.6 0.93 0.71 0.92
T-34*f 4920 4.16 83.5 5.40 25.2 0.66 0.60 0.92
T-35** 4880 4.16 83.5 5.40 25.8 0.64 0.78 0.90
T-36* 3500 4.16 167 3.601 40.3 0.74 0.65 0.77

461
TABLE 8.3.2 Continued

Beam Ucyl 100A rgfyq a/d1 Qult Regan Laupa Hillerborg

st Qcalc Qcalc Qcalc
2 bd 2
No lbf/i n r 1 lbf/i n ki ps
Qtest Qtest Qtest

T-37. 4615 4.16 j 326 3.60 47.1 0.97 0.79 0.97

1
T-38# 4380 4.16 326 3.60 53.8 0.86 0.68 0.85

1 45 °
stirrups

t 6" (t = 31' in all other beams)

'' b= 12" in T-34, b= 42" in T-35 (b = 24" in all other

beams)

TABLE 8.3.3 Statistical analysis of comparisons with other

shear theories

Type of Number Method of average Standard Coefficient

beams of Qcalc of
Calculation Deviation
tested beams --M- Variation
tested Qtest

Rectangular Regan 0.85(0.84) 0.124(0.130) 0.145(0.147)

with shear
21 Laupa 0.71(0.68) 0.114(0.081) 0.160(0.119)
reinforce-
ment (19) Hillerborg 1.04(1.05) 0,112(0,109)1 0.108(0.104)
T-beams Regan 0.77(0.76) 0.103(0.107) 0.134(0.140)
with shear
32 Laupa 0.72(0.69) 0.159(0.121) 0.220(0.175)
reinforce-
ment (29) Hillerborg 0.88(0.89)1 0.080(0.081) 0.091(0.091)
Beams with- Regan 0.79(0.77) 0.109(0.060) 0.137(0.078)
out shear 11 Laupa 1.09(1.04) 0.158(0.128) 0.144(0.123)
reinforce-
ment (7) Hillerborg 0.92(0.88) 0.112(0.056) 0.122(0.063)
Petersson

Note Figures in brackets are relevant to the statistical ana-.

lysis of beams with a/d1 > 2.5.

8.4 Discussion of the comparisons Presented in sections

8.1,8.2, and 8 .3

A summary of the statistical results obtained in the compar-

isons carried out in sections 1r 21 and 3 appears in table (8.4.1)

462
 Qcalc
The first figure in that table indicates the mean value
tost
and the second the coefficient of variation.
Qcalc
TABLE 8.4.1 Summary of mean and coefficients of variation
test
corresaondinU., to. the various theories examined

Type of ACI SSG CEB Reganl Laupa Hillerborgl Proposed

beam theory

Rectangular 0.72 0.71 0.60 0.84 0.68 1.05 0.91

with
(. 11) (. 10) (. 15) (. 15) (. 12) (. 10) (. 06)
stirrups
T-beams 0.59 0.61 0.49 0.76 0.69 0.89 0.91
with
(. 08) (. 09) (. 16) (. 14) (. 17) (. 09) (. 08)
stirrups
Beams 0.74 0.70 0.77 1.04 0.88 0.80
-
without
stirrups (. 12) (. 09) (. 08) , (. 12) (. 06) (. 07)
-- ;
It can be seen that in the case of beams with stirrups the

proposed theory gives the lowest coefficient of variation as

Qcalc
well as the least variation in the mean value for differ-
test
ent beam sections.

In the case of beams without stirrups the only other theory

having as low a coefficient of variation as the proposed one

is Hillerborg-Petersson's. The lower Qcalc value obtained in

test
that case by the proposed theory is due to the fact that the

coefficients involved in the prediction of shear cracking load

were kept intentionally low; as has been discussed this was

done because of the relative importance of dowel and interlock

action and the sudden nature of the failure involved.

Judged on an overall basis the proposed theory is thus the

most accurate one for predicting the ultimate shear strength

of beams with shear reinforcement and provides adequate pre-

dictions of shear strength in the case of beams without stirrups,

463
albeit slightly conservative for the reason just given.

Because of the limited number of tests no definite evaluation

of the particular performance of each theory in respect to

every parameter affecting shear strength can be made, although

some indications can be obtained regarding their overall per-

formance.

The A. C. I. and S. S. G. methods give very similar predictions

while the C. E. B. method seems to be a somewhat more conserva-

tive version of the same basic approach in which the ultimate

shear strength of a beam is calculated as being the sum of the

s
shear force supported at failure by the concrete and the shear

reinforcement. In the first two methods the effect of shear

reinforcement seems to be adequately taken into account, in

the sense that no systematic error seems to derive from it.

On the other hand there is a tendency in the CEB approach of

higher loads to be predicted when light shear reinforcement

is used, and the method seems more efficient for high degrees

of shear reinforcement; this can be explained by the fact that

it is based on the mörsch truss analogy which comes close* to

reality only in the presence of heavy shear reinforcement,

when the overall contribution of the concrete is relatively

small.

Actually the presence of important shear reinforcement

leads to high nominal shear stresses at failure and it would

seem that the limitations of the ACI code in that respect are

quite stringent for all sections while those of the SSG are

quite justifiably less severe. In the case of concrete of 4000

psi cylinder strength for exampleythe ACI recommends a maximum

464
nominal shear stress of 10JUcyl which is about 630 psi, while

the SSG recommends 5qc which in this case comes to a possible

maximum of 5x 155 = 775 psi. In rectangular beams a maximum

of 670 was obtained but in T-beams nominal shear stresses in

excess of 800 psi*wore sustained and strict application of the

ACI recommendations could lead to substantial underestimates

of the strength of T-beams whose unrestrictodly calculated

strength exceeds the maximum allowable. It is not suggested

that this upper limit be removed altogether but it should

certainly be less stringent for the case of T-beams. Actually

the reason for its existence lies in the consideration that in

beams with a high degree of shear reinforcement the increase

of strength due to the increase in shear reinforcement is not

a linear function of the latter as in the case of low degrees

of shear reinforcement and as implied in the code equations.

Although this is true in the case of rectangular beams, where

the increase in strength tends to become asymptotic, and is

taken into account by the proposed theory through the shear-

compression equation which normally governs in the presence of

heavy shear reinforcement - this is done by relating the con"

tribution of the stirrups to a varying crack length - it is

,

much less applicable in T-beams where a fixed crack inclination

was found to exist in almost all cases. The latter implies a

more or less linear dependency of shear strength on the degree

of shear reinforcement. This consideration, combined with the

fact that higher shear strengths are intrinsic to T-beams leads

to the suggestion that this limit could actually be removed al-

together in the case of normal T-sections; furthermore the web

465

I
crushing criterion will always ect. as a. safeguard in this

region of high stresses and high degrees of shear reinforcement.

When applied to rectangular beams, no systematic error in

connection with the degree of shear reinforcement seems to

exist in Hillerborg's theory and only, a very slight one in

Regan's where the effect of increased shear, reinforcement is

not quite adequately reflected in the predicted strength;.. this

has been, corrected in the proposed equations by taking into

account the, slight, beneficial effect of the stirrups on the,

_

compatibility, conditions. Hillerborg's theory on the other hand

has a tendency to-be unsafe but this is probably due. to an

overoptimistic evaluation of the effect of other parameters.

Since both theories were developed primarily for. rectangular,

beams their application to T-beams results in a number of dis-

crepancies, making an evaluation of their performance inýthat

case somewhat conjectural. Even Laupa's attempt to take account

of the influence of, in T-beams, which actually results

_flanges
Qcalc
in similar average values for both rectangular and T-beams,
Qtest

cannot be considered correct since the type of failure pre-

dicted by that method (shear-compression) did not correspond

to the actual physical behaviour, of the beams. The above

.

result can thus be considered coincidental. or, due. to the skill-

ful selection of the empirical constants involved.

Tho. influenco of the main reinforcement in. rectangular beams

is taken proportional to the cube root of the ratio of main

steel by the proposed theory, Regan and SSG.and its effect seems

to be adequately represented in this manner. Hillerborg's &

Peterssonts dependency on the square root of this ratio seems

466
to produce a tendency to overestimate its effect. This is not

very significant in rectangular beams where very high values

of rst are unlikely to be encountered. In T-beams however high

values of that parameter are common and could theoretically

lead to unsafety in Peterssonts case and logically even in

Hillerborg's, if it were not for the increased resistance of

T-beams - in the range of application of Hillerborg's theory -

not taken into consideration by that method. In the case of

the ICI code this effect is not really taken into account be-
Qdl
cause it is included in the term 2500 rst which is small
G1

'compared to the other terms of the equations. The progressively

higher values thus obtained by the ACI method for de-

Qcalc
test
creasing values of rst may theoretically have serious consequ-

ences in the case of beams with very low percentages of main

stool (r < 1%). This is also true in the CEB method which
st
basically ignores the effect altogether.

In the case of T-beams the effect of the main steel is role-

tively small and its influence together with the influence of

the ratio is. ignored in the present theory for beams with
d
1
s/
shear span to effective depth ratio greater than 3.0. In general

longer beams failing in shear contain more main steel than

shorter beams, as they would have failed in flexure otherwise.

In such cases the slight beneficial affect of the increased

main steel ratio is counteracted by a slight reduction of

strength due to the increasedd ratio of long beams. These

1
two actions compensate each other in the range of a/d1 values

greater than about 4.0; in the case of a/dl smaller than about

4.0 however, the increase due to rst is dominant and this re-

467
suits in a slight systematic tendency to underestimate the

shear strength of heavily reinforced beams in that range. How-

ever the error is rather small and on the safe side and since

the failure crack does not enter into the flange until failure

is imminent it may be assumed to be due to the increased dowel

action intrinsic to heavy main reinforcement, particularly if

the latter is arranged in more than one layer. In the present

state of knowledge of dowel action it would not be justified

to alter the basic equation for the sake of the rather small

differences involved in this particular case.

In short beams where the effect of the d ratio is more

1
evident all the code methods are conservative; as even the
increase predicted by the ACI through its term is quite
Qrd
1
small and insufficient to represent the actual increase in

strength of the beam. Laupa's method, while showing a tendency

to overestimate the strength beams, is on the other

of short

hand rather oversafo when applied to longer boams. Although

both Hillerborg's and Regan's seem to give good results

methods
Qcalc
in the case of short beams, in terms of values, their
test
performance in the region of low a/dl values should be com-

pared to their overall performance. It can be seen that al-

RQ
though these two methods give quite satisfactory values
tost
of the order of 0.8 to 0.9 in the case of short beams, these

values are quite different from those obtained when all the

test results are considered. They are higher than the average

values in the case of Regan's theory and lower in the case of

Hillerborg's.

In general it can be seen that no single equation or method

468
can do justice to the whole range ofd values - this is also
1
illustrated by the four tables of statistical analyses pre-

sented in sections 2 and 3 where it can be soon that when the

short beams are added to the results analysed, increased coeffi-

cients of variation are obtained in every case -. It is necessary

therefore that at least two equations or methods be formulated

to deal with cases situated on either side of a boundary value

of a/d1 of approximately 2.5 to 3.0. In this thesis the upper

range of a/d1 values has been investigated and shear strength

equations proposed reflecting the effect of the various para-

meters in that range. It is therefore necessary to undertake

the same task in the lower range as well, if rational equations

to be developed. The scatter of the existing results is

are

particularly important in that region and does not allow a

comprehensive the influent parameters to be

evaluation of all

carried out at present.

The influence of the concrete strength seems in general to

be represented by all theories in a more or less adequate

manner except in the case of the CEB method where the linear

relationship of the concrete term to the cylinder strength may

Qcale
lead to higher than average values in the case of beams
test
with high strength concrete and low degrees of shear reinforce-

ment. The slight overestimations or underestimations of. the

other theories are not conducive to the detection of any system-

orror, although a slight tendency for the ACI equation to

atic

underestimate the influence of the concrete strength may be

expected.
the/
When the various theories are compared in respect to influ-

469
ence of the ratio of compression reinforcement the following

general discrepancies can be detected.

1. When the method used ignores its effect altogether (Codes,

Hillerborg) then in the presence of heavy compression rein-

forcement the following tendencies exist;.

Qcalc
a) reasonable values when the degree of shear reinforce-
tost
ment is not high :
Qcalc
b) lower values when the degree reinforce-
of shear
test
ment is high.

2. In the case of the shear-compression methods that take

into account the effect of the ratio of compression steel

(Regan, Laupa) the following tendencies can be detected when

heavy compression reinforcement is present. -

a) high values when the degree of shear reinforcement

Qcalc
test
is low,
Qcalc
b) more correct Q ---- values when the degree of shear rein-
test
forcement is high.

Actually although 2b is obvious from the relevant tables

in the case of Regants theory, it is less evident in the case

gcalc
of Laupats sinco the value corresponding to a high degree
test
of shear reinforcement and heavy compression- steel is higher

than might be expected. This however does not invalidate the

tendency implied in 2b because of the following consideration.

The reason for the higher than expected value relevant

test
to this discussion, is due to the tendency, intrinsic in Laupa's

theory to overestimate the effect of the degree of shear rein-

forcement in the region of high rgfyq values; this tendency

results from the linear effect, implied in that theory, of

470
the degree of shear reinforcement on the increase of shear

strength.

The introduction of the shearing mode of failure in the

proposed theory eliminates these systematic discrepancies.

B. 5 Influence of CornprQs ion Flanges Comparison of

Rectangular and T-sections

In the course of tho tests reported in this thesis it was

observed that T-beams with shear reinforcement generally ex-

hibited higher shear strengths than equivalent rectangular

beams of breadth equal to the width of the web. This tendency

also appears in the preceding comparisons where it can be seen

that if a method makes no distinction between these two sections,

its application to the beams tested produces different values

Q
of average intrinsic to each type of beam'; the moan
Qcalc
tost
values obtained by these methods in the case of rectangular

beams are substantially higher than the' corresponding values'

for T-beams. This of course implies a higher shear resistance

on the part of the T-beams.

The above discussion is illustrated by table (8.5.1) where

it can also be seen that smaller differences are obtained when

the methods used take the effect of the flanges into consider-

ation.

471
TABLE 8.5.1 - Influence of flange in T-beams

Method Q Q A Approximate Remarks on the

calc B- ctealc B indication
A- of
Qtest Q
test increase of method
(Rect. ) (T--beam) in
strength
T-beams (p)

A. C. I. 0.72 0.585 1.23 23% flanges ignored

C. E. B. 0.595 0.49 1,22 22%

S. S. G. 0.73 0,61 1.20 20%

Hillerborg 1.05 0.89 1.18 18%

Regan 0.84 0.76 1.10 Not applicable flanges considered

Laupa 0.68 0.69 0.985

Proposed
theory 0.91 0.91 1.00 """
1

The increase in strength implied in these comparisons is of

the order of 20/. The fact that the proposed theory gives the

same Qcalc value in both cases indicates that the equations

test
developed in chapter 7 reflect this increase quite adequately.

The influence of the Compression Flange can also be ascer.

tained by comparing the test results of equivalent Rectangular

and T-beams as indicated in table (8.5.2).

472
TABLE 8.5.2 Influence of Compression Flanges coma arison

of roctancaular and T-sections

Beam Nos Approximate l100A rf a/d Q

I- - st q yq (ký 1 T
cy li nd er brd1 2 qR
strength lbf/in
(Psi)

R7 + T2 4500 » 5000 1.46 3.36 1.01

- »
R14 + T10 55.7 90 0 to 0.97
R8 + T3 133.5 " " 1.32
R9 + T5 167 1.33

R12 + TB 4500 » 5000 4.16 83.5 90 0 3.60 1.13

R13 + T9 167 " It 1.03
R28 + T6 326 it 1.14

R17 + T13 2000 2500 1.46 83.5 90° 3.36 1.28

R27 + T27 4.16 167 " 3.60 1.39
R20 + T25 6500 - 7500 1.46 83.5 " 3.36 1.28
R21 + T26 4.16 167 3.60 1.20

R23 + T30 4500 » 5000 1.46 83.5 90° 2.24 1.22

R22 + T31 1.46 83.5 " 4.48 1.19
R26 + T29 4.16 83.5 2.40 1.03
R24 + T19 4.16 83.5 " 5.40 1.13

R11 + T14 4500 - 5000 1.95 83.5 90° 3.36 1.22

R18 + T22 1.46 83.5 °

45 3.36 1.29

This table indicates an average increase of 19.5% in the

case of beams with shear reinforcement. If the short beams

are excluded the increase is 20.7%. This value of the increase

is of course only a general average and the actual increase in

a particular case can have a different value. This is however

taken into account by the proposed theory as can be seen by the

statistical analysis of the results appearing in this table, to

s
which the prediction of the proposed theory have been applied.

473
 Rectangular T-beams
Beams

calc 0.895 0.89

average Q
t es t

Standard deviation 0.069 0.078

Coefficient of variation 0.076 0.087

It can then be seen that the increased strength of T-beams

as implied in the proposed theory is justified; furthermore

this increase is predicted rationally since it is not detri-

mental to the value of the coefficient of variation which

remains small at all times,

ak6 Application of the roposed theory to T-beams tested

other investinators

The ultimate strength values predicted by the proposed

equation are compared to the actual test results for a number

of normal size T-beams with shear reinforcement that are re-

ported as having failed in shear or whose failure load was less

than their flexural capacity.

These comparisons, together with the details of the beams

considered, are presented in table (8.6.1). It can be seen that

the proposed theory gives quite adequate predictions since an

Qcalc
average value of 0.99 is obtained (the coefficient of
Qtes
test
variation being only 08) in the comparison of the theory with
.

25 beams tested by a number of investigators.

474
TABLE 8.6.1 Comparisons of proposed theory with test results

of other investigators

Beam Ucyl b br It dl d%E rgfyq Qcalc

Qtest
u lt Q -"-
2
in in ! in in in lbf/in lbf test
LEONHARDT A ND WALTHER (60)
TA3 2190 37.80 6.30 3.15 14.76 15.50 355 62400 1.08
TA4 2190 37.80 6.30 3.15. 14.76 15.50 219 52800 89
.
TAll 3540 37.80 6,30 3.15 14.76 15.50 355 76600 91
.
TA12 3540 37.80 6.30 3.15 14.76 15.50 219 60800 81
.
TA15 2480 37.80 6.30 3.15' 114.76 15.50 355 66900 1.02
(
TA16 2480 37.80 6.30 3.15 14.76 15.50 355 65900 1.04
OZDEN (85)
T-9 4420 15.75 4.3313.15 11.74 12.50 239 32200 1. Q1
T-10 4830 15.75 4.33 11.74 12,50 171 28800 0,96
T-15 4420ý15.75i6.3 3.15 11.74 12.50 118 31000 0.97
!
LEONHARDT & WALTHER (67)
ET2 3725 11.8 5.91 2.95 11.80 12.3 166 29000 1.08
ET3 3725 11.8 3,94 2.95 11.80 12.3 248 28100 1.04
GURALNICK (41)
la2R 2620,23.00 7.00 4.00 12.10 12.10 218 37900 1.11
lc2R 4930 23.00 7.00 4.00 12.70 218 50400 0,95
1d2R 4930 23.00 7.0014.00 112.10 12,10 218 49100 0.95
REGAN 94
!
D4 3830116.00 4.0011.5017.50 7.50 137 13450 0.92
TAYLOR (117)
ST-2 3145 12.004.50 3.00 10.20 10.50 198 23900 1.04

ST-3 3145 12.0014.50 10.20

3.00 110.20 10.50 187 23700 1.02

ST-6 2705 12.00 4.50; 3.00 10.50 142.5 19600 1,05

RDT `&'WALTHER(76)
LEONHA
TB1 12390 37.80 6.30 3.15 14.76 15.50 300 64000 0.93
T94 +2665 37.80 6.30 3.15 14.76 15.50 300 55200 1.08

TBS 1 2665 37.80 6.30 3.15 14.76 15.50 300 55200 1.08

T62 2390 37.80 6.30 3.15 14.76 15.50 300 53000 1.12
TB3 2370 37.80 6.30 3.15 14.76 15.50 300 65200 1.07
T89/1 2745 37.80 6.30 3.15 14.76 15.50 300 59500 1.00
TB10/11 2335 3.15 14.76 15.50 300 1 515001 1.15
37.8016.30

475
Table 8.6.1 Continued

Beam Ue b br t dl d* r fy Qcalc
Yl g q Qtest
2 ult Q
in in in in in lbf/in test
HAGBER6*
L3 2860 13.80 3.94 3.94 ; 11.42 12.62 615 17740 0,94
L4 2990 13.80 3.94 3.9' '11.42 12.62 506 17900 0.96
L5 3180 13.80 3.94 :.. . . 11.42 12.62 506 19900 0.88
L6 2790 13.80 3.94 3.94 11.42 12.62 506 19160 0.88

Mean Qcalc/Qtest 0.99

Standard Deviation = 0.08, Coefficient of Variation = 0.08

* Hagberg T. Versuche an indirekt gelagerten Plattenbalken..

Institut fur Beton und Betonkonstruktionen, Technische

Hochschule, Trondhciri

_calc
The value of the average obtained in this case is
Qtest

somewhat higher than that obtained in the tests reported in

this thesis. This difference is perhaps due to some extent to

the influence of scale effects that may exist between the pre-

sent series of tests and those carried by Leonhardt. The

out

value of d1 in these two cases is 10" and 15" respectively.

From the results of three series of tests where scale effects

(105)t
were observed - i. e. Busch Hangli and Mayor Leonhardt

and Walther(64) and Acharya(l) - it would seem that scale

effects do not disappear completely until d1 becomes greater

(28)
about 15". Do Cassiots28ý tests on the some subject

suggest that they are quite important for dl smaller than 6"

and in these comparisons only normal size beams were considered.

It remains however that there is some tendency for higher loads

to be sustained by smaller beams, even in the range of variation

of d1 used in the comparisons.

476
 Q
Since the difference the mean in
value is rather
Q
tecalcst
small and in view of the fact that no significant scale effects

can be expected to occur in beams with d1 greater than the

maximum 15" appearing in the comparisons, it can be reasonably

assumed that the proposed theory All apply adequately to the

normal size T-beams used in practice.

8.7 Comparison of the proposed web crushing equation with

the provision of the Draft Unified Code of Practice for

Structural Concrete (section 6.3.4. Shear

In section 7.5 the web crushing strength of a reinforced

concrete beam was found to be-given by equation (7.5.5) as:

75
gulf =s 12.5 + 2.5(rq%)IT- (1 + tan
budlt 1
r

where bRd* and Ucyl are expressed in lbf/int

r
The unified code proposes an upper limit for q=b of
r1
4qc which in this case can be construed as a safeguard against

web crushing. The value of qc expressed in N/mm2 is given in

a table form for varying ratios of main reinforcement and con-

crete strength.

Transforming equation (7.5.5) into SI units and rounding of

the constants the following equation is obtained (with slightly

conservative values for the constants involved so that all the

experimental points of fig. 151 in section 7.5 are above the

line representing the expression for K defined in section 7.5.3).

(1 J-Ucyl (1 + 0.75 tan L,') (8.7.1)

bud* = + 20rq) .....
r Qlt
where bud-* and Ucyl are expressed in N/mm2
r
For design purposes it can hardly be economical to increase

477
a beam's strength by adding more shear reinforcement beyond

the amount at which web crushing would first occur (rq

approx. 0.005). Thus for practical use the upper limit of shear

strength imposed by web crushing expressed in terms of cube

strength may well be taken ac: -

Ru FU (1
+ 0.75 cote) (8.7.2)
= qu .....
bd w
r
The results of using this equation are compared with the

recommendations of the Draft Unified Code below.

Characteristic Maximum q quit N/mm2 from equation (8.7.2

St reng t h of permitted by
Concrete N/mm2 Unified Code for p1. =90° for c _45
N/mm

15 3.20 3.87 6.77

22.5 3.60 4.74 8.30

30 4.00 5.4B 9458

37.5 4.60 6.12 10.71

45 4.80 6.71 11.73

52.5 4.80 7,25 12.66

It can be seen that the Code values give a partial factor

of safety for materials of from 1.2 to 1.5 for vertical stir-

rups which values are acceptable. The corresponding factors

for 450 stirrups are from 2.1 to 2.6 and ara unnecessarily

restrictive. It appears that the Code should take some account

of the superiority of inclined stirrups in respect to web

crushing, as equation (8.7.2) is itself a consorvative simpli-

fication of equation (8.7.1), the safety of which is clearly

demonstrated by figure 151 and increased by the slight trans-

formation of the constants mentioned previously. The consorva-

tism is not excessive since most of the test pieces had rela-

478
tively stiff flanges which doubtless increased the beams'

strengths - the equation necessarily predicts zero strength

when br_0 while in fact there would still be some strength

even in the absence of a web.

479
 CHAPTER 9

SERVICEABILITY CRITERIA

The equations developed and discussed in the preceding

chapters are relevant to two load stages; namely shear cracking

and shear failure. Any design method however, while it can be

based on these equations and appropriate safety factors, must

also ensure that the behaviour of the beam at possible inter-

mediate stages is not unsatisfactory. In other words, criteria

other than strength must also be satisfied at service load

level. The two criteria that must be satisfied at this level

are deflections and cracking.

Because of the rather low limit of ratios at which shear

1
failure can take place, (max. a/dl of 7.0 to 7.5 in the case

of T-; beams, substantially lower in rectangular beams) deflection

conditions at working load level are not normally critical in

beams failing in shear. On the other hand shear cracks tend to

become wider than flexural cracks and may in fact reach limit-

ing values for crack widths - as prescribed by the codes - at

early load stages.

In this chapter the behaviour of the beams tested - all of

which failed in shear - is examined in terms of these two

factors and a formula for predicting shear crack widths is

presented.

9.1 Shear Crack Widths

Although crack widths have been measured at times by

shear

various investigators - by Leonhardt and Walther in references

480
(66) (69) for has been made at
and example - only one attempt

producing a formula for them. Perhaps the fact that even in

the case of flexural cracks the existing formulae show a sur-

prising scatter when applied to results other than the ones

that helped to derive them, has prompted the various research

workers to make only general statements on the subject of

shear crack widths. This is unfortunate in view of the present

realisation that the necessity of limiting crack widths can play

a decisive part in the designs of reinforced concrete structures.

To be on the safe side one could of course follow Peredy and

Visy's(88) concept that shear cracks should not be tolerated

at working loads. This conservative suggestion could possibly

be acceptable for the case of rectangular beams but it would

often be uneconomical to limit the ultimate load of T- and

I-beams to approximately twice the load causing shear cracking.

In general there are two main approaches to the approxi-

mation of the load-shear crack width relationship and these

are shown graphically in fig. 184in terms of the relationship

implied in these methods between load taken by the shear rein-

forcement and the total load applied.

True
shear y
taken by
stirrups
Borishanski proposed
/(

Applied Shear
Fig. 184 Idealisation of the basic approaches to the shear

crack width problem.

481
 The only method presently in existence gives crack width

in terms of the applied load disregarding the fact that prior

to shear crackingno load is taken by the stirrups in fact. In

the proposed method this is taken into account by establishing

a relationship between crack width and Q-Qcr. The latter of

course is not rigorously true but the two possible objections

to it can be shown to be of importance by the following

minor

considerations.

a) Although small inclined cracks exist at loads just prior

to the shear cracking load Qcry their is Very small
width and

always well within the allowable limits.

b) The shear taken by the concrete Qc can be greater than

Qcr' especially in T-beams. However the bulk of the increase

in the load carried by the takes the

concrete place once stir-

rups have yielded; and such late load not important

stages are
for crack widths which have by then the
exceeded permissible
limits. Up to the level the the load
of yielding of stirrups

carried by the concrete can be reasonably by Qcr"

approximated

Because of the fact that it is the only approach presently

available, Borishanski's method is presented in the following

section. However, because of the poor correlation between the

predicted and measured widths of the cracks a new formula is

proposed based on observations during the tests and the con-

sideration on Qtr described in the preceding paragraph.

9.1.1 Calculation of shear crack widths according to Borishanski

The derivation of the equations in this method is parallel

to the general approach used in deriving expressions for

flexural cracks, i. e. based on the spacing of the cracks; this

482
actually constitutes one of the criticisms of the method as

not only the stresses causing the cracks are different in the

two cases but the conditions of spacing and opening of the

cracks are not compatible either.

The crack width w is taken as

f
ýEsl`
st
where lt spacing of cracks

fs = stress in reinforcement

t1 _ coefficient dependent on load stage

The spacing of flexural cracks is lt =T rl where u is the

ratio of the cross-sectional area of the reinforcement to its

perimeter, q is dependent on the surface of the

a coefficient

bars (for plain round bars q=1.0. for deformed bars n, = 0.7

and for smooth wires R=1.25) and p the reinforcement ratio

in a directly tensioned member.

A similar equation for lt is applied for inclined cracks

If
with the introduction the properties of the concrete and

the transverse reinfdreemente The essential difference between

the two cases is justified by the fact that in a tensile member,

after the first crack has been formed, the concrete is stressed

only by bond with the reinforcement, while in the web, in

to bond, the concrete is stressed by its contribution

addition

to the support of the shear force. According to Borishanski,

it would at first thus appear to be possible to take lt as

being for inclined than for vertical cracks. However

smaller

he notes that the bond between the concrete and the reinforce-

an inclined crack the spacing

ment is not symmetrical about and

483
of inclined cracks is therefore more irregular than that of

vertical cracks. This second consideration was considered to

be dominant and it is proposed that lt should be taken to be

30% greater for inclined than for vertical cracks.

When the shear reinforcement consists of one type of stir-

rup the equation for the spacing of inclined cracks is thus

lt=3

AA
where u=xx and 0 is the bar diameter
xbs
When lt calculated by the above expression exceeds dl it is

recommended that its value be limited to dl +30g,,

When stirrups and bent-up bars are-used together, the com-

bination of their effects is assumed to be algebraic rather

than and in that can be
geometric, case - inclined stirrups

treated as bent-up bars to this method

according -

xxo0

In the above the suffix "x" applies to the stirrups and the

suffix "o" applies to bent-up bars.

The crack width equation also contains a term fsp the

stress in the transverse reinforcement. To evaluate this it

is assumed that from the stage of inclined cracking up to

failure the shear is distributed between the concrete and the

reinforcement in a constant ratio. At failure the reinforcement

yields; thus
RaQ

5 Qult

where Ra is the design stress of the reinforcement.

This hypothesis is not always correct. It is more accurate

484
the nearer Q is to Qult'

Another term in the crack width equation is the coefficient

kýi which also varioa as Q approaches Qult'

It is therefore assumed that the product w fs be given the

following value: -

(-Q9---) 2
ý1 f=k. R
ax .
ult
If the shear reinforcement consists of vertical stirrups

only, the plane of minimum strength method, as prescribed in

the Russian Code gives the following

Quit ' 0.6 b d12 Rh qx

where Rh = design compressive stress for concrete - taken as

Ucyl by Borishanski in reference (17) -

axRax
qx = stirrup force per unit length of beam ýS
x
R
0, =X Nxs S_x_Rax
replacing qx_-: -cs
.xx
the expression for ý1 f, becomes
22
41 fk- -------- kR
b2dl2Rhpx hpx

Putting this value into the expression for crack width

k2
w_E-. lt
Wx
ct
In this equation k is a coefficient accounting for a number

of factors not otherwise considered - i. e. non uniform distri-

bution of forces between the stirrups, influence of main rein-

forcement etc. Its value is established empirically and is

taken equal to 4.0.

The final equation is thus

42
1
Est Rhyx t

485
in the of bentup bars the corresponding equation is
presence
42
W +T 1
RhEst(NXý t

Unfortunately the correlation of this formula with Boris-

hanski's own results is not very happy, as cen be seen from

fig. 185. In this figure the widths at design loads are com-

pared with those predicted by the crack width equations. The

abscissa is the percentage of transverse reinforcement, and

the ordinates are the crack widths. The actual widths are

indicated by circles and the calculated ones by crosses. The

two points for each beam are connected by a line, the length

of which is thus the error of the equation; full lines are

used for the cases in which the actual crack width was greater

than the calculated one. From the figure it can be seen that

in many cases, particularly when the percentage of transverse

reinforcement is small, the errors are considerable, - in some

instances as much as 0.3mm4 As the percentage of transverse

reinforcement increases the crack widths decrease and smaller

absolute values of discrepancies are obtained. Viewed in terms

of percentage of error though, the error is just as important;

furthermore the error is not always on the safe side as in a

number of cases the calculated width was smaller than

substantial

the actual one. It remains however that in the majority of

beams the calculated widths were greater than the actual.

The above remarks are also reflected in the comparison of

the predicted by this formula with the crack widths mea-

values

during the tests reported in this thesis, in beams that

sured

did not fail shortly after shear cracking. Inspite of many

486
discrepancies the general tendency of that formula was to pre-

dict crack widths varying from approximately one and a half

times the actual width at high load levels in beams with a

high degree of shear reinforcement to about two and a half

times the actual width at relatively low levels of loading in

beams with an intermediate degree of shear reinforcement. These

figures result from calculations based on an Rh value equal

to the cylinder strength (such a value for Rh was used by

Borishanski in the comparison of hie theory with other test

results which appeared in a CEB Bulletin (17)).

Because of the fact that such predictions of large crack

widths would tend to make crack widths the only design criterion

it was decided to try and establish a formula reflecting more

closely the results of the present series of experiments. It

must be pointed out however that the tests examined had stirrup

spacings S O. 5d1. When larger spacings are used cracks of greater

width are likely to occur. Since the maximum spacing recom-

mended is 0.75d1, a linear interpollation between the pre-

dictions of the proposed formula and that of Borishanski should

produce acceptable answers, in the region 0.5d1 <s0.75d1.

9.1.2 Proposed equation for redicting shear crack widths

During the tests reported in this thesis the shear crack

widths were measured in all three types of beams used; i. e.

1) Rectangular beams with plain round mild steel stirrups

2) T-beams with plain round mild steel stirrups

3) Thin web I-beams with high yield deformed stirrups

If the shear cracking load of a beam with stirrups is de-

fined as the load at which an otherwise similar beam without

487
 would fail in diagonal tension, then small, inclined
stirrups

cracks may exist at lower loads.; These cracks are never of, such

width as would render a-structure unserviceable. In design one

is concerned with load stages between shear cracking (as de-

fined) and the load at which the cracks attain widths of the

order of 0,008" to 0,0151, (0.2mm to 0.4mm). Such widths are

generally attained prior. to the yielding of the stirrups. In

this range of loading shear crack widths seem to depend on:

1) The shear force carried by the stirrups, which may bo

represented by Q-Qcr*

2) The ratio of shear reinforcement,.

3) The quality of the concrete

4) The spacing of the stirrups

The tests relate to the case of closely spaced stirrups

(s 0.5dl). When wider spacings were used the value of rqfyq

were so low that yield of the stirrups and failure of the beam

took place soon aftor'the shear cracking load, and no shear-

cracking was present at the service load (= approximately

0.5 x ultimate load). If higher rgfyq values were used with

wide spacings, the critical cracks would lie between stirrups,

causing low values of Rult9 and thence of the service load;

this again normally means that no shear cracks would be present

in the service condition. If however in a given case of high

rqf value and wide spacing none of the above is satisfactory,

Yq
an approximation to the shear crack width might be obtained

by an interpollation between the proposed and Borishanski's

in the section.
method as discussed preceding

When considering shear cracks it is not possible to apply

488
the type of theoretical approach used for determining the

widths of flexural cracks, because the spacing of the cracks

is variable during thu relevant load stages and the cracks

are not of even approximately equdl widths. In rectangular

beams one crack tends to predominate, while in T-beams new

cracks are progressively formed and those that are wider at

early load stages are not necessarily the widest later on.

Proceeding empirically by plotting crack widths against

the parameters considered the following relationship was ob-

Q-Qcr 2
tained. ý- -I
brdI

moo 0
100rq(Ucyl)

where w= width of widest shear crack in inches and both

(Q-Qer)/brdl and Ueyi are in lbf/int.

If the units are kgs and cm, as in Borishanski's equation

the expression becomes:

(Q-Qcr)/brdl 2
w=1.3 -ý -- - (9.1. la)
.....
'1000rq(Ucyl)
Q-Q
where w is in mm and bd and Ucyl are both expressed in
r1
kgf/cm2.

Figures (18? ) and (188) illustrate the above relationships

for various types of beams with the crack width expressed in

ins and mms respectively (if s. i. units are used and Ucyl and
Q-Qcr
expressed in N/mm2 a change in the constant of equation
b
rd1
9.1.1a to 30 instead of 1.3 will produce an approximately

equivalent equation in these units).

It is worth noting that equation (9.1.1) draws no distinction

between vertical and inclined stirrups and the irrelevance of

489
 ý 1-0
LE
U

0.6
D5
V xýF 'i,
%f#
.ýO
Ix
L 0.2 jj 1l YX iIx
ý lý s
li

0.2 0.4 0.6 0.8,1.0 1.2 1.4 1.6 1.8 2.0

Perc e ntag2 of transverse reinforcement

Fig. (185) Comparison of Borischcnski's crack width formula with

his own tests.

Note on - Fig. 086) The curves for maximum crack widths are
similar albeit with higher ordinates

So
0
r

.
3:

U
v
L
U

Urn
U
L
U

0 40 80 120 160 200 t 240

total load on bzcm in tonnes

Fig. (188) Crack widths of b earn. foiling by web crushing

69
Tests by Lconhardt

49O
 "a

20 20
0
width of width of
veldest shear 3 widest shear
treck in in Y.10 crack in in x 103

O
oOO
0ý
'"0 10
/31
CO
o

poo 0
p ov o (QOcý/brd1 pD0 (Q -Q. )/brd
1GJ'q (Ucyi) 3 100rq(Ucyi)

10 20 30 40 10 20 30 40
Fig. (a) Rect. beams Fig (b) T beams

KEY

OV CIAO beams width 90° stirrups

0
fl 0o beams with 45° stirrups

Theoretical lines represent equation (9.1.1)

0
w: 15A2

0 0cr)/ brdj
where A1(
Q
1000 1O0rq (Ucyl )6
Q'Qcr/ bA
A0
900rq (Ucyl)
°ol
10 20 30 40
Fig. (c) I beams

Fig. (187) Crack widths

4<
S
 "ýo
0.5 as

C D
0.4 0.4
width of width of -
widest shear widest shear
crack in mm crack in mm
ý c 0
0.3 0-3

0.2 0.2
0
V O
00 C>O
0.1 0.1
0- 0 1 I
(Q-Qcr)/ OD
br dt (Q-Qcr)/brdl
1000 rq (Ucyi )''S 0 1000rq(Ucy1)3

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.1 0.2.0.3 0.4 0.5 0.6 07 0"E +
F; g. (a) Rect. beams Fig. (b) T beams E

n.

ov
0.4 KEY
width of
widest shear 0V 13Ao beams width 90° stirrups
crack in mm
o
O °00 beams 450
with stirrups

Theoretical lines represent

equation 2
v (a-Qcr)/ br d,
w: 1.3
o p [iooo
rc (Ucyi)113
A
O.1 "
DO
0 (Q-Qcr)1brdi
v
10OOrq(Ucyi)ý3

01 0.2 0.3 0.4 0.5 0.6 0.7

Fig. (c) I beams

Fig. (188) Crack widths

I

ýý
inclination in this context is borne out by the test results

of rectangular and T-beams failing in shear. The result is

understandable since with crack inclinations of the order of

2: 1 (0: 27°) even more for the inclined

or stirrups, both
stirrups
vertical and 45°/make much the same angle with the crack. This

however is not so in the case of web crushing failures where

the crack widths are much smaller when inclined stirrups are

used. However in that case they are even smaller than the

flexural crack widths as was also noted by Leonhardt - see

fig. (166) and it can be assumed that if a beam has an

-

extremely high degree of inclined shear reinforcement, the

flexural cracks will govern if web crushing would take place

in a beam of equivalent rq value but with vertical stirrups.

The equation is applicable to rectangular, T- and I-beams

of normal proportions. When an exceptionally stiff flange is

provided the crack widths seem to be somewhat reduced.

The experimental results for T-beams show a greater scatter

than those for the other two types. The reason for this is

probably that in rectangular beams the same crack is generally

widest at all stages, and in thin-webbed I-beams all the cracks

are of much the same width, while in T-beams new wide cracks

overtake the original ones, so that the observations of the

greatest widths relate to a series of different cracks. The

consequent measurement of different cracks cannot be expected

to give as smooth a curve as is obtained for the other types

of beam. However, expressed in terms of the shear force Q

producing crack widths of a specified magnitude within the

range (0.2 to 0.4mm or 0.08 to 0.015") the maximum

allowable

493
error in Q is only about 15%.

Study of-the shear crack width equation together with the

ultimate load equations shows that with the ratio 'collapse

load/service load' equal to 2t an allowable crack width of

0.01011 or 0.25mm will not be reached at the service load so

long'as mild steel is used for'the shear reinforcement. If

other steels are used the design of shear reinforcement may

be governed by the crack width criterion.

9.2 Deflection of beams failing in shear =

The load deflection relationship for a structural member'

is an important one. Load-deflection curves are useful-in

studying the behaviour of baams at working loads and for esti-

mating the range of acceptable working loads. The load de-

flection curves obtained from tests of beams failing in shear

differ from those of beams failing in flexure, since for beams

failing in shear part of the ductility is contributed by"the

opening-up of diagonal cracks. Typical load-deflection curves

illustrating the effect of the various parameters are presented

in-figures (189) to (195) for beams failing in shear.

Although this is not always' apparent from all the 'load--

deflection curvest-the latter can be construed to containtý

theoretically at least, four more or less distinct stages.. -

1) The first stage corresponds to that part of the curve-before

flexural cracking. In this stage the beam behaves-elastically,

and the load deflection. curve is astraight line, theýdeflection

being proportional to the load. The extent of this stage depends

largely on the tensile strength of the concrete and the shear

494
 span. Knowing the properties of steel and concrete, the flex-

Ural cracking load and the amount of deflection up to this

load can be calculated on the basis of an uncracked section

analysis.

2) The second load stage represents that part of the curve

between the beginning the flexural

of cracking and the form-

ation of a major inclined crack. After the appearance of flex-

ural cracks, the deflection curve takes the first change of

curvature and follows a smooth curve deviating only slightly

from the original straight line. The presence of shear rein-

forcement has no important effect on the load-deflection curve

in the first two stages and the behaviour of beams failing in

shear and flexure should be reasonably similar.

3) The third stage of the curve that part the

represents of

curve after the formation of the major diagonal crack. In this

stage the load deflection curve undergoes a second change of

curvature of varying importance, depending on the shear span

ratio and the degree of shear In beams without

reinforcement.

shear reinforcement this is relevant only to beams

short
(2-<3) as the longer fail in diagonal
d1 ones tension near that

load stage and the working load calculations to

are relevant
the previous stage. In short beams without reinforcement the

deflection values beyond this stage exceed the values calculated

from bending alone because considerable shear deformations occur.

The load-deflection curve in this stage becomes more erratic

as it is influenced by the width and number of diagonal cracks

and the extent of their penetration into the compression zone

or their extension in the web-flange junction. Because of the

495
fact that deflections are used as a criterion only in connection

with working loads and they are hardly likely to govern in

the short beams under discussion, measurements were not taken

up to failure. However this was done by Leonhardt and Walther

(64)
who reported that in beams with of 2.0 and 2.77 the
d
1
shear deformation attained, shortly before failure, the magni-

tude of the flexural deformations. They also add that the shear

deformations do not assert their influence until loads in ex-

cess of the working load are reached. Even if working loads

higher than those considered by Leonhardt are used, the increase

in deflection due to the shear deformation will still be quite

small since its increase is gradual and only a small fraction

of the final value would be manifest at working load levels.

From a prediction point of view this small increase can be cal-

culated by interpol ation with a final value of twice the bend-

ing deformation but this need be done only when the increase

in load above shear cracking is very considerable. In practice

short beams are not critical from the deflection criterion even

if this small increase, if it takes place at working load, is

taken into consideration.

For beams with web reinforcement, the widening of the dia-

gonal crack is controlled by the shear reinforcement and in most

s
cases the change of slope of the load doflection curvo hardly

noticeable. In general there is loss of a change when more

stirrups are used but the difference is rather small as shown

by these tests and Leonhardt's series E(64) for rectangular

(6B)
beams and TA for T-beams. A similar trend can also be ob-
(117)
served in the deflection curves for T-beams reported by Taylor.

496
This moans that calculations taking into account the bending

deformation only. cannot be very much in error except perhaps

when very light stirrups are used. In that case however the

failure load is smaller and at service load levels the increase

due to the shear deformation is unlikely to have materialized.

Scordeli's(20) tests of similar beams with and without shear

reinforcement (series OA and A) show that the deflection curves

in the two cases are almost parallel. Naturally the curves

intrinsic to the beams with shear reinforcement extend further

than those of boome without shear reinforcement which failed _

in diagonal tension.

If however instead of increasing the amount of stirrups, a

high degree of shear reinforcement is obtained simply by re-

ducing the web area and the beam is likely to fail in web

crushing, the shear deformation can be quite important. Leon-

hardt and Walther have shown that if the shear reinforcement

remains the same, the deflection increases with decreasing

(67) (web
thickness series ET. Dilger33ý has investigated this
-

question in the case of web crushing failures and evaluated

the shear deformation by using a modulus for shear deformation

dependent on the percentage of shear reinforcement. His results

for Leonhardt and Walther's beam Tl show shear deformations

almost of the order of the bending deformations even at working

load levels. In general the relative importance of the shear

deformation seems to depend on the web thickness, shear rein-

forcement and a/dl ratio. It can thus be seen that a beam with

a larger rgfyq, or rather rq, value than another will not

automatically deflect less even at the same a/dl value, unless

497
of course the two beams have the same web thickness.

In the present tests for example, the deflection behaviour

of beam T-9 having an rq value of 0.417% and a web of 6" was

almost identical to that of beam WI having an rq value of

2.17% and a web of 2.511. On the other hand the behaviour of

beam T6 or T32 with rq = 0.815% and br =611 was slightly better

than either T-9 or W1. All three beams had the same a/dl ratio,

amount of main steel and concrete cylinder strength - see fig.

(194).

In general the slight change of slope of the load-deflection

curve at shear cracking indicates that the extension of the

curve of stage 2 represents adequately the load deflection

behaviour at the early stages of stage 3 corresponding to a

maximum possible working load. However the load-deflection

curve at stage 2 is basically a flexural one and the small

change in curvature also implies that the effect of the shear

deformation at working load levels is If thus the

small.
flexural deflection behaviour can be adequately predicted for

stage 2 the same method can be extended to give an approximate

value, by default, of the deflection at load level,

working

when the lattor is in excess of stage 2. A further refinement

of this question does not seem warranted because of the fact

that it is only in rather short beams (tl < 3.5 to 4.0) that
1
web crushing takes place and deflection is not then a critical

factor. The only other possibility of increased deflection lies

in the presence of only light shear reinforcement, but as

0
mentioned previously, failure in that case, in any long beam

where deflection can be critical, will take place shortly after

498
shear cracking, thereby reducing the working load calculations

to loads below shear cracking.

4) For beams with web reinforcement failing in shear there is

a second change in the slope of the load deflection curve when

the stirrups start yielding. Since this also means that failure

is imminent no measurements were taken of the rather large de-

flections that resulted. The magnitude of these deflections is

quite irrelevant to the use of deflection as serviceability

criterion since at working load level the stirrups will always

be in the elastic range.

The four stages described above are not always distinguish-

able in the load deflection curves drawn from readings at

reasonable load intervals. Because of the great number of

beams tested and the secondary nature of this problem it was

decided not to multiply the number of load stages in order to

obtain a better representation of the actual behaviours. It

remains however that a better understanding of the beam de-

flection behaviour is obtained when the various load-deflection

curves are viewed from the angle of the four stages described.

Although the above stages describe the general pattern of

the load-deflection curves, various parameters affect the

particular load-deflection curve of a given beam. To illustrate

this the load-deflection curves are drawn'in different groups.

As far as possible each figure contains the data for beams

having the same or similar values of the variables not con-

sidered so that the differencJin the curves reflect roughly

the effect of different factors on the deflections of beams.

499
How each factor affects the load deflection curves is briefly

described below.

a) Concrete properties The relevant fig. (189) shows the way

in which, in all cases, the deflection increased as the

concrete strength was reduced.

b Shear-Span The relevant fig. (190) indicates the manner

in which the deflection increases with increasing a/d1 ratio.

).
2. Percentage of tensile reinforcement The relevant fig. (191)

indicates the manner in which deflection increases with de-

creasing ratio of main reinforcement.

Al Percentage of compression reinforcement Both Ramakrishnan's

and Wilby's tests show that this parameter has no effect on

the deflection of T-beams; the relevant fig. (192) refers

therefore only to rectangular beams, where the manner in which

the deflection decreases with increasing ratio of compression

reinforcement is indicated

e) Cross-section The effect of the shape of the cross section

can be discerned in most of the above figures where deflection

of similar rectangular and T-beams are presented. It can be

seen from these figures that tho addition of a flange stiffened

the beam considerably from the inception of loading. Figure

(193) illustrates further the manner in which this effect takes

place by comparing not only rectangular and T-beams but also

T-beams of varying flange width and breadth.

f Web reinforcement The effect of the stirrups in roducing

deflection, as discussed previously in terms of various stages,

is illustrated by fig. (194). The last stage, involving the

yielding of stirrups and increased deflection is not shown in

500
this or the preceding figures as it is the load-deflection

characteristics in the range of possible working load levels

that are of any significance in terms of serviceability.

In two tests of beams failing in web crushing Leonhardt

compared the effect of the stirrup surface on deflection by

placing plain stirrups on one side and deformed stirrups on

the other side of a beam loaded by two point loads and measuring

the deflection near the middle of each shear span(66). From an

examination of the load-deflection curves obtained, it appears

that deformed bars resulted in a fixed reduction of the deflect-

ion of the order of lmm at all but the very early load stages,

when vertical stirrups were, used. A similar but increased re-

(= 2mm) was obtained were

duction when inclined stirrups used.

The deflection at failure at these measuring points was about

10mm in the case of vertical stirrups. At that load the corre-

sponding deflection in the beam with inclined stirrups was 7.5mm.

When inclined stirrups are used there is a tendency for the

deflection to be reduced, which confirms the findings of other

(66)(78)
investigators and is generally related to the better

overall performance of beams with inclined stirrups and the

direction of the latter in respect to all the inclined cracks.

In fig. (195) the deflections of beams having stirrups inclined

at 450 are compared to those of similar beams with vertical

stirrups. It can be noticed that this decrease in deflection

is rather random in the case of T-beams in the sense that the

beam in which this reduction is a maximum does not exhibit the

maximum increase in strength. On the other hand the reduction

does seam to increase with increasing loads and amounts of

501
 30 P
30

20 20 c,

FZ'i a1d1=3.6
10 10 (rgfyq_157)

rq : "417°lo
Ast_ߢlörrnm

0.10 0.20
0.10 020
Mid-span deflection in ins Mid-span deflection in ins

30

20

Note : the number in brcckcts

indicates ultimate load

10

010 0.20

Mid-span deflzction in ins

Fig. (169) Effect of Concrete Strength on Daflcction

50,
 N
n
E co
ö E Ö
I,
V I-
v O
CD N '9"
CO
H O
N N
L0" Q
If) r
0
H

ico
A
Ö
th .
r-
_

N k- (0
u c 1ý - c) a ýA
" Co
::. CO 0 YD-
uN CO '
r
C [T 0 u
C
ag 4.0
V L Q
(V d vL C
IN w O
C) Ü
Z7

O C OU
o ß
a
1 C7

6
11_ cri
u
, CS

1 CV "0 CyS
94
N- cl ö " v O
n f. - N
N
iý N
Oý B N
CD
N
Hý
O F \ 0
N
0 Ö
c.1

0
00 `- 0
N
N
suoI ul pco1 suo}.. U! PDO-1

Fig. (190) Effect of a! d1 on d7flection

so
 30 30

0
,J

T8(25T), rst. 4.16°lo

c q0
R12(22T), rst 4.16°l0 T3(21T), rst_1.46%
20
"- T7(22T),
0
J R11(18T),
r'st'1.95°1o v rs -30°l0
0
R10(15T),
10 rse0.975°lo 10
a /d1 : 3-3-3-6
aid1 =3-3-3-6
(rgfy03.5) (rgfy- : 83.5)
q
rq : 0.208111o rq : 0.208°l0

0.1 Mid-span deflection in ins 0'1 0.2

Fig. (191) Effect of ratio of main steel on deflection

30 30

c C
0 O
4J
C
R15(28T),rs6:2
.G
301),r, ý0.37°I°
R13(.
2C ., 20 R25(21T).
v
v g 2'Ö1%
0
rre R12(22T),
J
rsc=0.37°I°

10
a/d1 =3.6 a/d1 : 3.6
(rgfy,,: 167) (rq fyq: 83.5)
rq: 0.417°J° rq : 0-20811*
Ast, 8016mm Ast : 80,16rm

. 0.1 0.2 0.1 0.2

Mid-span deflection in ins Mid-span deflection in ins

Fig. (192) Effect of ratio of compression reinforcement on GefI cticn

I-
ý.I
 4or T38 (48T) 40
t_6
T6 (41 T)
t= 3p
T36 (35T)
::: 30 t_6"

T9 (31T)
t=3°
6T) 2
R 0

s1 °!o
20 R13 (30T)

'o
v
0
J
rII old, = 3.6 /j a/d'=3.6
10 10 /
(rgfyq=326) (rq fy,: 167), rq = 0.417°/0
rq_0.815°1° Ast :8 016 rnm
Ast: 8016 mm br_ 6b= 24`'
t'r=6" b: 24"

0.1
0.1 0.2
0.2 0.1 0.2
Mid-span
Mid-span deflection
deflection in ins
ins Mid-span deflection in ins

N
C
O
0
4'
4'

30
.S

8
J

T34 (225T)
20 b: 12"T35(23T)
T19(22-5T) b42"
b24"
R24 (18.5)
rsc-2 61°lA
10 aldi: 5.4
(r9 fY9= 83.5)
rq: 0.208°I°
Ast :8¢ 15rnrn
br-_6'', t: 3°
01 0.2
Mid - span deflection in ins

Fig (193) Effect of cross section on deflection

sn- J
 30 ý-- - ::

R28 (36T)
C
rc. 4315°I°

20 R9(21T) 20
v0-41701z R15 (28)
rq _ 0.417°I/
R14 (IST) 0
0
rq= 0.14?1. R25 (21T)
10 10 rq="206°Ib

a/d, =3.36 ald,=3.6

Ast = 3ý16mm Ast = 8t 16 mm

I
0-1 0.2 0.1 0.2
. .
I
Mid-span deflection in ins Mid-span deflection in ins

f ý-

T6 (41T)
rq: "815°I°

41
30 3 W5 (33T)
.
rq-3.25°%°
C
0
4J Wi (32T)
0 T5 (28T) rq=2.170/3
T8 (25T)
.rq . "417°l0 "ý
2 rq. "2tß$°13 T9(31T)
20 ö
'---*T3 (21T) ra: 0.417°1°
rq:-"2C3°l° J

T10 (17.4)
rq_ O.14°l°
10 101
' -'-T18(15T)
a/d, 3.36 old =3-6
rq n01
T2(11T) Ast=3¢16rnm Ast: 15mm
rO

0-1 a2 0-1 0.2

Mid-span deflection in ins Mid-span defiecticn in ins

Fig. (194) Effection of shear reinforcement on dcflaction

 N
ö O
tD
r
(D 0
(Y) N CO
6
I-
ý
ý O LýQ N
` ` N
ý-
O
\ v p
rtr
\ \ 0
ýo
O
- N
co . -O to
i- \ r to

Nom \ V N
`` \ I'
^ LCT' Q
f-
d
2 ý

cli

°tt) ýý C7 L'ý Q ýi
o

) E
E
\ý MO
CO CO co
p E f
N

lb 10 0
... oO f.. ' 1'
1o 8,
0) M %% v s.ý Q
\ cc 0) M co ` ýý
ýy
4°Lr) `\ "-
.0 It 4j T-
%

%
d\ 6
`
cc d ýý
r

E
E E
EN (0 ö0 cV
F- C_0 cl) Co ö
,° H N -oý
\ Z> o I- t;
CV t!
co O (y) V) +. a
O ei
t7
A Ö Q
Lff
La <

!,
ý'
ý0 N 3
N
N
ad

0 00 000
co N
co 04 r-
SUO4 U pao"j

Fig. (195) Effect of inclination of shear reinforcement on deflection

r"D ?ý
-
shear reinforcement, provided of course that in the latter

case all the other parameters are kept constant.

9.2.1 End Rotations

Rotations were measured by means of clinometers positioned

over the supports with a view to observing any influence of

shear cracking and shear reinforcement on flexural stiffness.

It is realized that the study of ultimate rotation could not

be suitably undertaken in simply supported beams under main-

tained loads, and the object of this instrumentation was to

observe 'elastics behaviour prior to hinge formation.

The rotations recorded reflect the obvious influences of

main steel ratio and span which require no further discussion;

the only remaining conclusion that can be drawn from these

readings are

1) Compression steel was effective and reduced rotations.

This accords with the arguments put forth in chapter 7

on the ability of this reinforcement to resist flexural

but not transverse actions see fig. (196) Beams R13y

-

R159 R16.

2) The amount of shear reinforcement did not seem to have

any material effect on rotational stiffness. Considering

that deflections were somewhat influenced by shear rein-

forcement, this indicates that stirrups resist shear de-

formation but are insignificant in comparison with the

main reinforcement in resisting rotation see fig. (196)

-

Beams R121 R28

508
 dO
N
y\ _
X
ýº 0
n"- r. c L
p
n
D'
to C.
C

I!
' ( N N
ý1 \ Ö
(,()
t) \ ý, N
p On ý
\ .

suol U! pooh Io}o}

0 N
9
C9
Q.

1
O

ti..
°o
to ý
\ Ö
11
10

I L O
\ Ö
x
..c . {ii /1 M
%.i LU
Dý.
p UI PDOJ JD; O;
CO NN SUO;
° L111
L
N
.-pNO
N
c ýý
1ý

CD

su0I ui pool 10404 ý.

0
04
\ r
1ý X
N
x

C
` D tf) 0 U-)
*j
II
0
. 6-1 0
04

04 CO
N
0 [ 0 ö
' n
1. S .. 41
. L4')
1,cß N E N ° ! ri
ýQ ý° 0
L m
x

suo} ut pooh 1o1o1 suo} ui pooI lo; 03

I _` C
o
C'') N N

.=y, ý ý
 CHAPTER 10

CONCLUSIONS AND SUGGESTIONS FOR FURTHER RESEARCH

10.1 Conclusions

As had been mentioned in the Introduction - chapter 1- and

when reviewing the principal existing shear theories - chapter

2- one of the basic criticisms of existing methods for calcul-

ating the shear strength of beams with shear reinforcement is

the difference between the assumed physical behaviour of the

beams and the behaviour exhibited during tests. The two most

obvious examples of this discrepancy are the ascdmed cracked

state of the beams and their failure modes.

In this thesis the introduction of a twofold failure mode

provides a rational correspondence between the actual failure

modes and the ones assumed in calculation. The question of

cracking is also rationally treated since 1) A variable crack

length is used in the shear-compression equation and is derived

from compatibility and rotational equilibrium conditions and

minimum strength considerations. 2) On the other hand in the

shearing equation where vertical equilibrium conditions prevail

the critical crack length is based on consideration of aggregate

interlock and the vertical displacements intrinsic to this

action.

Furthermore not only does the use of two equations, reflect-

ing two modes of failure, put forward in this thesis, enable the

ultimate shear strength to be given as the lesser of the two

calculated values, but a comparison of the two values thus ob-

510
tained can provide a good indication of the exact manner in

which failure will take place.

An important implication of the above approach is the fact

that for. the first time a rational explanation is provided for

the influence or lack of influence of the compression rein-

forcement on the shear strength of a beam. This has been a

controversial point among investigators as both an influence

and. aýlack of influence of this parameter have been observed

in tests. The reason for this diversity in behaviour has been

found to, lie in the fact that in the presence of a high degree

of shear reinforcement the governing equation is normally that

of shear-compression involving longitudinal failing compressive

stresses, in the area containing the compression reinforcement.,,

The latter, inevitably takes up part of the compressive stresses

acting on the compressive zone thereby delaying the attainment

of failure stresses in the compressed concrete. On the other

hand when only light shear reinforcement is used, particularly

if the beam is not too, longp it is the shearing equation that

will normally be the governing one implying equilibrium of

vertical stresses which are hardly influenced by the presence

of compression reinforcement at load stages near failure.,

Another implication of the proposed approach is that it

leads to predictions of increased shear strengths of varying

importance in the case of T-beams as opposed to equivalent

rectangular beams. Although this has been observed by previous

investigators it has never been fully explained in more than

empirical terms of. arbitrary application.

In addition to the equations developed for the cases des-

511
cribed above, two other modes of failure have been analysed

and relevant equations proposed. These are the diagonal tension

failure of beams without shear reinforcement and the web

crushing failure of beams with a very high degree of shear

reinforcement.

It must be noted that the field of application of each of

the equations, intrinsic to a particular mode of failure is

not bounded by any arbitrary limits but stems from consider-

ations of minimum strength predictions.

The equations developed and the suggested manner of applic-

ation have resulted in very adequate strength predictions when

compared with test results of normal (a/d1'> 3.0) beams in

which every major parameter likely to affect the shear strength

has been varied. In the case of short beams guidelines have

been proposed for the type of solution envisaged but further

tests are necessary before a final proposal can be made in

that case.

In addition to this general approach to the problem of shear

involving four modes of failure attention has also been focused

in this thesis on some other relevant aspects of the problem

of shear and concrete proposals made to deal with them. These

aspects include the horizontal shear in the flanges of T-beams,

the widths of shear cracks, and the detailing of both the main

and shear reinforcement.

Although concentrated loads and simply supported beams were

used in the tests reported in this thesis, and which form the

backbone of this investigation, the theoretical approach adopted

is quite general and could be extended to cases of uniform loads,

512
axial loads' prestressed members, continuous members, and beams

having shear reinforcement consisting of bent-up bars, when

enough relevant data become available. In this context the

latter means investigation of all the relevant parameters as

has been done in the present series of tests. So far only the

question of bent-up bars has been fully investigated and as

mentioned in chapter 7 the proposed approach gives not only

adequate strength prediction but also reflects the actual be-

haviour of these beams. In all the other cases further research

is still necessary ff a general is

solution warranted.

These and some additional aspects of the shear problem

where further research is still necessary are briefly described

in the following section. A recommendation for research on

short beams has already been made in chapter 7 section 7.8.

10.2 Suggestions for further research

10.2: Y Agqregate Interlock

The need for more basic research on this action has been

mentioned repeatedly throughout this thesis particularly in

connection with diagonal tension failures where the small

widths of the flexural cracks, existing prior to the appearance

of the critical inclined crack and ensuing failure, must allow

important transfer of force, sthrough aggregate interlock. Some

aggregate interlock forces are also assumed to exist in the

upper portions of critical inclined cracks prior to shearing

failures in rectangular beams. To obtain additional information

on this action a basic test is suggested in fig. (197). The

preformed crack can be obtained through splitting of the con-

513
 n
13
----- - ý-----iý
Jr

\ c.

croTs-in

I'.

OflC1 C,O h

Os ýi
ti
c'ýtc! w5 incýi ". cioii

of preform^. d cr cok

rcinforc

NOT--.
1C9
f1l'r1-, `)2r of tcE, ts sliC,,:

Inc! r irirorc : rn' Ci'f

'i

aP
Grid
r

tvisiOf1

Fij. (97) the ý.ýuýýyOF 11e lo-k'

crate (in a manner analogous to the splitting or Brazilian test

of concrete cylinders) along the desired direction. The set-up

suggested can also be used for the study of aggregate interlock

in the presence of axial loads. In cases (b) and (c) of fig.

(197) concentric and excentric conditions of load application

should be studied, and the angle of inclination varied by ex-

tending or reducing the preformed slots. Exploratory tests on

this subject have just been started at Imperial College by far

H. Barn.

10.2.2 Effect of axial loading on shear strength

Influence of axial tension

The detrimental influence of axial tension on shear strength

was the probable cause of the well known failure of the roof

structure of the warehouse at Wilkins Air Base (USA). This

failure gave much impetus to American shear research but that

research has done virtually nothing to shed light on the in-

fluence of axial tension. Some codes of practice do specify

reductions of allowable shear forces when tension is present,

and the percentage reductions are greatest at points of contra-

flexure.

In view of the above some research on this subject has

recently been conducted at Imperial College, using mainly re-

strained beams tensioned along the length from an internal

support to a section of maximum positive moment. The method of

tensioning was by jacking apart between two sets of monolithic

cross-pieces. From such tests on beams without shear reinforce-

ment by F. N. N. Sayani reported in a 1968 University of London

1. Phil thesis the following conclusions can be drawn:

515
 1) Axial tensile loading has a negligible effect on the

diagonal tension strength of beams without shear reinforcement

provided, of course, that the longitudiml reinforcement is

sufficient to prevent a purely tensile failure. Axial tension

can cause either vertical or steeply inclined cracks to form

at loads well below the shear cracking load of a beam free

from axial load, but interlocking of aggregate across such

cracks present premature shear failure. In the case of normal

fii1 (>2.8)
ratios the axial tension causes a slight decrease
Qd1

of shear strength due probably to the general widening of all

cracks.

2) When a beam without shear reinforcement is sufficiently

short, that if free from axial tension it would not fail imme-

diately upon shear cracking, then axial tension reduces its

ultimate shear strength but the resultant strength is never

below the ordinary shear cracking load.

3) In cases whore axial tension reduces shear strength the

loss can be considerably offset by the provision of additional

"compression" reinforcement.

It would seem thus that the influence of axial tension on

the shear strength of members without shear reinforcement is a

very limited one. It is difficult to attribute this phenomenon

to factors other than aggregate interlock. To clarify this

point it would be valuable to have direct test evidence of

interlock behaviour and strength for various crack widths, aggre-

gate sizes and angles of application of load. Such tests are

suggested in section 10.2.1. Their correlation with tests of

actual beams as described above is a sine qua non condition for

516
the understanding of the problem in question.

Turning briefly to the effect of axial tension on the shear

strength of beams with shear reinforcement, the limited number

of tests available show 'that in the longer restrained beams

axial tension causes only a slight reduction of shear strength.

In shorter beams axial tension causes significant reduction in

shear strength which are only partially offset by the provision

of additional longitudinal "compression" reinforcement.

Influence of axial compression

The most recent of the limited tests relevant to this problem

are those presently being conducted at Imperial College by'

H. Bara on beams with a given degree of shear reinforcement.

These tests show that in the longer restrained beams small

amounts of axial compression give a considerable increase of

shear strength and seem to transform the mode of failure from

one of 'shearing' to one of 'shear-compression'. Large axial

s/ further
compression have little/effect since the increase of shear

strength leads to increased bond stresses and bond failure.

The application of axial compression to short beams causes

marked increases of strength, again up to the point at which

bond failure occurs. Bond failure is far less of a problem in

simply supported beams than in restrained ones because of the

beneficial influence of the transverse compression at the support

plattens, an effect that would presumably baJost if the beams

were supported by cross members.

It is not possible on the basis of existing results to make

proposals concerning design equations for members subject to

axial loads but a number of pointers to the development of such

51?
equations exist. The most striking difference in the appearance

of compressed beams failing by shear is the increase of the

depth of the compressive zone as opposed to otherwise similar

beams without compression. Since both the 'shearing' and 'shear-

compression' equations contain terms reflecting modified

elastic theory neutral axis depths it seems likely that the

effect of axial loading can be accounted for by its effect on

the elastic neutral axis depth using the dimensionless ratio

Ndl
(m) at the failure (for beams the
section point-loaded section

of maximum moment). The use of this ratio would give an at

least qualitatively correct account of the results obtained

in the tests, as the effect of the axial load N would be

significant in short beams where G1is low, and much less im-

portant in longer beams.

In view of the above it would seem that research is still

required in the following fields if a rational theory for the

effect of axial loads is to be obtained;

1) Aggregate interlock tests as suggested in 10.2.1 and

application of the results obtained to the tests of

beams subject to axial loading.

2) Further 'axial compression' testing of beams with varying

ratios of main and shear reinforcement.

10.2.3 Effect of manner of load application and support

condition

There seem- to be three broad groups of loading arrangements

as far as the shear failure or reinforced concrete beams is con-

518
cerned. Beams loaded by one or two point loads appear to belong

to one group while beams subjected to more than three point

loads or to a uniformly distributed load form the second group.

The third refers to beams loaded indirectly (through

group

nibs etc. ).

Distributed loads

The experimental part of this research has used load systems

belonging tothe first group. Although it is possible to extend

the theoretical approach to cover distributed loads by examining

failure conditions at every section the exact approach in such

cases must be substantiated by tests in which the actual

behaviour of the beam should be closely observed. In

physical

the process will be one of applying the equation deve-

general

loped at various sections in order to determine the critical

section at which the shear resistance of the benm'as a whole

is at a minimum. To render the suggested tests more meaningful

that such is valid several

and verify an approach as suggested

beams should also be subjected to a combination of concentrated

and distributed -loads.

Beams loaded indirecly!

(37)
Tests by Ferguson and Thompson and R. Taylor(115) in

beams without shear reinforcement and by muto and Kokusho(81)

in beams with shear reinforcement have shown that the shear

strength of short beams can be considerably reduced in the case

of indirect loading and/or support conditions. An explanation

of this phenomenon has been given in chapter 7 section 8 in

.

terms of freedom of the concrete to fail at its minimum strength.

A more precise evaluation of this effect is however warranted

519
and this would require test of beams with varying shear/span

ratioJsince, as expected, the effect of this loading arrange-

ment seems to be reduced at longer spans.

10.2.4 Prestressed Concrete Beams

A primary objective in any research associated with this

e
subject must be the proposal of formulae for prestress concrete

that would also be applicable to reinforced concrete in the

absence of the prestressing force Several attempts have been

.

made in the past to obtain this correspondence but the theories

that do so do not reflect accurately the'behaviour of rein-

forced concrete beams, as has been pointed out in the intro-

duction.

The proposed theory could be extended to prestressed beams but

a series of tests are required to investigate and if necessary

modify some of the assumptions intrinsic to the proposed app-

roach. In some preliminary tests on prestressed T-beams at

Imperial College the proposed theory seems to apply both

physically and numerically once the shear cracking has occurred,,

provided that the calculated strength is well in excess of the

higher shear cracking loads exhibited by prestressed beams.

These higher shear cracking loads depend on the degree of

prestressing and as a result the more or less direct application

of the proposed equation in the fiold of T-beams is greater

in cases of partial prestressing. If failure occurs near

the shear cracking load, then the aggregate interlock effect

must be taken into account and this will require some of the

knowledge gained by the research suggested in (10.2.1). In the

case of rectangular beams a reassessment of the effect of the

520
main reinforcement could be expected es well as some modific-

ations to the aggregate interlock considerations put forward

in this thesis for reinforced concrete members.

10.2.5 Continuous beams and influence of shear on moment-

rotation characteristics

In the case of rectangular sections the proposed theory

which is basically an improvement-on Regan's theory could

be applied to a continuous beam by considering the shear re-

sistance of various sections along the span and arriving at

minimum strength solutions. Regan's original theory has given

good results so the proposed theory should a fortiori also be

applicable. In the case of T-beams very few results exist to

date. Hognestad(48) tested four continuous T-beams under uni-

form loading, three of which failed in shear but the mode of

failure is not very clear and the region of the span covered
was
by shear reinforcement/variable. The main conclusion derived

from these tests was that the ACI/ASCE formula for shear applies

in the case of continuous T-beams under uniform loading. For

a basic shear theory a more systematic approach is warranted,

starting with tests on beams with concentrated loads. The

writer has performed one such test of a three span beam point-

loaded in the middle of the interior span in a previous in-

vestigation(89) and obtained a shearing failure in the vicinity

of the load application area of the middle span where the shear

reinforcement present was lass than that near the supports.

In the case of web crushing the proposed formula should be

satisfactory since it is based not only on tests of simply

supported beams (Plocas, Leonhardt) but also on Robinson's

521
tests of restrained beams.

There exists however another problem associated with inde-

terminate members, stemming from the limit stete methods which

may allow considerable increases of load after the first yield-

ing of the main steel and requiring plastic hinge rotation at

sections subject to high shear forces. The particular points

that should be investigated in connection with this problem are:

1) possibility of reduction of shear resistance due to the

formation and rotation of plastic hinges.

2) possibility of reduction of the rotational capacity of

a plastic hinge due to the presence of a high shear

force

3) In the case of a disadvantageous interaction between

shear and hinge rotations the possibility of its counter-

action by the provision of extra stirrups should bo in-

vestigated. In other words if the effect of shear on the

rotational capacity proven to be significant, an ex-

pression should be derived for calculating the additional

shear reinforcement necessary to achieve given totations,

over and above the shear reinforcement calculated normally

for a member not requiring appreciable plastic redistri-

bution of moments to reach its flexural capacity.

Finally the effect of the curtailment of tho main roonforce-

ment on the shear strength of continuous beams should also be

investigated.

10.2.6 Repeated loading

It is imperative to study the effects of repeated applications

of load on the shear characteristics of a beam because a member

522
is more likely to be subjected to this rather than the static

type of loading. Although some research has been done on the

subject it has not beon systematic and conflicting conclusions

have been drawn from tostß by various investigators.

G. Alatore and J. Casillas of tho Nations Autonomous Uni-

versity of Mexico tested 26 pairs of rectangular beams with

and without stirrups. One beam of a given pair was tested

statically, the other subjected to a 100 load cycles at 80%

to 90% of ultimate load. The authors reported that the cracking

by this load cycling and also concluded that

was not affected

the repetition of a short number of high loadings does not

cause a significant decrease in the strength of beams failing

in shear.

A similar conclusion of no effect is reached by Robinson in

his tests thin I-beams failing by web crushing (BQ

of webbed

see chapter 6 and chapter 7 section 7.7). In these

series -

tests 1000 load cycles were carried out at several load stages

corresponding to 3/6t 4/6p 5/6 of the expected ultimata load.

On the other hand in rectangular beams without stirrups

Chang and Kesler(24) reported that after 1000 cycles the crack'

ing load is reduced to about 70% of the static value while at

the same time the ultimate load is reduced to only about 80%

of the static ultimate. In other words hot only does a reduction

in strength take place but the cracking load is reduced more

rapidly than the ultimate failure load.

A reduction in the shear strength due to repeated loading of

even a very limited number of cycles is reported by Rusch(104)

who states that if beams are tested through 10 load cycles at

523
various load levels up to 70% of the failuro load a systematic

reduction in strength will result.

It can thus be seen that there is a definite need for further

experimental evidence on this question if the conflicting data

listed above are to be rationally assessed. Efforts should

particularly be made to evaluate the effect of repeated loading

on the various failure modes.

Another point of some significance is that the shear crack

widths intrinsic to these tests have not been measured or at

least have not been reported. The tests reported in this thesis,

where shear crack widths were measured, were monotonic tests

to failure. However in one or two tests in which a very small

number of unloading and reloading cycles were used the following

features were noted. In the earlier stages of the reloading

cycle crack widths were greater than those found in the original

loading, but when the maximum load of the original cycle was

approached the reloading cycle crack widths were virtually the

same as those in the first loading. It is not thus expected

that any drastic revisions of equation (9.161) will be necessary

for the case of repeated loading but this of course must be

substantiated by appropriate tests.

In the case of post-tensioned thin webbed I-beams with

(90)
nominal stear reinforcement Price reported that diagonal

tension cracking is possible under repeated loads of lessor

magnitude than the static diagonal tension cracking load. This

seems to be duo to the effect of the repeated loadings on the

crack propagation since he reported that a prior history of

repeated loading on a beam in an uncracked state has no signi-

524
ficant effect on the static diagonal cracking strength. He also

observed that the width of the shear eracks was adversely

affected by repeated loading and that inclined cracks, once

formed did not close completely on unloading.

As far as the ultimate failure is concerned, Price notes

that a prior history of repeated loading has no significant

effect on the static ultimate shear strength of beams in sub-

sequent test to failure, oven if cycling is carried out at

load levels beyond the diagonal cracking stage and causes

fatigue fracture of some of the stirrups. This last conclusion

seems surprising and may be due to the questionable bond pro-

perties of the beams tested.

Price's conclusions are based on tests of two beams reported

to have failed in diagonal tension. It is more likely that

these beams which failed after shear cracking did so in the

shearing manner described in this thesis. In any case the

conclusions drawn from these tests can only be very tentative

and cannot really be used to evaluate the effect of repeated

load on a given mode of failure.

Finally the effect of load cycling on the bond characterist-

ics of the beam's reinforcement should receive very close

attention in any research concerned with repeated loading.

525
 REFERENCES

1. ACHARYA D. N. The influence force

of shear on the ulti-

mate strength of flexural members in reinforced concrete

structures. M. Sc. Thesis, University College, London 1964

2. ACHARYA O. N. & KEMP K. O. Significance of dowel forces

on the shear failure of rectangular reinforced concrete

beams without web reinforcement. A. C. I. Journal, October 1965

3. A. C. I. STANDARDS 1964

4. ACI-ASCE JOINT COMMITTEE 326. Shear and diagonal

tension, 1963

5. ADEPEGBAD. A theory of failure for reinforced concrete

beams without web reinforcement subject to bending and

shear. Ph. D. Thesis, University of Leicester, 1966

6. ADEPEGBA D. & SWAmY N. Deformation characteristics of

reinforced concrete beams under combined bending and shear.

Indian Concrete Journal, October 1967

7. AL-ALUSI A. F. Diagonal tension strength of reinforced

concrete T-beams with varying shear span. A. C. I. Journal

may 1957

8. ANSONM. An investigation into a hypothetical deformation

and failure mechanism for concrete. Magazine of Concrete

Research, June 1964

9. ARROYOJ. Dowel action in reinforced concrete beams. DIC

Dissertation, Imperial College 1969

100 BAKER A. L. L. An analysis deformation

of and failure

characteristics of concrete. Magazine of Concrete Research

November 459

526
11. BAKER A. L. L., YU C. W. & REGAN P. E. Explanatory note on
0
the proposed Unified Code clause on shear in reinforced

concrete beams with special reference to the report of the

Shear Study Group. The Structural Engineer, July 1969

12, BAUMANN T. Versuche zum Studium der Uerdubelungswirkung

der Biegezugbewehrung eines Stahlbetonbalken. Bericht

Nr 77, n1aterialprufungsamt fur das Bauwesen der Technischen

Hochschule Munchen 1968

13. BAUS R. Essai do traction directe sur le beton.

Memoires - Verhandeligen, Universite do Liege

14. BERNANDER K. G. An investigation of the shear strength

of concrete beams without stirrups or diagonal bars rein-

forced with high tensile steel with various rib patterns.

R. I. L. E. M. Symposium Stockholm 1957

15. OJUGGREN U. Failure in combined shear and bending.

Nordisk Betong 2/1962 (in Swedish; translated by P. E. Regan)

16. BORISHANSKI M. S. Calculation of reinforced concrete

elements subjected to transverse forces. Moscow 1964.

(in by A.
Russian; translated Placas and V. Arrangelovic)

17. BORISHANSKI M&S. Calcul a 1'effort tranchant des elements

flechis en beton arme d'apres les normes sovietiques.

CEB Bulletin d'Information No 42, may 1965

18. BRESLER B& PISTER K. S. Failure of plain concrete under

combined stresses. ASCE proceedings, Vol 81, April 1955

19. BRESLER B. & PISTER K. S. Strength of concrete under

combined stress. A. C. I. Journal, September 1958

20. BRESLER B. & SCORDELIS A. C. Shear strength of reinforced

concrete beams. A. C. I. Journal, January 1963

527
21. BRITISH STANDARD CODE OF PRACTICE C. P. 114: 1957. The

structural use of reinforced concrete in buildings (Reset

and reprinted 1965)

22. BROCK G. C. The split cylinder analogy in short beams.

Communication appearing in the Shear Study Group Report

(reference 107)

23. BRUCE R. N. The action of vertical inclined and pro-

stressed stirrups in prestressed concrete beams. P. C. I.

Journal, February 1964

24. CHANG T. S. & KESLER C. E. Static and fatigue strength in

shear of beams with tensile reinforcement. A. C. I. Journal

June 1958

25. CLARK A. P. Diagonal tension in reinforced concrete beams.

A. C. I. Journal, October 1951

26. COfITE EUROPEEN DU BETON. Conclusions techniques de la

12e session pleniere. Lausanne, April 1968

27. COIITE EUROPEEN DU BETON. Recommendations pratiquos

unifiees pour le calcul at ltexecUtion des ouvrages en

beton arme. (en vue d'un reglement international de beton

arme) 1964

28. DE COSSIO R. D. Influence of the size and shape of con-

crete beams and slabs upon their resistance to shear fail-

ure. Ingenieria, Universidad Nacional Autonoma do Mexico,

April 1962 (in spanish)

29. DE COSSIO & LOERA S. Discussion of reference (54). A. C. I.

Journal, December 1966

30. DE COSSIO R. D. & SIESS C. P. Behaviour and strength in

shear of beams and frames without web reinforcement.

528
 A. C. I. Journal, February 1960

31. DE61ORIEUX J. ffl. Poutres an double to an beton arme.

Annales de 1'Institut Technique du Batiment at des Travaux

Publics, No 234, June 1967

32. DIB E. Unpublished test results of research currently in

progress at Imperial College. Proposed Ph. D. Thesis

33. DILGER W. Anfangliche und nachtragliche Durchbiegung

infolge Querkraft bei Stahlbetonbalken im Zustand II.

Beton und Stahlbetonbau, Heft 9,1967

34. ELECTRICITE DE FRANCE (E. D. F. ) Influence do la methode

d'essai dans la determination de is resistance a la traction

du beton. Report No 1156-la of the Laboratoire Regional

d'Electricite de France, Aix-en-Provence; January 1966

35. EVANS R. H. Stress in the steel reinforcement of rein-

forced concrete structures, The Structural Engineer,

September 1935

36'. EVANS R. H. Experiments on stress distribution in rein-

forced concrete beams. The Structural Engineer, March 1936

37. FERGUSON P. & THOMPSONJ. N. Diagonal tension in T-beams

without stirrups. A. C. I. Journal, march 1953

38. FERGUSONP. Some implications of recent diagonal tension

tests. A. C. I. Journal, August 1956

39. GILKEY H. J. Experiments with concrete in torsion. A. S. C. E.

Transactions Vol 100,1935

40. GOODE C. D. & HELMY M. A. The strength of concrete under

combined shear and direct stress, magazine of Concrete

Research, June 1967

41. GURALNICK S. A. High strength deformed steel bars for

529
 concrete reinforcement. A. C. I. Journal, September 1960

42. HARRIS A. J. The failure of concrete under compound

stress. Symposium on the strength of concrete structures;

Session B papor No 5 (C & CA publication) London April 1964

43. HILLERBORG A. Privately circulated (in Swedish)

papers

translated by P. E. Regan and appearing in references (94)

and (107)

44. HOFGECKJ. A., IBRAHIM I. 0., MATTOCKA. Shear transfer

in reinforced concrete. A. C. I. Journal, February 1969

45. HOGNESTAD E., HANSON N. W., & McHENRY D. Concrete stress

distribution in ultimate strength design. A. C. I. Journal

December 1955

46, HOGNESTAD E. What do we know about diagonal tension and

web reinforcement in concrete. University of Illinois

Engineering Experimental Station Bulletin, 1962

47. HOGNESTAD E. High strength bars as concrete reinforcement.

Journal of the P. C. A. Research and Development Laboratories

Vol 3' No 3f September 1961

48, HOGNESTAD E. HANSON N. W. MAGURA O. D. & MASS M

t , 4A,

Shear strength of slander continuous reinforced concrete

T-beams. Journal of the P. C. A. Research and Development

Laboratories, September 1963

49. HOLAPJDI. Design of the compressive zones of concrete

members subject to pure bending in accordance with Nor-

wegian Standards. Nordisk Betong 4/1962 (in Norwegian

translated by P. E. Regan)

50. JAGER K. Grundsatzliche Bemerkungen zu den Ergebnissen

der Stuttgartur Schubversuche. C. E. B. Bulletin d'Information

530
 No 42, July 1964

51. JONES L. L. A theoretical solution for the ultimate

strength of rectangular reinforced concrete beams without

stirrups. C. E. B. Bulletin d'Information No 42, July 1964

52. JONES R. The ultimate strength of reinforced concrete

beams in shear. Magazine of Concrete Researchp August

1956 and CEB Bulletin d'Information No 40, January 1964

53. KANI G. N. J. The riddle of shear failure and its solution

A, C. I. Journal, April 1964

54. KANI G. N. J., Basic facts concerning shear failure.

A. C. I. Journal, June 1966

55. KANI G. N. J. Discussion of "How safe are our large rein-

forced concrete beams by G. N. J. Kani, A. C. I. Journal

March 196711. Author's closure, A. C. I. Journal, September

1967

56. KANT G. N. J. A rational theory for the function of web

reinforcement. A. C. I. Journaiy march 1969

57. KESLER C. E. Statistical relation between cylinder, modi-

fied cube)and beam strength of plain concrete. A. S. T. M.

V54,1954

58. KHAN m. H. Shear strength of concrete beams reinforced

with bent-up bars. B. Phil. Thesis, Imperial College,

London 1969

59, KREFELD W. J. & THURSTON C. W. Contribution of longitud-

final steal to shear resistance of reinforced concrete

beams. A. C. I. Journal, march 1966

60. KREFELD W. J. & THURSTON C. W. Studies of the shear and

diagonal tension strength of simply supported reinforced

531
 concrete beams. A. C. I. Journall April 1966 and Columbia

University in the City of Now York, Department of Civil

Engineering and Engineering Mechanics report; June 1962

61. LAUPA A., SIESS C. & NEWMARKN. Strength in shear of

reinforced concrete beams. University of Illinois Engi-

nearing Experimental Station Bulletin 428; March 1955

62. LARSSOiJ L. E. On the anchorage of high quality corrugated

reinforcing bars in concrete beams. R. I. L. E. mj Symposium

Stockholm 1957

63, LEONHARDT F. La roduction de la couture dteffort

tranchant dans los structures on beton arme - Interpre-

tatlon des resultats dtessais a l'aide dune generalisation

de l'analogio du treillis. C. E. B. Bulletin d'Information

No 49, may 1965

64. LEONHARDT F. & WALTHER R. The Stuttgart shear tests 1961

Cement and Concrete Association Translation Nr 111,

-1964
65. LEONHARDT F. & WALTHER R. Contribution to the treatment

of shear in reinforced concrete. Technical Translation

1172, National Research Council of Canaday 1965

66. LEONHARDT F. & WALTHER R. Uborblick uber neuere Schub-

versuche durchgefuhrt am Otto-Graf Institut Stuttgart.

C. E. B. Bulletin d'Information Na 37, October 1962

67. LEONHARDTF. & WALTHERR. Schubversuche an einfeldrigen

Stahlbetonbalken mit und ohne Schubbowehrung. Deutscher

Ausschuss fur Stahlbeton, Heft 151,1962

68. LEONHARDTF. & WALTHERR. Schubversuche an Plattenbalken

mit unterschiedlicher Schubbewehrung. Deutscher Ausschuss

fur Stahlbeton. Heft 156,1963

532
69, LEONHARDT F. & WALTHER R. Versuche an Plattenbalken mit

hoher Schubbeanspruchung. Deutscher Ausschuss fur Stahl-

beton. 152,1962
-Heft

70. LEONHARDT F. & WALTHER R. Geschweisste Bewehrungsmatten

als Bugelbewehrung - Schubversuche an Plattenbalken und

Verankerungsversuche. Die Bauteknik, October 1965

71. LEONHARDT F., WALTHER R. & DILGER W. Schubversuche an

Durchlauftragern. Deutscher Ausschuss fur Stahlbeton.

Heft 163,1964

72. LIHER1ITE R. Present day ideas of concrete technology

R. I. L. E. M. Bulletin Nr 18, June 1954

73. LICHTENBERG S. A statistical hypothesis for failure in

concrete. Nordisk Betong 4/1959 (in Danish; translated

by P. E. Regan)

74. LYSE I. & JOHANSEN R. The relationship between the cube

and cylinder strength of concrete. R. I. L. E. M. Bulletin

Nr 14, march 1962

75. MATTOCK A. Discussion of reference (80). A. C. I. Journal

December 1957

76,11cHENRY D. & KARNI 0. Strength of concrete under com-

bined tensile and compressive stresses. A. C. I. Journal,

April 1958

77. m00DY K. An investigation of reinforced concrete beams

failing in shear. Ph. D. Thesis University of Illinois

1953 (as quoted in reference 65)

78. MOODY K., VIEST I. m., ELSTNER R. C. & HOGNESTAD E. Shear

strength of reinforced concrete beams. A. C. I. Journal

December 1954, January, February and March 1955

533
79.61ORETTO 0. An investigation of the strength of welded

stirrups in reinforced concrete beams. A. C. I. Journal,

November 1945

80, PA0RR0W0. & VIEST I. M. Shear strength of reinforced

concrete frame members without web reinforcement. A. C. I.

Journal, March 1957

81. G1UTOK. & KOKUSHO S. Contribution a l'etude de la re-

sistance a 1leffort tranchant des poutres en beton arme

Bans diverses conditions de chargement - Essais d'effort

tranchant de poutres en beton arme suivant uns methods

nouvelle de chargement. C. E. B. Bulletin d'Information

No 40p January 1964

82, NATIONAL BUILDING CODE OF CANADA National Research

Council of Canada, 1965

83. OHJA S. K. The shear strength of rectangUlär reinforced

and prestressed concrete beams. Magazine of Concrete

Research, September 1967

84, OHNO K4 ä ARAKAWA Shear tests on reinforced concrete

beams using a special type of loading. R. I. L. E. M. Sympo-

sium Stockholm 1957

85. OZDEN K. An experimental investigation on the shear

strength of reinforced concrete beams (with a preface by

K. W. Johansen). Faculty of Civil Engineering, Technical

University of Istanbul, 1967

86. PAULAY T. & FENWICK R. C. The nature of shear resistance

in reinforced concrete beams. The Institution of Engineers

Australia; Civil Engineering Transactions October 1967.

(also: mechanisms of shear resistance of concrete beams.

534
 A. S. C. E. Journal of the Structural Division, October 1968)

87. PETERSSON T. Shear strength of reinforced concrete

rectangular and T-beams. Statens Räd for Byggnadsforsk-

ning Report 111, Stockholm 1964. (in Swedish; translated

by P. E. Regan)

88, PEREDY J. & VISY Z. The economic design of members

subjected to shearing. Concrete and Constructional Engi-

neering, February 1964

89. PLACAS A. La theorie de la plasticite applique au beton

arme. m. A. Sc Thesis, Ecole Polytechnique, University of

Montreal 1967

90. PRICE K. Fatigue characteristics of bonded post-tensioned

concrete I-beams. Ph. D. Thesis Imperial College, London

1969

91. RAMAKRISHNAN V. Behaviour and ultimate strength of rein-

forced concrete in shear. Publication of the P. S. G. XN

College of Technology, Coimbatore, India, January 1969

92. RAPES S. Comportement des barres relevees dans les

poutres an beton armes Annales de 1'Institut Technique

du Batiment at des Travaux Publics. No 234, June 1967

93. REEVES J. S. The strength of concrete under combined

direct and shear stresses. Cement and Concrete Association

Technical Report TRA/365 November 1962

94, REGANP. E. Combined shear and bending of reinforced

concrete members. Ph. D. Thesis, Imperial College, London

1967

95. REGAN P. E. & PLACAS A. Shear strength of reinforced

concrete beams, CEB Bulletin d'Information No 701

535
 December 1968

96, REGLES B. A. 68. Regles pour le calcul et 11execution des

constructions en beton arme. Institut Technique du

Batiment at des Travaux Publics, May 1968

97. RICHART F. E. An investigation of web stresses in rein-

forced concrete beams. University of Illinois Engineering

Experimental Station Bulletin 166; 1927

98. ROBINSON J. R. Essais a l'effort trenchant de poutres

ame
a mince en beton arme. Annales des Ponts at Chaussees,

march-April 1961

99. ROBINSONJ. R. LRadherence des barres groupees par deux

au contact. Annales de l'Institut Technique du Batiment

et des Travaux Publics. No 108, December 1956

100. ROBINSON J. R. Cours de beton arme, tome I. Ecole

Nationale des Ponts et Chaussees, 1562-63

101. ROBINSON J. R. La resistance a l'effort tranchant des

poutres en beton arme; Une tentative de classement des

faits experimentaux; Esquisse dºun programme d'essgis.

C. E. B. Bulletin dtInformation No 37, October 1962

102, ROBINSON J. R. Premiers resultats dtessais de contures.

C. E. B. Bulletin d'Information No 49, may 1965

103. ROBINSONJ. R. Rapport explicatif, Commission V Effort

Trenchant. 12th C. E. B. plenary session, Lausanne, April 1968

104. RUSCH H. Observations concerning repeated loading of

beams reported in "Compte-rendu de la session pleniere

de Monaco 13 Jan 1961". C. E. B. Bulletin d'Information

No 33p July 1961

105. RUSCH H., HAUGLI F. N. & MAYER H. Schubversuche an Stahl-

536
 beton-Rechteckbalken mit gleichmassig verteilter Belastung.

Deutscher Ausschuss fur Stahlbeton, Heft 145; 1962

106. BUSCH H. & LEONHARDT F. Actions at contraintes tangentes -

Proposition pour la nouvelle redaction des recommendations

FIP/CEB. CEB Commission, V on Shear, July 1969

107. SHEAR STUDY GROUP REPORT The shear strength of reinforced

concrete beams. A report by the Shear Study Group of the

Institution of Structural Engineers. Series No 49t

January 1969

1086 SORETZ S. Compte rendu des reunions communes de la

Commission II "Aciers, Adherence, Ancrages" et de la

Commission V "Effort Trenchant". 7th CEB plenary session

Luxemburg, April 1967. C. E. B. Bulletin dtInformation

No 37, October 1962

1099 SUBBIAH K& SMITH R. B. The influence of shear on the

moment of resistance of reinforced concrete beams. The

Structural Engineers November 1958

110. SUNDERLAND A. Shear tests of reinforced concrete beamsi

Ph. D. Thesis. University of Leeds, 1947

111, TAUB J. L'influence de l'effort tranchant supporte par

l'armature longitudinale dune poutre an beton arme sur la

fissuration et la rupture par cisaillement. C. E. B. Bulletin

d'Information No 40, January 1964

112. TAUB J. & PJEVILLE A. M. Resistance to shear of reinforced

concrete beams. A. C. I. Journal, August, September, October,

November and December 1960

113. TAYLOR H. P. J. Shear stresses in reinforced concrete beams

without shear reinforcement. Cement and Concrete Association

537
 Technical Report TRA/407, February 1968

114. TAYLOR H. P. J. Distribution of shear forces in rein-

forced concrete beams carrying shear. Cement and Concrete

Association publication PP/48 November 1968 of a paper in

an International Conference on shear, torsion and bond in

reinforced and prestresses concrete at P. S. G. College of

Technology, Coimbatore, India

115. TAYLOR R. Some shear tests on reinforced concrete beams

without shear reinforcement. Magazine of Concrete

Research, November 1960

116, TAYLOR R. The design of stirrups ih reinforced concrete

beams. C. E. B. Bulletin d'Information No 41f April 1964

117. TAYLOR R. Some tests of reinforced concrete T-beams

with stirrups. Magazine of Concrete Research, December 1966

118. VAN DEN BERG F. J. Shear strength of reinforced concrete

beams without web reinforcement. A. C. I. Journal, October,

November and December 1962

119. WALTHER R. Calculation of the shear of rein-

strength

forced and prestressed concrete beams by the shear failure

theory. Cement and Concrete Association Translation Nr 110;

1964

1201 WATSTEIN D. & 1ATHEY R. G. Strains in beams having dia-

gonal cracks. A. C. I. Journal, December 1958

121. WARDM. A. The testing of concrete materials by precisely

controlled uni-axial tension. Ph. D. Thesis Imperila

College, London 1964

122. WILBY C. B. The strength of reinforced concrete beams in

shear. Magazine of Concrete Research, August 1951 and

538
 M. D. Thesis, University of Leeds 1949
123. WINTER G., URQUHART C., O'ROURKE E. & NILSON A. Design

of concrete structures. 7th edition, 1964 (McGraw-Hill)

124, ZIA P. Torsional strength of prestressed concrete

members. A. C. I. Journal, April 1961

125, ZSUTTY T. C. Beam shear strength prediction by analysis

of existing data. A. C. I. Journal, November 1968

126. ZWOYERE. M. & SIESS C. P. Ulvimato strength in shear of

simply supported beams without web reinforcement. A. C. I.

Journal, October 1954

--I_.

539
 APPENDIX A

MODIFICATIONS TO EXISTING SHEAR COMPRESSION THEORY

In discussing the results of chapter 4 it has been pointed

out that in the cases where a shear-compression type failure

occurred, Regan's theory would seem to be the most applicable.

not only because it gives fairly satisfactory values with

Qtest
test
varying parameters but also because it describes more adequately

the behaviour of the beams failing in this manner. Strain

readings show that the compression thrust does slope downward

once shear cracking has occurred; the reduced neutral axis

tends to be greater and the length of the shear crack smaller

when increased amounts of shear reinforcement are used. These

and other physical aspects implied in the solution obtained by

that theory have been satisfactorily verified as was discussed

in chapter 4.

Since this type of shear failure is the most common one in

rectangular beams (or continuous T-beams in the regions over

the inner supports) failing in shear, it was considered perti-

nent to investigate the possibility of modifying Regan's theory

in order to improve its failure prediction capabilities.

A detailed study of the method and of the relevant tests

has revealed three areas where modifications can be substanti-

ated; implementation of these modifications leads to a some-

what better agreement with test results.

The areas susceptible to modifications are:

1) The value of the stress block factor-k1k3=0.67 in the

failure criterion

540
 2) The influence of stirrups in the compatibility equation

3) The influence of inclined stirrups

In the following paragraphs these points are examined and

methods for improving the original theory suggested. A statist-

ical comparison of the theoretical and actual test results in-

trinsic to each of the proposed modifications is presented in a

table form at the end of this appendix.

A. 1 Failure criterion

The pertinent data for the rectangular beams without heavy

compression steel having failed in shear compression in a more

or less definite manner, are summarized in the table (A. 1.1)

below.

TABLE A. 1.1 (see also fig. 198)

Beam d1 c y x r± Ocyl Ospli

! ins lins; ins ilbf/inilbf/in Ocyl Ocyl
ins lbf/in lbf/in t

R8 2 112.58 15.5 865 3500 3870 410 0.223; 0.904

13.548.544.523
R9 690 3700 4290 460 0.161$ 0.86
R10 i1,6 118 `6 12 825 4440 4295 387 0.186 1.03
Rll 12.3 21 9 6 595 2920 3800 363 0.1565, 0.77
1
R12 28 18 6 12 930 4120
1r 4920 510 {
0.189 0.84
R13 3.25 17 110 9 870 4200 4680 430 0.166 0.9
,
R15 2.6 12 6 16 885 3890 4330 470 0,204 0.9
.
11,03
R17 4.051 12 6 18 400 1910 1850 220
j10,216
R18 2.1 20 6 8 835 3650 4540 457 0.184 10.8041
1 3
R19 3.45 18 6 12 615 3610 4390 420 [0.14 0,8
R20 1.5 21 9 6 900 4440 6230 558 0.14510,713
R21 2.0 16 i 8 12 1300 7400 16980 566 0,18611.06
R22 2.6 15 9 24 570 4100 4280 410 96
.
R27 5.0 7 8 21 410 12250
198 223 0.207 135
1 .

541
 The apparent depth of the reduced area of the compressive

zone that crushed (ndl) cannot be determined exactly. However

by examining visually the actual area that failed and also

guided by the strains measured in that vicinity an approximate

value of the ultimate compression zone can be obtained.

Considering these average' compressive and shear stresses in

the that failed, values o'and - can

concrete area of average

be calculated, and consequently values of the ratios and

Uý.
7 cyl
ov'-lusted. These are plotted on fig. (199).
Ucyl
Examining fig. (199) it can be noted that the shear stresses

obtained are quite compatible with a compression failure, as

fig. (200) shows that average compressive stress values near

the cylinder strength value can be obtained in the presehce of

shear stresses of the order of 0.15 to 0.25 Ucyl. Furthermore

-
if the failure criterion is compression and is plotted
ö,
against when o assumes the value of Ucyl (see fig. 201) the

for t in the that

values obtained range we are concerned are

in good agreement with the results.

Examining the average compressive atresses at failure it

would appear that they are higher than the 0.67 Ucyl value

normally used (Regan, CEO) and closer to 0.9 Ucyl for normal

concrete. This seems to be possible because of the fact that

the stresses at the crack level do not necessarily have to be

zero as would be the case with a flexural crack since the shear

crack ants into an already compressed zone. Furthermore there

seems to be a tendency for the stronger concretes to have

average compressive stress values somewhat smaller than those

encountered in normal strength concretes. The opposite, to even

3(:2
 0.2

o"

01

or

ucyl

05 10 15
O'

_U
cyl
Fig. (199) Compressive and shear stresses at failure
.
presents tests on beams
ýý
,i =4-:
 nd,

Fing. (198)

0.3 p
W
O
O
O
O"
OO
p Op
"""OO
CQýO" G"
00
0.2 Oe"""OO"
A

N goo oa%""
e oe4 oe .ea
A qo ".. ,
°Or)A
ee oc, 0 12 eo00" ýA
01 Anw" -
Aý
84
Good & Helmy
"o0o Reeves
o Bresler & Pister
A Tsuboi & Suenaga

O. 4 0.6 0.8 1
.00.2
SC
.Ucy
Fig. (200) Compressive and shear stresses at failure
tests on specimens

0.35
0.30
025
0.20
0.15
0.10
0.5

Fig. (201) plotted against z (for compression failures)

'r Ucy
{

ý; r ý.ý
 higher degree seems to be true in the case of low
a slightly
(49)
strength concretes. This has also been observed by Holand

in his criticism of the CEB proposals. The reason for that

situation seems to lie in the favourable stress distribution

due to the greater ductility of low strength concrete and the

unfavourable stress distribution due to the brittleness of

high strength concrete. This observation tends to explain why

the experimental ultimate loads obtained in shear-compression

failures of beams differing only in concrete strength vary less

than would be anticipated by theories using a unique stress

block factor for all concretes. The values of 0.85 Ucy for

high 1.00 Ucyl for low strength concretes

strength concretes and

by the author1s tests seem to conform to the above

suggested

observations and can therefore be adopted for the particular

cases of shear-compression failures in which they are applicable.

When the value of the average compressive stresses to be

used at failure is known a solution is obtained once the area

on which it can be applied is also known. To determine the

latter an expression involving compatibility is required for

the reduced neutral axis depth, before applying equilibrium

conditions and establishing a solution in terms of the ultimate

load.

Regan's compatibility condition is:

-n
n0
n
n- Qc Qdl
o n0 +m
no +m

In order to verify this compatibility condition and ensure

that the new values for the stress block factor k1k3 proposed

do not have any detrimental effect, the two sides of the

545
equation are calculated in table (A. 1.2) using on one hand the

experimental values of n and on the other hand the values for

no andd given by Regan charts modified to take into account

1
the proposed k1k3 values. The results are plotted on fig. (202)

and the correlation seems acceptable in view of the approxi-

mations involved in measuring the experimental value of n.

The two points that seem to be at a certain distance from

the theoretical line belong to beams R15 and R19 which aro

somewhat different from the rest. Beam R15 has four times the

compression reinforcement present as in the other beams and

R19 has heavy inclined shear reinforcement. The question of

compression reinforcement is dealt with in the main text and

that of the inclined stirrups in a subsequent paragraph where

it is shown that it is only when heavy shear reinforcement is

present that the effect of inclined stirrups is significant.

It can be seen that beam R18, having light shear reinforcement,

albeit inclined, conforms to the general pattern.

546
 q)

J
D

.n
0

d
6

(1)
Ö

0.1
O

Fig. (202) Verification of compatibility conditions c. f table A. 1.2

TABLE A. 1.2 Verification of compatibili condition

see fick. 202.

Beam 1000rSt% Qdl 1 Qdl no

no rgfyq c
n cf
dl j no+
u n1 ---- -; P(I P1 'd Qd
1c
cyl 0 1-
.Ucyl n
ý o+ 111'dl

i2,2 '
R8 0.279 0.47 0.40; 0.016 0.298 ý 0.91 0.52

0.252 `"0.4610.7 0.029 i1.3 0.298' 0.73 0.63

R9
R10 0.168 10.39: 0,3 0.014 11.95 0.298! 0.99 0.39
4 12.2 10,298' }
R11 0.380 0.531 0.28 0.016 1: 34 0.40

R12 0.625 0.51 0.461 0.01251 2.7 ! 0.2B 1,20 0.50

1
R13 0.667 0.. L2' 0.53 0.0264 1.6 0.28 1.275 0.49
1
R15 1, 0.710 0.631 0e411 0.0286 1,45 ; 0.28 1.04, 0.60
i
R17 0.585 0.60 0.63; 0.033411.4 0.2961 1.00 0.60

R18 0.238 0.44 0.45- 0,013612,0 0.298 1,155 0.39

110.298ý 0.43
R19 ! 0.247 0.45 0.71,10.0284[1.3 1.05
1
R20 0 173 0.40 0.35! 0,01 2.0 : 0.298 1: 18 0.34
.
R21 0.442 0,5510#360,0177 2.1 10.28 ; 1.13 0.49
4
0.252 0.46.0.521' 0.014511.5 10,22 0.84 0,55
R22
R27 1.554 0.7610.6610.0625.0.9 0.28 1.01 0.75

With this compatibility condition and the equilibrium equat-

ions, solutions in terms of ultimate load can be found using

Regan's charts modified for the proposed k1k3 values. This

modification can be brought about by keeping the charts as they

to the
are but using a fictitious cylinder strength equal
1.00,0.9
actual cylinder strength multiplied by a factor of F.-671 0.67
0'85 (3000
and 0.65 for weak (<'3000 psi), normal to 6000) and high

() 6000 psi) strength concretes respectively.

Application of the above-in the relevant beams increases the

Rcalc
value of from 0.80 to 0.84 and, what is more important,
test
reduces the coefficient of variation from 11.6% to 7.7%

(see table A. 4.1 at the end of Appendix A).

548
A. 2 Influence of stirrups on the compatibilit condition

Regan's basic compatibility condition is

c- ;-cc
tl

6
st 1-n

which is approximated by tho use of reference neutral axis

no, to the following expression

n
no (A. 2.2)
no-no+me .........

Equilibrium conditions are then applied and a solution for

the ultimate load obtained by finding the critical value of

c for which the load passes through a minimum value. The

value of c in the equilibrium equation is related to the amount

of shear reinforcement present and in this manner the amount of

shear reinforcement does influence the ultimate neutral axis

depth, since the compatibility condition contains the term c.

No account is however taken of the influence of the stirrups

in the deformation equation, the increase in deformation in

the concrete being considered counterbalanced by an increase

in the steel deformation in that case.

It is the object of this section to investigate this point.

This can be done in two ways.

a) By comparing the deformation of the concrete block above

the failure crack when the contribution of the stirrups

to the magnitude of the top fibre strains is taken into

account in varying degrees.

b) By investigating directly the influence of the stirrups on

n
the compatibility n
condition
0 no +m

549
A. 2.1 Comparison ofýto fibr© strains

As stated previously an expression for can be derived based

n_, cc
on the condition
lýn st

Tests have shown (reference graphs of main steel strains in

chapter 4) that the force in the main steel does not vary much

in the below the critical shear crack (within 10%). Any

region

increase in the neutral axis depth factor n, brought about by

considerations will depend thus mainly on tho inte-

additional

of the strains above the shear crack at failure. It is

gral

therefore necessary to dorive expressions for calculating the

values of these strains.

The value of the top fibre strain along the crack length

inclination of the force which

depends on the main compressive

is in turn affected by the presence of shear reinforcement. It

is in this manner that the stirrups affect the compatibility

and it is the aim of this section to determine to

condition

what extent they do.

A. 2. l. a Inclination ofmain compressive force

Three cases can be considered:

1 The slope of the compression thrust is unaffected by

case

the of stirrups. In this case the direction of the

presence

thrust is line with a slope whose value is

a straight FQ
cc
This is the procedure followed by Regan as far as the strain

compatibility is concerned.

2 The slope of the thrust line is a straight line again

case

but by the of which determine its

affected presence stirrups

the head of the shear crack. It is equivalent to a

slope at

transformation of equation (A. 2.2) into

550
 n IT0
= (A. 2.3)
no --- Qst .......

An evaluation of the effect of equation (A. 2.3) on the ulti-

mate strength is performed-in section A. 2.1. c but for the com-

parison of strains an expression must be derived for the slope

f'; It should be added here that for the present purpose of

.

comparing top fibre strains, the simple CEB - Regan parabolic

stress strain relations can be used provided they are employed

to evaluate the strains in all throe cases.

Q-Qst (Q-Qst)la (Q-Qst)d1(1- n)
ý. e
tan ý. _Fmr.
1
cc

since M can be written as Qdl . Qd

8
(Q-Q)dl(l Qdl
tan ,)3 st, (1-8
- Q
Qd1
Qdl

3 This is the case that is actually closet to the truth.

case
dx thrust is
The slope of the compression variable, changing

the of shear reinforcement that comes into play

with amount

along its path.

qW = (c-x)rgfyqb

QC =Q- (c-x)rgfyqb

QC
tan = F
cc

Fcc
Yb+xr9f
ygb
dx FCC

yd dx
. 2
_ x)+
(Y4b nd
+ 2F 1
-F cc
cc

551
 The equations presented for cases 1,2, and 3 give the slope

of the compression thrust in the block of concrete above the

failure shear crack, or in other words the point where, accord-

ing to the case examined, the centre of compression lies in

any vertical section.

To perform the comparison for the three cases involved a

solution is first obtained by Regen's method which strictly

applies only to case 1. Using the value of n and crack length

c implied in that solution, top fibre strains are evaluated

along the crack length c in all three cases, using the corre-

sponding slope equation; top fibre strain profiles along the

same crack length are thus obtained. Comparison of the integral

of the strain profiles provides an approximation of the effect

of the stirrups. Before discussing the results of such com-

parisons however and presenting an illustrative example, an

explanation must be given of the manner in which the top

strains are obtained when the distance of the top fibre from

the centre of compression is known.

A. 2. l. b Evaluation of top fibre strain

A parabolic C. E. B. stress strain relationship in the concrete

is assumed.

f= f0 (2e - e2)

where f is the stress corresponding to any strain 'C and

r"
e; U
0
Assuming that sections of concrete above the crack, that were

plane in the unloaded stage remain plane during loading, and

neglecting the tensile zone of the concrete, the direct force

in the concrete at any section is:

552
 0
0
0

Fig. (203) Thrust line trajectories

see section A. 2.1a.

Fig. (204) Effect of stirrups on slope of thrust line

see section A. 2.1a. ccse 3
,

4
®o 0 _r0 0
1
-r-

2 2" 2 2w 2"

Fig. (205) Section; considered in the ilfust^ative excmple

s. e section A. 2.1c.
 Fcc = f0 bn d1 (e - e2/3)

where In' and t0 refer to the section considered

Since Fcc is constant along the length of the shear crack

Fcc/fob 2/3)
- d1 (a -0

The distance from the centre of compression to the top fibre

is
nd1(1 - ©/4)
3(1 - a/37

The moment of the force in the concrete about the main rein-

forcement is
dl (1 e/4) I
- ndl
M Fcc 31-
e/3
f Fcc(1 - o/4)
FccdJ. (! '-(37372'
fobdl 3 '-e 3e
LLL j()
+ c-
Fcc
fobd
m _.... _._ Z 2
Ucylbd1 Ucylbdl

T
where (e) is a function of e; the relationship between

and e is shown on fig. (206).

3n-Ocyl
Substituting
bdc
11
n (e)1 (a)
M3n -2 .........

If one now considers a section where the centre of compress-

ion is at a distance tyt from the top fibre

m= Fcc (d,
- y)

Fcc (d,
~ Y)
m=2
u
ylbdl

(1 -d)......... (b)
-3n
1

Combining equations (a) and (b)

(e) 3........... (c)

21n

554
 (o) M (e) for
From equation can be calculated values of y

and fig. (206) can be used then to determine the corresponding

values of tee.

A. 2.1. c Comparison of Loa fibre strain profiles and impliod

ultimate loads

In any given case the slope of the main thrust line can be

evaluated by the methods described in section A. 2. l. a for each

of the three cases considered. This procedure gives the value

of yj the distance of the centre of compression from the top

fibre, at any vertical section along the crack length. Applic-

ation of the findings of section A. 2. l. b gives the correspond-

ing top fibre strain at that section. When this is done all

along the crack three different top strain profiles are ob-

tained. Theße three different strain profiles are shown in

fig. (207) for the case of an illustrative example with the

following characteristics.

b= 6" 0.525
0

d1= 10" c 10"

from Regan's charts n=0.33

rSt=1.5%

rqfyq = 167 lbf/in2

Ucyl = 4000 lbf/in2

Q 15500 lbf

QW = 10000 lbf

The difference in steel deformation at sections at the head and

start of the crack = 11%

The increase in total steel deformation along crack length

Ix 11% = 5.5%

555
 20

18

1.6

14

1.2

1-0

0.8

0.6

0.4

0.2

0.3 . 0.4 0.5 0-6 07 0.8 0.9

Fig. (206) function ý (e)

Variation of the with the strain ratio e
 Table A. 2.1 gives a summary of the top fibre strain calcul-

ations. These are based on the subdivision of the concrete

block above the shear crack in the manner shouwn. in fig. 205.

As already mentioned the resulting strain profiles are plotted

in fig. 207.

TABLE A, 2.1: Summar of calculation of top fibre strain pro-

.

-'files for differen t thrust line tra edtories

Case 1
I Case 21
Section y

(1) 1.24, 35.0

1,35.0 {
(2) 31.76'0.79 13.1 1.4510.66'0.710X24., 8
0.545
(3) 2L2811.02 ; 0.400 1.65 120.4
114.0
110.5
(4) 2.81 1.26 0.300 1.86 0.85 jO.. 490 ;17.1
"
13.33 I8.7
(5) 1.50 jO. 25O1 2.06 0.94 10.430 i115.0
i 17.5
(6) 13.86 1.75 j 0.210 2.27 1.0310_395 j13.8

ýT--ý
Case 3
Section ye

(1) 4` 35.0
(2) '11.4910.68 0.820 23.6
(3) 1.811O. 82 j0.325 17.9
(4) 12.2011.00 10.490 14.0
(5) t2.67! 1.21 0.385 1100
(6) 13.2111.46 0.3101 8.9

Calculations of the integral of the area under the strain

profiles in each case gives the following results

case 1 145 stirrups not involved in compatibility

,6 cc

case 2 A\cc 205 stirrups affect compatibility but thrust

line remains a straight line

case 3 = 180 stirrups affect compatibility thrust

cc
line having varying slope

557
 0 0.2 0.4 0.6 0.8 1.0

Fig. (207) Concrete strain profiles above shear crack corresponding to

different thrust line trajectories

S 58
Comparison of cases 1 and 3

n cc
1-n
:1 st

or approximately

3/2
(ý) =KQ)
st
ca
n =K
st
The increase of n between cases 1 and 3 will then be

cc
r1_
6c
41cc 190
_ 145x1.05
"()=1.12 _
3I
1 x 1.05
cc
st
1
L 1
st

The strength of a beam failing in shear compression can be

taken approximately proportional to (n)i. This is the power to

which the term ms appears in the proposed simplified equation

(7.4.9) which has been developed as an approximation to the

exact solution.

Thus the resultant increase in strength obtained when

stirrups are taken into account in the compatibility condition

in the manner of case 3 is

(1.12)1 = 1.08 or about an 8% increase in load.

In this case the percentage of load taken by the stirrups

is about two thirds of the total load, or about 66%. The per-

centage increase in load therefore, when the effect of the

stirrups is taken into account in the compatibility condition

in this manner' seems to be approximately equal to the square

root of the ratio expressed as a percentage of the load taken by

559
the stirrups to the total load.

Similar calculations have been carried out for beams with

other characteristics and the results obtained confirm the

above approximation.
Q-Q
this increase but rQ
Adopting using as a sufficient

approximation to the exact value of the average

QW Qoalo
test
value of the relevant rectangular beams listed in section A. 1

becomew now 0.90 and the coefficient of variation 7.5%.

Comparison of cases 1 and 2

The increase of, n in this case is

205
(145x1.05 - 1.23

Increase of strength:

(1.23)' = 1.15 or an increase of 15%.

Such an increase would tend to produce unsafe results as a

safety margin of about 10% is needed in any theory to account

for small dowel and residual interlock actions.

A. 2. l. d Discussion

Comparing the three cases presented and evaluating the

results they produce in terms of ultimate loads compared with

actual test results it appears that the case in which a variable

inclination of the thrust line is used to determine the top

fibre strain profile gives the most satisfactory results. This

is to be expected as the thrust line cannot have the same slope

all along the crack length since it is affected progressively

by the shear reinforcement. The difference however that this

refinement produces in terms of increase of load carrying capa-

city of a member failing in shear compression, is quite small

and will not in any practical case exceed 8%. Regan was thus

560
quite justified in neglecting it but the correction factor that

is necessary in order to enable this effect to be taken into

account is now presented in a simple manner that is readily

applicable.

The final proposed equation for the effect on the ultimate

load of the compatibility role of the stirrups is then

Q_Qcr j
1
Qfinal
-Q1 + 100 x 100)
4

where q is the ultimate shear strength calculated without

taking into account the contribution of the shear rein-

forcement in the compatibility of deformations.

A. Investigation of the affect of taking directly account

22.2.

of the stirrUp s in the compatibility condition A. 2.2

The basic compatibility equation is equation A. 2.2 of

section A. 2
_n
no
n0
n0
.
+
P.
11
If stirrups are taken into account as affecting compatibility

in a manner other than their role in influencing the ultimate

neutral axis depth by affecting the length c of the critical

shear crack, this equation becomes

-n
no ýQ_QS
= (A. 2.3)
"..........

o+ m
n0

ný + (Q brgfyqc)n1
..
rf
taking w= b9ß
cyl
n
no - bwUý
lc Qdl
c
Q

561
 Qd
putting Q=K Ucyl b no dl gives K=
U----
cyl 01
n
no
no bwUcylc Qd1
(1 ) c
n r KUcylnod1 f1 dl
o

n
no (A. 2.4)
na Rd1 .......
(1 wc c
no + ^ Knc . dl) . dl

Equation A. 2.4 provides an expression for the value of the

ultimate neutral axis depth factory and equilibrium conditions

can now be applied. This will be done for the case of vertical

stirrups; the question of inclined stirrups and their influence

is dealt with in section A. 3.

Referring to fig. (142) of section 7.2 and expressing the

condition of rotational equilibrium (equation 7.2.3), relevant

to shear compression failures, the following expression for i

is obtained:

FM la +cb rgfyq "2

cc
c2b rf
k1k3Ucyl bn d12 (1-0.4n) v4
+24

m2 rgfvn
klk3 n (1-0,4n_)
2= +c2 2U -ý
d1 n
Ucylbnodl 0 cyl 0
2
klk3 (1-0.4n) + (d )
2w
010

=1kk ýa
Ucylbnodl M, 13 no
_2
(1-0.47n) (c
+ d1 )-u=K
2n0 j

(A. 2.5)
.......

replacing by the expression appearing in equation (A. 2.4)

0
we obtain
Qdl n°(1-0.4n)
K °21)
Cd W
Lk1k3 Qdl 2°
(1 W° )- °+
n+
0 _. Kn dl rfl . dl (A. 2.6)
.........

562
 For given values of nog shear span to effective depth ratio
Qd
d (or to
and shear reinforcement cylinder strength ratio
r--j1)

wo a relationship can be established between K andd ' and

1
the value of K obtained for progressive values ofä The
.
1
curve thus obtained passes through a minimum value of K which

constitutes the minimum strength solution. In other words

equation (A. 2.5) contains two unknowns and thus a further con-

dition is necessary if a solution is to be found. This second

condition is provided by the fact that what we are interested

in is the minimum value of K- that is the shear force Q-since

Qbn (2)
K=U ' compatible with equation that can produce
cyl o
failure. The additional condition can therefore be expressed

as
C id)

The ultimate shear force that can be withstood will thus be

Quit Kmin dcyl b no di

To carry out a comparison a solution for Q in terms of

Ucylbnodl must also be found for the same given parameters

n
the n_ °Qo This
but using compatibility condition . of
o no+ m

course can be readily done by Rogan's charts, which are based

on that condition.

The determination of Kmin and of the corresponding solution

with which a comparison can be made has been performed for

Qd
1 followed
varying values of nog and w. The exact procedure
rfl

is illustrated by typical examples in A. 2.2. a.

These calculations have shown that the increase in predicted

loads obtained when equation (A. 2.3) instead of equation (A. 2.2)

563
is used as a compatibility condition are of the order of 20%;

this increase does not seem to vary very much with the amount

of stirrups present. Furthermore this increase is of the same

order as the increases obtained when considering case 2 in the

preceding section. This is not surprising as the two are really

two equivalent ways of expressing the same effect. Similarly

to case 2 of section A. 2.1 the results obtained hore tend to be

unsafe since this procedure actually overestimates the contri-

bution of the stirrups. It would therefore seem appropriate to

take account of the influence of stirrups on the compatibility

condition by their effect on the slope of the main thrust in

the manner indicated for case 3 of section A. 2.1 and evaluate

it in the simplified way suggested there - section A. 2.1. c.

A. 2.2. a Illustrative examples

Genoral Equations
Qd j' no (lý-o. 4n )2
K= 1j klk3 +d r (A. 2.6)
fý11 Qd 2n .....
no Kno dl Al dl

Assuming a parabolic distribution of compressive strains at

failure with the same stress block factors as in the preceding

paragraph and in Regan's approach, - in order to render the

comparison meaningful - and taking no(1-0.4n) ti 0.9n the

equation becomes
Qdl 0.6no 2w
K- --- +
m Qdl dl 2no
(1- wc
no+ Md1 Kno dI)

0,6n Qd 2
olc. m
K-
M- cwc pill, d 2n
noQdl 10
dl(1- Kno dl)

564
Exam la 1
Qd1
no = 0.6 -=0.3 w=0.07

Kc0.360.11? (d )2x0.0583
c+0.3
2.0+ d1(1-- d1
K.
1

Mathematical transformation of the equation gives

234
K2(2.0+ Kt0.152(ä )+0.0175(4 )+0.36j + 0.002045(- )
-c)-
1111

=0
A graphical solution is then used to find the minimum value

of K. For every value of Ti, the resulting quadratic equation

in terms of K is solved and tho value of K corresponding to

d
that particular value of is plotted, see fig. 208. The
1
curve obtained passes through a minimum which constitutes the

minimum strength solution that satisfies all failure conditions.

Qutbd
from fig. 208 Kmin =U 0.16
no1
cyl
(94) Q
utbd
From Regan's charts U = 0.132
no1
cyl

increase in ultimate load w 20%

Example 2
Qdl
no =0.6 m=0.3 w=0.02

0.36 (dc20.0167
K _
0.033 c)+0.3
2+c - ý'
dl(1 K dl

mathematical transformation gives

234
)- Kt0.36+0.043(° d1 )+0.005(°-) )=0
K2(2+ d1 d1
J+0.000167(- d1

The solution for minimum K appears on fig. 209 giving

K= d=0.134
bnRU
o cyl 1
ult
From Rogan's charts d=0.109
0 cyl 1

565
018

0.17

0.16

oiý

0.1 02 03 04 0.5 0.6 0.7 0.8 0.9 1.0

F'ig. (208) Minimum strength solution Example 1; w_ 007

566
 0

Fig. (209) Minimum strength solution Example 2; w0.02

 0

i
i

I-
V

d
a

Fig. (210) Minimum strength solution, example 3; w=0.01

X09
increase in ultimate load ='22%.

Example 3
Wdd
1
no = 0.6 = 0.3 w=0.01
2
K=c0.30.0167 + 0.3 äx0.00838
2+ d (1- Kd
11

mathematical transformation gives

23
K(2+-2--) )=0.36 + 0.0217 (d )+0.0025(- )
111

The solution for minimum K appears on fig. (210) giving

K- 0,116
bnQUlt d
o cyl 1

From Regan's charts bnRült d=0.094

0 cyl1

increase in ultimate load = 23%.

A. 3 Influence on the ultimate load of the inclination of

shear reinforcement

It is intended in this section to investigate tho change in

ultimate strength brought about by the use of inclined stirrups

in a beam. This can bo done by establishing a procedure for.

calculating the shear strength of beams having inclined shear

reinforcement of a given ratio rqI based on a spacing measured

perpendicular to the direction of the stirrups M followed by a

comparison of the results obtained in such a case with those

produced when vertical stirrups with the same shear reinforce-

ment ratio are used.

The method for calculating the shear strength of beams having

inclined reinforcement will be developed for tho case of a 450

inclination. The same procedure can be followed for the cases

569
of inclinations other than 450. As it will be soon however, the

difference in ultimate loads due to the stirrup inclination

is quite small, up to a maximum of 15% for practical cases, and

a linear interpollation between 900 and 450 inclination will

give quite acceptable results for any intermediate inclination

(103)
of the shear reinforcement. Actually Robinson suggests

such a linear interpollation for the case of web crushing

the differences are much more important. Because of the

where

magnitude of these differences the equation that has been

developed for in (7.5.5) takes direct account

web crushing

of the exact stirrup inclination. In the present case however

of compression failure such a refinement does not seem

shoar

in view of the small differences in ultimate loads

warranted

involved.

Shear inclined at an angle other than 450 is

reinforcement

used in practice and few test results exist dealing

rarely

with that case. Some of the ones that do exist are more rele-

vant to web crushing and are discussed in the appropriate

Most of the tests relevant to the present discussion

section.

carried out by iloretto(79). In his tests of 44 simply

were

beams with welded stirrups Moretto found that beams

supported

having stirrups inclined at 67°30' exhibited higher shear

than those whose stirrups were inclined at either

strengths

450 or 90°. However a doubt must be cast on his findings because

he also found that stirrups inclined at 45°' 900, and 67°301

yield in that order, which in view of the above could imply

that beams with vertical stirrups may have higher shear strengths

than those inclined 45.° This is to

with stirrups at contrary

570
the findings not only of the series of tests reported hero but

with the findings of other researchers having investigated

also
(91)
this (IYloody(78)9 Wilby(122), Ramakrishnan ). one
question,

reason for Moretto's odd findings may lie in the inconsistency

the he used; for example stirrups inclined at 67°301

of spacing

0.765 of the spacing (horizontal) used in the

wore spaced at

45 0 in terms perpendi-
case of stirrups; viewed of a spacing

cular to the direction of the shear reinforcement it represents

a 17% increase. Another reason may be the variable strength of

in the beams tested. Although the differences in

concrete

duo to that variable are not very important in beams

strength

failing in they may become relatively important when

shear,

differences in strength, such as the ones produced by in-

small

clined stirrups, are involved. This is more true in the rather

used by A1oretto where the relative importance of

short spans

the stirrups decreases in comparison with the other parameters

affecting shear failure.

In
support of the writer's attitude the findings of Taub-
(112)
be quoted. In a review of relevant
and Neville should also

tests, including those of beams with 67°30' stirrups by Moretto

ones by Bach and Graf on

and Evans as well as some earlier

beams with bent up bars inclined at loss than 45°9 they conclude

that the most effective inclination of the web reinforcement is

0 to the the beam. Furthermore the CEB(26)

at 45 axis of accept

linear intorpollation between 450 and 900 in all their recommend-

which are expressed in terms of these two inclinations.

ations

A. 3.1 Ultimate load of beams whose shear reinforcement is in-

45 ° to the
clined at main reinforcement

571
i

In shear-compression failures the governing equation is the

equation of rotational equilibrium - sea section 7.2-. Equation

7.2.3 can be simplified in the present case to

!ýI =tx1t f1
cc- aw

where fl F z
w sq sq

The term Fro x 1a is the normal term expressing the moment

of the compressive force in the concrete about the axis of the

main steel. The term [ must be examined in more detail.

w
Calculation of the torm rn
w

The number of stirrups involved at failure can be taken

c+0.75d
approximately as (see fig. 64a and section 7.3(b)).
s1
c+0.75d1
Vertical force of stirrups Vq = Agfyq sin ^-, (ý---ý----)
.
c+0s75d
Horizontal it is it Hq = Agfyq cos w. (ý---r-----)
s

Taking moments about A

L
mwU9.2+Hi

c+0.75di dl
Agfyq sin (S )(2c +2)
ý^ý
Af
but ßY3.
_
bs sin . -c rf
gyq
br f sin2v.
rn (c + 0.75d1) (c + dl)
= --ý-Y ------,
br f
fli =4 --yq (c + 0.75d1) (c + d1)

Returning now to the original moment equation

mim cc 1aw +m
f.1 k1k3Ucylbnd1 (1-o. 4n) + few

ýtl (1-ý0.4n)_m +W2n

klk3
2;
Ucylbnadl o Ucyln0bd1

Qd br f (c+0.75d )(c+d )
(1»0.4n)_ ya 12 1
k1k3 n +4 _ .ý
U---- bd-_ ml
cyl a1a4 Ucyl nc b dl
,}

572
 Using now the compatibility condition

-- n
no
no Rdl
c
no+ 111 d
1
rf
and replacing by wt the expression becomes
cyl
Rml n (1-0.4n)
1 30 (d
+ 4n + 0"75)(d +l)i
U bn d gd
cyl c1 1co1j
no+ G1 " dl

using the same stress block factors and approximations as

in section A. 2.2, wo obtain

0.6n Qd
Q0'+1U . (E- (A. 3. ß.)
4n0 dl + 0.75)(c dl + 1).,
Ucylbn0 d1 C. m
no QAl +
11

With the condition of minimum load at failure,

additional
ýQ=0,
that is a solution can be obtained in terms of
: 1_
d Qd1
for given values of nog and w. This can be
Uc b
Ucylbna --- ,
1
done in section (A, 2.2).
graphically as explained

It is interesting to note that the critical crack lengths

implied in tho solution of (A. 3.1) are greater than those corre-

sponding to the solution of beams with vertical stirrups having

rf rf
tho same q For axample for values of 0.02,
values.
cyl cyl
o, 04 and 0.08 corresponding c/d1 values in booms with vertical

stirrups are 1.70,1.0 and 0.6 while in the case of beams with

stirrups inclined at 450 these values are 2.29 1.2 and 0.8

respectively. These tendencies are substantiated by test results

see chapter 4 section 4.8.

A. 3. Developmment of an approximat© equation for the inareas©

r2
in load due to the 450 inclination of the stirrups

The solution obtained by the use of equation A. 3.1 can be

573
compared to corresponding solutions obtained in the case of

vertical stirrups and the p3rcentage increase in ultimate load

determined. This
comparison appears on fig. (211) for varying
Qd1
values of w and - (or a/dl) and a typical n0 value of 0.5.

It can be seen from fig. (211) that there is no increase in

load for low ratios of shear reinforcement, and that the in-

crease obtained when higher ratios of shear reinforcement are

used in mainly pröportional to that ratio. This increase is

somewhat affected by the shear span to effective depth ratio,

but it beams (d y 3) the

would seem that in normal main effect

of that variable lies in determining at what ratio of shear

reinforcement an increase commences to take place. To avoid

having too many specialized equations or charts for what in

normally reinforced beams is a rather small increase in strength,

Qd1
the latter is plotted against the product 100w in fig. 212.

It can be seen that a good approximation can be obtained for

the increase in terms of this product by the single logarithmic

curve 24log10(100w -F- x -S).

The shear strength of beams having stirrups inclined at

450 can thus be expressed in terms of the shear strength of

an otherwise similar beam having vortical stirrups, the shear

Rd1
reinforcement ratio w and the ratio m.
0 0 [2040 Qml}
90 90 (167w (A. 3.2)
Q450 = Qult + Qult log10
ult .......

The validity of this formula has bean verified in the case

of different values of not within the practical range, i. e.

ne = 0.4 and no = 0.6.

It must be emphasized however that this formula should be

applied only to beams possessing shear span to effective depth

574
1

11

"01 "02 "03 "04 "05 "06 "07 "08 "09

Fig. (211) Percentage increase in ultimate strength

resulting from the us: of inclined (45°)
instead of vertical stirrups

57
 V

t)
1-

0
I-.

LO
Ö

0
L

Fig. (212) Proposed cpproximate equation for the

increase in strength of beams with inclined (45°) stirrup'

ý;
ratios greater than 3.0. In shorter beams the relative import-

ance of the stirrups decreases in respect to the influence of

a
such other parameters as and ratio of main reinforcement;
1
it follows that the relative increase in total strength duo to

the inclination of the stirrups cannot be as great as the use

of the proposed formula in this region would indicate. In that

case a solution could be obtained by finding the minimum value

of Q that satisfies equation (A. 3.1) in the described in

manner

the previous section. (Actually in the case of beams it

short

may not always be possible to obtain a solution in the usual

manner as the value of c corresponding to a minimum strength

solution may be greater than the shear span. In such a case

(A. 3.1) directly by replacing a

equation can be solved dl
by
dl

- see Also section 7.0 on short beams in general )

-.
d
Qdl
The curve corresponding to a value of 0.4, or in

d=2.5, (211)
other words appearing on fig. helps to illustrate

this discussion on short beams. It can be seen that in the

region of norrnal shear reinforcement ratios, the increase duo

to the inclination of the stirrups is not very different from

that of longer beams having a/dl values of 3.3.

Equation A. 3.2 could therefore be used even in that case

Qdl
provided an upper limit of 0.3 is placed on the value of d

to be used in that equation.

Application of the above in the case of the relevant beams

listed in section A. l gives the final statistical values for

the beams that failed in shear compression

average gcalc = 0.91

test

coefficient of variation = 0.06

577
 It can also be noted again here that the crack lengths

implied in the solution of equation A. 3.1 tend to be longer

than those obtained in the solution of equivalent beams with

vertical stirrups. Relevant tests (see chapter 4 section 4.8)

confirm this tendency as well as the dependance of the increase

in shear strength on the degree of shear reinforcement.

A. 4 Statistical analysis

Table A. 4.1 contains a summary of the statistical analysis

of the predictions of the proposed modifications. The main

statistical data relevant to each of the proposed modifications

are presented separately so that a clearer picture of their

effects may emerge from that table which thus becomes self-

explanatory.

578
 TABLE A. 4.1: Summate of values obtained after modifying
Qcalc
test
the exist, shear c_ompression theory
A ýT
Beam 8CD
0.89 10.95 0.95
R8 '0.89
R9 0.80 0.84 0.93 0,93
.
} R10 0.85 0.85 0.90 10.90
i 0.86 , 0.94 1
Rll 0.88 0.94

R12 0.86 '1.00 1,1.00

10,94
0.69 0.77 I
R13 0.83 0.83
I 10.79 !
R15 0.70 '0.85 0.85

R17 0.72 0.61 0.87 1,0.87

;
R18 0.92 : 0.89 10.95
40.95
R19 10.86 0.74 10.85
10.79
R20 0.79 0.92 0.98 10.98
,
R21 0.62 0490 10.97 0.97
10.81 10&87 0.87
R22 0.80
R27 0.71 t O, 75 0080 ! 0.80
;
average
Qcalc
Qtest 0.91
0.87 `0.84 10.90
}
Standard
I deviation] 0.093 0.0650.068! 0.058
I
Coeffo of
variation 0.116: 0.077 0,0750.054

KEY

A= Regan's original theory

B= Modification of stress block factor (A. l)

C=B and taking into account the stirrups' effect on compati-

bility (A. 2)

D=B, C and influence of inclination of stirrups (A. 3)

579
 APPENDIX B

STRAIN READINGS ON BEArl D-2

Complete range of strain readings recorded during the test

of beam D2 (see chapter 4 section 4.2 and fig. 34).

The strains in this table are given in in x 10-6/in and

designate tension (+) and compression (-). The positions of

the strain gauges are shown in fig. 34.

gauge] Load in tons

no !
of 23456.77.5 8 a. 5 9 ý. 5 X10 0.5
fig 34,

! +70 ; +115, +162 +206 425043131+349 ++898

p +412 +482 +579 +713 +815
'+123 '+288
1 +52 +88 , +156 +190 +229 +250 ý +325 1+365 +414 +454 +493
1+806 ,
2 : +1781+266! +355t+456 +5541+664: +726 1+6621+716
+8811+959 +10564-1124 +12031
1+280 1+931
3 +123; +200 "+355 1+434'+518 +560 +610 +785 '+837
f+1027
4 +1491+239'+335+4251+5041+591'+634; +6861+73711+791 +859, +914
5 i+3 i+5 +6 j+9 ; +14 +17 +18 +12 +13 +14 +11 ++21 IF
1-1
6,0; 00 -1 !-2 -5 -7 +2 -2
te-1 1-8 1I
!
7 -3 -3 -1 +13 +30 1+45 iL
-1 -2 -2 -4 ++7 1+41
! 1+45
8 +5 +11 +15 1+19 +25 +35 +64 +65 1+39 -9 +2
1,
'+36 +1098+1557+1826'L
9 +68 +106! +146, +182'+271 +354 +536 +738
1+10 i+29 1+19
! t+10
10 '+3 +4 +5 +6 +8 +15 +74 +30
11 1,+2 . +1 j-1 -2 -4 1+6
+1 +2 i +l -3 -4
12 1-81 ',-1111-139'-160i 201, -215I. -2331-249i-261; -274; -283'
-54
! !
13 1-83 .:-105'-127 !
-153 -165 -174 -184 -189 1-195 -199
-39 1-60
14 ----^-
15 ; +-74 1-68
-18 -29 "-39 '-49 -60 -72 -76 -76 -67 -68
0 413 ! +81 1+146
16 s-1 10 , +2 +3 +7 +35 +57 +88
17 !^^-
- --
18 1-68 1-103;
-1421-179i-218: -260"-283--304i-329; -351'-375; -391!
1-40 ! 1641-175'-187 196; -208 -2141
19 -61 -83 -103'-127'-151
1-20 : t-48 1-73 1-61
20 -28 -39 : -58 , -71 1-77 -76 -82 ! -80
1 '+12 421
21 1+3 : +4 +6 +8 +9 +10 +19 -1 -40 1-4
-61
22 i+10 +22 " +25 ' +24 +21 +12 '-5 1-41 8
i+14 +18 -25

580
 w~
gouge, Load in tons
no - ---,.. _.
2345677.5 8 005'9 9.5 =10 10.5
of
fig 34,23
YS-17-25
3 8-58
+1 +1+1 -1 -2 -7 -12
1+59
24 +37 +97 {+1461-62
+177 j
-7 -4 -1 +3 +7 +13 +19
i
1 1-83
25 -23 1-55 -66 +--78 ý-85 -86 -78 -68
-34 " "-45
1 26 ! -39 ',
1-58 -79 1751-176
-98 -118 -140'-1511-159'-167'-172
1-1591-200 `-2331-244 1-255
27 '-52 1-77 j X106 -132 -206 -218 :-263
28 i-64 1-134 -1691- 2041-241!
-98 -262; -280'-300'-319i-3404-354;
1-387
29 -69 -107 #-146 -1661-226! -269, -290}-313 337 -351 -406
30 -77 I-118-163 1-250{i_296 -320 {-341 -391 i»432 1-473'
{ '-206 , (-363
1-64 j-240 1-357
31 -97 -133 t-166 -203 1-276 =-314 f
-260 L-295 -335
1-48 -97 ! ", 1-213 j-222 j-"229
32 i-72 -123 -148 -176 -190 -200 -237
33 1,-36 1-65 1-82
-49 -100 -117, -126j-132, -1391-140,, -1281-113,
1-33 11-66 !
34 -26 a-166 1-43 J-51 -71 `-66 ; -25
. -19 -68 -72 -45
z
;.. 299 1-3431-3671-398
35 -79 -120 -208 1-253 -320 1-437 1-468
1-61 4-92 1-127 j-3381+
36 )-160 -193j-229: -248; -263 +-282 -301 11-319
? 4i !
37 -44 -90 -114-138-164-175-187-3.99 -204 i-201 -209)
11-65 1-121
11-62 i-126
38 -30 -45 -79 i-95 -112 -132 -122 -99 f 104
.
. 1-55 .
39 -63 '-68
, '"-67 1-64 i-39
-19 -28 -37 -46 -17 -12
!0 j+713
' 1-4 1+15 1+29
40 -5 -2 +3 . +9 1-49
+49 11+162 +470
!
41 1-9 1-15 R-33 -39 +29
1-6 -21 1-26 1-43 +1 +16
11-16 , r i
42 j-7 -23 ,-25 1{-21 1+14 +1 !'_7
-4 -12 -20 11-26
1+1 1+2 1+19 434 j-17
43 +1 t+3 t+5 +7 +11 11+20 4.5
1+7 1+14 1+23 1+84
44 +30 +41 +54 +67 +90 1+144 +197 +107
1+2 j+46
45 +4 +4 +5 1+8 +14 +21 +49
! I-11 1-9 ±+53
-1 -3 -6 r9 -13 -13 +1 ,+51 +36
,
1iº
67 ; +210 +3291+445j+559, +672; +784{'+8481+919. +985 +1055+1148+12171

581