Sunteți pe pagina 1din 331

Cunningham, Lee Scott (2000) Automatic design of concrete structures using a strut & tie approach. PhD thesis.

http://theses.gla.ac.uk/1726/

Copyright and moral rights for this thesis are retained by the author A copy can be downloaded for personal non-commercial research or study, without prior permission or charge This thesis cannot be reproduced or quoted extensively from without first obtaining permission in writing from the Author The content must not be changed in any way or sold commercially in any format or medium without the formal permission of the Author When referring to this work, full bibliographic details including the author, title, awarding institution and date of the thesis must be given

Glasgow Theses Service http://theses.gla.ac.uk/ theses@gla.ac.uk

Automatic Design of Concrete Structures Using a Strut & Tie Approach

Lee Scott Cunningham

A thesis submitted for the degree of Doctor of Philosophy

Department of Civil Engineering University of Glasgow

June 2000

(D Lee S. Cunningham 2000

This thesis is dedicated to my grandfther, JamesMcKeirnan.

Tro

Ovoya

TOI) OE01),

rav icat irov irov avat, irov 77, Oa eivat, aAVa ical wyrya
In the name of God, who is, and was and is to come, A and S2

Contents
Acknowledgments Summary i ..............................................................................................

ii ................................................................................................................

Notation

iv .................................................................................................................
1 ..........................................................................................................

Chapter 1 Introduction

Chapter2 The Strut & Tie Method


2.1 2.2 2.3 Introduction 3 .................................................................................................... Background 4 ..................................................................................................... Strut-tie Terminology ..................................................................................... .6 2.3.1 Structure's B and D regions 6 ................................................................ 2.3.2 Identification of D-regions .7 ................................................................. 2.3.3 General Principles of Strut and Tie model 11 ......................................... 13 Dimensioning Struts, Ties and Nodes ............................................................ 2.4.1 General Procedure 13 .............................................................................. 2.4.2 Strut and Tie Types 13 ............................................................................ 2.4.3 Node Types 14 ........................................................................................ . 2.4.4 Dimensioning of Ties 19 ;.................................................. ...................... 2.4.4.1 Reinforcement Ties 19 ................................................................ 2.4.4.2 ConcreteTies 19 .......................................................................... 2.4.5 Dimensioning of Struts 20 ....................................................................... 2.4.6 Dimensioning of Nodes 26 ...................................................................... 2.4.6.1 Singular Nodes 28 ....................................................................... 2.4.6.2 SmearedNodes 31 .......................................................................

2.4

Chapter3 The Finite Element Method


3.1 3.2 Introduction 34 .................................................................................................... General Theory 35 .............................................................................................. 3.2.1 Isoparametric Elements 37 ...................................................................... 3.2.2 ShapeFunctions 38 ................................................................................. The Layer
Approach

3.3

3.3.1 Assumptions 41 ...................................................................................... 3.3.2 Displacement Representations 41 ........................................................... 3.3.3 Strain Displacement Relationship 42 ...................................................... 3.3.4 Cartesian ShapeFunction Derivatives 43 ...............................................

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

40

3.3.5 Stress-StrainRelationship 46' .................................................................. 3.3.6 Element Stiffness Matrix and Force Vector 47 .......................................
3.4 3.3.7 Numerical Integration ........................................................................ Non-linear Solution Techniques .................................................................... 3.4.1 Standard Newton Raphson ................................................................. 3.4.2 Modified Newton Raphson ................................................................ 3.4.3 Incremental Procedures ...................................................................... 3.4.4 Convergence Criteria ......................................................................... 49 50 51 51 52 52 .

Chapter 4 Model Visualisation and Direct Design


4.1 4.2 Introduction 56 .................................................................................................... The V isualisation Process 56 .............................................................................. 4.2.1 Theory 56 ................................................................................................ 4.2.2 Examples 57 ............................................................................................ 4.2.3 Application to Slabs 62 ........................................................................... Direct Design of Slabs 62 .................................................................................... . 4.3.1 Assumptions 63 ...................................................................................... 4.3.2 Yield Criteria for Reinforced Concrete Slabs 65 .................................... 4.3.3 Design Equations 70 ............................................................................... 4.3.4 Procedurefor Placing Reinforcement 71 ................................................ 4.3.5 Multiple Load Cases 72 .......................................................................... Inplane Application 73 ........................................................................................ 4.4.1 Yield Criteria 75 ..................................................................................... . 4.4.2 Design Equation Derivation 78 .............................................................. . 4.4.3 Boundary Curves 81 ............................................................................... .

4.3

4.4

Chapter5 Material Behaviour and Numerical Modelling


5.1
5.2

Introduction

Concrete Constituent Behaviour .................................................................... 5.2.1 Uni-axial Compression ...................................................................... 5.2.2 Uni-axial Tension .............................................................................. 5.2.3 Bi-axial Stress ....................................................................................

84 ....................................................................................................
85 85 91 91

5.3

5.2.4 Compression Softening 92 ...................................................................... Cracking of Concrete 92 ........................................................... ....................... ... 5.3.1 Discrete Crack Model 95 ........................................................................ 5.3.2
Smeared Crack Model ........................................................................ 5.3.3 Fracture Mechanics Model ................................................................ Present Concrete Model .................................... ..... ........................................ 5.4.1 Yield Criterion ................................................................................... 5.4.1.1 Compression-Compression Yielding ................................................. 5.4.1.2 Tension-Compression Yielding ......................................................... 5.4-1.3 Tension-Tension Yielding ................................ ... .. ..... ...... ................. Concrete Non-linear Behaviour ...................... ............................................... 5.5.1 Compressive Stress-Strain Relationship ............................................

5.4

5.5

95 96 96 96 99 100 100 102 102

5.6

5.7

5.8

5.5.2 Tensile Behaviour 104 .............................................................................. 5.5.2.1 Single Cracking 104 ...................................................................... 5.5.2.2 Double Cracking 105 .................................................................... 5.5.2.3 Tension Stiffening 106 .................................................................. 5.5.2.4 ShearRetention 107 ...................................................................... Modelling of Reinforcement 110 .......................................................................... 5.6.1 SmearedModel III .................................................................................. 5.6.2 Discrete Model 111 ................................................................................... 5.6.3 EmbeddedModel 112 ............................................................................... 5.6.3.1 EmbeddedElement Geometry 115 ............................................... Applications of Numerical Model 120 ................................................................. 5.7.1 Simply Supported Slab Tested by Hago 120 ............................................ 5.7.2 Comer SupportedSlab Testedby McNeice 127 ....................................... 5.7.3 SlabsTested by Taylor et al 130 ............................................................... 5.7.4 Deep Beams Tested by Khaskheli 134 ..................................................... 5.7.5, Corbel Tested by Niedenhoff 146 ............................................................ 5.7.6 Frame Tested by Stroband & Kolpa 152 .................................................. Conclusions 158 ...................................................................................................

Chapter6 Slab Design


6.1 6.2 6.3 Introduction 159 ................................................................................................... Effect of Mesh Size on Visualisation 160 ............................................................ Model Design and Performance 170 .................................................................... 6.3.1 Slab SMI 170 ........................................................................................... 6.3.2 Slab SM2 180 ........................................................................................... 6.3.3 Slab SM3 190 ........................................................................................... 6.3.4 Slab SM4 200 ........................................................................................... 6.3.5 Slab SM5 210 ........................................................................................... 6.3.6 Slab SM6 220 ........................................................................................... Adapting Load Path for Design 227 ..................................................................... Conclusions 234 ...................................................................................................

6.4 6.5

Chapter 7 Experimental Program


7.1 Introduction

7.2 7.2.1 7.2.2 7.2.3 7.2.4 7.3 7.3.1 7.3.2 7.3.3

Preparation 236 of Models ..................................................... .... . ... ... ................... Formwork 236 ...................................................................................................... Concrete 236 ......................................................................................................... ReinforcingSteel 238 ........................................................................... .. ... .. ...... ... StrainGauges 238 ............................................................ .. .. ................................. Experimental Procedure 239 ................................................................................. DoubleSidedCorbels 240 ................................. .. .. ............................................... SingleSidedCorbel 240 ..................................... .. . ............................................... ComerJoints 240 ..................................................................................................

236 ...................................................................................................

Chapter 8 Strut & Tie Design


8.1 8.2 8.3 Introduction 245 .................................................................................................... Deep Beam BI............................................................................................... 246 Corbel C2A 255 .................................................................................................... 8.3.1 Strut-Tie Model and Design 255 .............................................................. 8.3.2 Numerical & Physical Testing 258 ............................................................ 8.3.3 Comparison of Strut-tie Design with Direct Design 265 .......................... Corbel C3A 270 .................................................................................................... 8.4.1 Strut-Tie Model and Design 270 .............................................................. . 8.4.2 Numerical & Physical Testing 271 ...........................................................

8.4

8.5

CorbelC4A 279 ....................................................................................................


8.5.1 Strut-Tie Model and Design 279 ............................................................... 8.5.2 Numerical & Physical Testing 279 ........................................................... Comer Joint FJ IA.......................................................................................... 287 8.6.1 Strut-Tie Model and Design 287 .............................................................. 8.6.2 Numerical and Physical Testing 291 ........................................................ Comer Joint FJ2B 297 .......................................................................................... 8.7.1 Strut-Tie Model and Design 297 .............................................................. 8.7.2 Numerical and Physical Testing 300 ........................................................ Conclusions 305 ...................................................................................................

8.6

8.7

8.8

Chapter 9 Conclusions
9.1 9.2 9.3 9,4 Summary 307 ........................................................................................................ SlabDesign 307 .................................................................................................... Strut & Tie Design 308 ......................................................................................... Suggestions for FurtherWork 308 ........................................................................ 309 .............................................................................................................

References

Acknowledgments
Firstly I wish to express sincere thanks to my supervisor, Dr P.Bhatt, for his valuable supervision and guidancethroughout the course of this work.

I would like to thank Prof. N.Bicanic, the Head of the Department, for making the facilities available. I am grateful to Dr A. Agar, Dr C.Pearceand Dr D. Phillips for their

usefuldiscussions andadvice.
I am most grateful to the staff of the structural laboratory, especially: Mr R.Boyd, Mr A. Burnett, Mr I. Gardner, Mr R.McCaskie, Miss J.McCulloch, Mr S.McLean, Mr W. Thomson, and Mr A. Yuill, for all their assistancein preparing and testing the models.

K. McColl. Thanks facilities, I Mr For all his assistance to thank with computing wish issues. for his early andusefuladviceon programming arealsodueto Dr A.Bensalem, To my friends and colleagues, Dr Xiaowei Gao, Dr Hashim Musavi, Dr Mohamed
Rouainia and Dr Ben Zhang, I wish to express thanks for our discussions and friendship.

I would like to expressthanks to all my friends and colleaguesin the department, especially the following: Ali Al-Nuaimi, FrancescoBasile, Igor Bojanic, Alan Cuthbertson,GraemeForbes, Domenico Gallipoli, Vassiliki Kochila, Boon-Tiong Lim, Andy Macauley,Andrey Shvidchenko, Xiao-ya Tao. Their friendship will be forevermaintained. Special thanks,are due to my parentsfor their vital encouragement and support
throughout all the years of my studies. Finally, I wish to thank God for giving me strength and lighting my path.

Summary
The major part of the work presentedin this thesis is an investigation of the strut and tie method for designing 2-D in-plane, reinforced concrete structures. Two important issues relating to this method are addressed.Firstly, the issue of visualising an appropriate strut and tie model is dealt with. In many situations it may be difficult to visualise an appropriate model for a given structural system. Here, a convenient method of visualising strut and tie models is presented.Using elastic finite element analysis, low stressedparts of a structure are removed in a step by step process until the main stresspaths, which representthe ties and struts, are defined.

The second important issue to be addressedis that of serviceability of the designed structure becausethe strut and tie model naturally representsa great departure from the elastic stress distribution. Since the strut-tie model is used to design for the ultimate load situation, it is necessaryto assessthe suitability of the same model in relation to serviceability characteristics of the resulting design. It is important that ductility of the structure should be maintained at ultimate loads while avoiding excessive deflections and cracking at service loads. A wide variety of structures were designed, and to assessthe performance of each design, non-linear finite element analysis was used. Verification of some of the numerical results was carried out through physical testing in the laboratory which also allowed the serviceability behaviour of the structuresto be assessed.. The test program comprised of three corbel joints and two frame corner joints.

It was concludedthat design from the strut and tie method can produce adequate both at serviceandultimate loads. In termsof ultimate load prediction, performance the strut-tie methodcan produceresultsof comparable accuracyto non-linearfinite elementanalysis
As an interesting extension to the work here, the same visualisation process is also applied to the direct design of reinforced concrete slabs and the resulting designs are testednumerically using non-linear finite element analysis.

ii

It was found that application of the visualisation processto the direct design of slabs can result in increasesof steel provision, over that which results from the initial elastic pattern. However, in areaswhere the steel can be orientated along the direction of the principal moment paths, a reduction in steel provision can occur. Where a predetermined steel layout is envisaged, the visualisation process can prove useful by directing the load paths accordingly.

iii

Notation
A, A,, A, y B [B]
C19 C2

C [D] [D] [D,, ] E E, Ei Er, fc fc1f fcc


fcd

fc, ft fy (F) IF) G H [K] M


MIP M2

Md

MP M11MY9 My
Mn, Mt, Mnt

M, *, M y Mu n,,, ny, n, N Ni P Pcr Pd PU q rr

Area of steel Area of steel in x-direction Area of steel in y-direction Shearretention factor at cracking strain of concrete Strain matrix Tension stiffening coefficients Strut Elasticity matrix Instantaneouselasticity matrix Rigidity inplane matrix for cracked concrete Young's Modulus Young's Modulus of concrete InstantaneousYoung's Modulus of concrete Young's Modulus of steel Effective compressivestrength of concrete Cylinder compressivestrength of concrete Intermediate yield surface strength of concrete Design compressivestrength of concrete Cube compressivestrength of concrete Tensile strength of concrete Tensile strength of steel Nodal forces vector in Cartesiancoords Nodal forces vector in local coord system, (n,t) Shearmodulus Strain hardening parameterfor steel Stiffness matrix Bending Moment Principal Moments Design Moment Plastic Moment Applied Moments at in point in Cartesiancoords Applied Moments at a point in local coord system (n,t) Design Moments in x and y directions Ultimate Moment Applied inplane forces in Cartesian coords Total number of nodal points Shapefunction associatedwith node I. Applied load First cracking load Design load Ultimate load Intensity of uniformly distributed load Rejection Ratio

iv

I/R I/Ry (R) T TC T, [T] U,V,W uO, vO, WO x,y, z Zi

Curvature Curvature at yield Residual force vector Tie ConcreteTie Steel Tie Transformation matrix Displacementsat a point in x, y, z coords Displacementsat a point in the referenceplane of a plate Cartesiancoord system Distance from the referenceplane to the centre of the ith layer

7XZq YYZ (8) Ecr co EX, Cy, 7Xy ex 0 Ocr OXI OY On, Ot
V

factor Designcompressive strength Shear retentionfactor Materialfactor in Cartesian Shearstraincomponents coords Nodal displacement coords vectorin Cartesian Crackingstrainof concrete Strainat peakstress of concrete in Cartesian Straincomponents coords. Yield strainof steel in a structure from c;maxvM VariationOfa evm Angle of the principalplane Angle of crackwith respect to x-axis Rotationsaboutx andy axesrespectively Rotationsaboutn andt axesrespectively
Poisson's ratio Local, (natural) coord system Steel ratios in x and y directions Stressat a point Stressvector Principal stresses Peak stress Normal Stress Octahedral stress Stresscomponentsin Cartesiancoords. Average von Mises stressin an element Maximum von Mises stressin structure Octahedral shearstress in xy, xz, yz planes respectively The shearstresses Reinforcement bar diameter Transverseshearrotations about xz and yz planes respectively ConvergenceTolerance Degree of transversereinforcement

4,71 pxPY

(; I 02 s

aP
On

a Oct ax, GY9 (FXY


vM CFe vM G max Toct

TXY, TXzf orxz oxgOY 15 (0

ChapterI

Introduction

Chapter 1 Introduction

design The strut and tie methodpresents the to of approach a rational and consistent load in With the this carrying approach, all parts a reinforced concretestructure. *compressive by is the the stress mechanism of structure represented approximating fields as struts,andtensilestress fields asties. The stressin the strutsand ties should the allowablecompressive or yield strengthof the not exceed strengthof the concrete steelrespectively. In the design of structuresby this method there are two important issues to be The first issueis that of the visualisationof an appropriate strut-tie model addressed. for a given structural system.In many structuresthere may be various load paths available and hence no unique strut-tie model exists. The secondissue is that of load in to the and ultimate relation serviceability validity of chosen models ductility is important It that the the of the of resulting structure. characteristics by ensuringthat crushingof concrete prior to yielding structureshouldbe maintained of steel is avoided at design loads. Since the strut and tie method involves a refrom the elastic pattern,it is necessary distribution of the stresses to determinethe extentto which that re-distributioncan be allowed for, while preservingthe required from the structure. Details of strut-tiemodeltheory are given in the next performance chapter.
The purpose of the work carried out here, is to addressthe two issuesdescribed above.

In this work, visualisation of strut-tie models was carried out using elastic finite element analysis in combination with a procedure adapted from optimisation

Chapter I

Introduction

techniques.The visualisation processis the subject of chapter 4. This processwas also applied to the direct design of reinforced concrete slabs. Direct design theory is also discussedin chapter 4. The effect of using the redistributed stressfields which result from visualisation, on the slab performance at service and ultimate loads was assessed using an in-house non-linear finite element program. A full description of the finite element method and the material model used for analysis is given in chapters 3 and 5 respectively. The applicability of the visualisation process to direct design of slabs and the performance achievedis discussedin chapter 6

Using the visualisationprocess,strut-tie models were developedfor a number of The test seriespresented here consisted structures. of deepbeams,corbeljoints and comer joints. Analysis of these structureswas carried out using the same finite elementprogramas described above.In addition, three corbeljoints and two frame comerjoints from the serieswere physically testedin the laboratory.Details of the 7 and8. test-series andresultsaregiven in chapters

In this thesis,all references are listed in alphabeticorder at the end. All figures are to the sectionin which they werefirst referenced. numbered according

Chapter 2

The Strut & Tie Method

Chapter 2 The Strut & Tie Method

2.1

Introduction

In the design of structural concrete, considerable effort has been spent on developing in level difference however design There the of a procedures. exists, safe and efficient i. in design the the e. certain parts are of of structure, accuracy employed each part designed from sound theoretical principles while other parts are designed from rules of thumb and past experience.Examples of this occur in the design of a column with a corbel joint. In this case, the column would be designed from bending theory, while the corbel may be designedfrom empirical formulae. Similarly, in the design of frame comer joints, the adjoining members of the frame are designed using bending theory, and the comer itself is treated empirically. All parts of the structure are of equal importance since their integrated behaviour will control the overall performance of the whole structure. It is therefore necessaryto develop a consistent and unified approach to the design of members.The strut and tie method provides such a consistent method.

Strut andtie modelsarediscreterepresentations fields which result of the actualstress from a given appliedload and supportconditionsin a structure. The modelsrepresent the load carryingmechanism the flow of internal forces of a memberby representing within a structure through struts (representingcompressive stresses)and ties (representing tensilestresses).

Chapter 2

The Strut & Tie Method

2.2

Background

Theprecedent for the strutandtie modelcanbe foundin the earlyinvestigations of Ritter (1899)andMorsch(1909).In their work a trussanalogy was appliedto a beam(fig 2.2). The truss analogyassumes that the crackedreinforced concrete is incapable the beamis madeup of a tensileforcesandhence concrete of carrying by diagonal numberof concrete strutswhich are separated cracks.The strutson interaction form trusswith the with thestirrups andlongitudinal reinforcement a plane followingcomponents:
togetherwith the concrete e top andbottom reinforcement actingas top andbottom chords.
o * stirrups acting as vertical tensile web members. concrete struts acting as diagonal compressionweb members.

Although aware of the possibility of varying angle inclination for the diagonals, Morsch proposedthe use of a 45-degreetruss in order to simplify design. Many designmethodsbasedon this model e.g. BS 8110, nationalcodeshave incorporated Eurocode2, however,only certain parts of the structureare dealt with. The truss analogywas later modified by Leonhardt(1965)by taking into accountthe thickness of the web on the internal distribution of forces.In this work, it was found that the diagonalshearcrackscould be inclined over a rangeof 30'45' depending on the web The applicationof ultimate strengthconsiderations thickness. to the truss model and the formulation of a scientific basisthrough plasticity theory of reinforced concrete was carried out by Thurliman et al. (1975), Muller (1976) and Marti (1985). This but was limited to certain specific cases work furnisheda logical designprocedure suchasa deepbeam.The strut andtie methodhassincebeenusedfor the analysisand designof deepbeamsand is adoptedby national codessuch as the CanadianCode, Rogowsky& MacGregor(186,1988),Cook and seeKong et al. (1977,1978,1990), Mitchell (1988),Tan et al. (1997,1998). Similarly the strut-tie methodhasbeenused for the designand analysisof pile caps,seeAdebaret. al (1990,1996),Siao (1993), Huanget al. (1998).The use of generalised strut tie models,able to be appliedto all In this work the parts of the structurewas proposedby Schlaichet al. (1987,1991).

Chapter 2

The Strut & Tie Method

elastic stress paths of the structure which have been realised through finite element

the strut-tiemodel. analysis, areusedto create


The visualisation of an appropriate strut and tie model is a significant problem. Development of the model is an iterative procedure and there is often no unique model associatedwith a given structure. Alshegir and Ramirez (1992a) developed a meansof constructing the strut and tie models by means of elastic finite elements and an interactive computer graphics package. With this method the user is able to superimposethe strut or tie member over the elastic principal stress plot in order to develop the model. The model when completed, is automatically discretised and analysed as a truss in order to obtain the member forces for design. Examples of the application of this procedure were presentedby Alshegir and Remirez (1992b) for the case of a pre-tensioned deep beam, and by Yun et al. (1994) with a corbel joint. This method is similar to that of Schlaich et al. in that the model is derived from the elastic principal stress pattern. The advantage of this method is that it makes for ease of computation since the model is automatically discretised and analysed and allows the model to be adjusted easily if members are found to be inadequate. However, one disadvantage of the system is that in many casesthe main stress paths are not clear and hence any number of strut-tie models may result. In the systemused in this thesis, the main stress paths are identified automatically, which then helps to identify a unique strut-tie model solution.

The procedure developed by Alshegir and Ramirez (1992a), was further extended by Yun (1997,2000), by carrying out non-linear analysis of the plain concrete structure, and using the resulting principle stress flows to generate the strut and tie model. Hence, in this case,the strut-tie model is derived from a redistributed stressfield. This may in some casesallow for easeof identification of main stresspaths. However, by deviating from the elastic pattern, it must be ensured that the resulting strut-tie model does not exceedthe ductility capacity of the structure.

Chapter 2

TheSftut & Tie Methcd

2.3 " Strut-Tie Terminology

2.3.1 Structure's B and D regions


Due to the presence of concentrated loads and changes in geometry, the stress and (1988) Schlaich distributions et al uniform. always a are not strain within structure depending B D into be two the that or of regions categorised one proposed structure D B distribution The the and regions of a typical on present. respective strain cantilever beam are shown in figure 2.3.1(a)

is linear A 'B-region' is an areaof the structure distribution the andstress where strain 'B' distributionis smoothor undisturbed (a). The be from figure 2.3.1 term ascan seen standsfor Bernoulli since in these regions Bernoulli hypothesisof plane sections remainingplane is assumed valid. In an uncrackedsection,the internal forces and stresses can be calculatedfrom moment,shearand axial forces using well defined formulaefrom bendingtheory.In a crackedsection,the trussmodel canbe appliedto derivethe internalstresses. becomes Wherethe straindistributionin a structure non-linear,the regionis described for discontinuityor disturbance in this case.Such asa 'D-region'. The term 'D' stands loads, bends disturbance distribution of stress a can occur at concentrated comers, and openings.Some typical D-regions of a structureare shown in figure 2.3.1(b). Because the strain distribution is significantly non-linearin theseregions,Bernoulli is no longervalid. The designof D-regionshasin the pastbeenbased hypothesis upon It is proposedto implement the strut and tie rules of thumb and past experience. methodin theseregionsas a meansof rational design. Using the strut and tie model, approach,the first stage of design would be to divide the structure up into its B and D regions.The truss model for the B-regions can be readily corresponding appliedandonly the strut andtie modelsfor the D-regionsneedto be developed

Chapter 2

The Strut & Tie Method

2.3.2 Identification of D-regions


The identification of D-regions within the structure is relatively straightforward when flow At D-region, the the the of the stresses of a structure. examining stress plot becomes disturbed, the direction of stress flow will change rapidly and the stresses become non-uniform. The extent of a D-region boundary can be defined through application of the Saint Venant principle (see ref. 102). This states that the stress distribution at a point far removed from the point of load application will depend forces. As a means distribution force the the of purely on and not upon actual resultant of illustrating this procedure the column shown in figure 2.3.2(a) was analysed using elastic finite elements. The figure details the elastic principal stress plot. It can be clearly seen that the stressesaround the concentratedload are high and have a steep gradient. The stressesgradually spread out moving away from the load point until a uniform state of stress is reached i. e. a B-region. The boundary between the B and D be load, h is lie distance h from to the can assumed at a applied where equal to region the width b of the column. The stressplots at sections 1-4 in the column illustrate the stressvariation in each region. At sections2 and 4, the stressis non-uniform, reaching a peak close to the point of load application. At sections I and 3, which are removed from the point of load application, the stressis uniform.

Chapter 2

The Strut & Tie Method

VT
fig. 2.2 Truss Analogy in Cracked Beam

D-regio

B-region

D-region

compression tension

fig. 2.3. I (a) Elastic Principal Stressesin a Cantilever Beam

Chapter 2

The Strut & Tie Method

h -4

h2

&h

Ea
h

t--h2 hl-

h,

i) Geometrical Discontinuities

-h-

hl

2xh

T-

7" ,;:!
IIIIIII1I11

t
ii) Statical and/or Geometrical Discontinuities

fig. 2.3. I (b) Typical D-regions

Chapici 2

'Flic Strul &, I ic Method

(distance

2 @0.3b -b D-region

along section)

/b

I'll

1 I (a-) .3

CFY 0

Section 1-1

Section 2-2 u

B-region

b) Vertical Stressesat 1&2

-5b 43
-l

O.8b

El compression m tension a) ElasticPrincipalStresses

-2

c) Horizontal Stressesat 3&4

fig. 2.3.2 Identification of D-regions

10

Chapter 2

The Strut & Tie Method

2.3.3 General Principles of Strut and Tie model


Consider the deep beam shown in figure 2.3.3(a). The beam is discretised into a mesh of many elements and the elastic principal stress diagram is shown in fig. 2.3.3(b). One of the elements is shown in (a), is orientated along its principal stress direction which is also the main direction of force flow. Since the tensile stresses at this location are known, the amount and corresponding orientation of reinforcing steel can be provided for. Similarly, the compressive stress at this point can be checked as to whether it is within the compressivestrength limits of the concrete.

The design of a whole structureby this elementby elementmethod would be too detailing. In tedious,time consuming,and could lead to difficulties in reinforcement fields alongthesepathsare replaced by a:system the strut and tie approach, the stress of struts and ties joined at nodes.The internal stresses acting on the strut and tie model can be found from the overall analysisof the structureusing equilibrium betweenthe appliedloads and inner forces.From the data,the struts,ties and nodal regionscanbe designed usingappropriate procedures. The strut andtie methoddictatesthat the structureis designed accordingto the lower boundtheoryof plasticity. In the caseof concrete, only limited plastic deformationis permitted and the strut and tie model have to be chosenin such a way that the deformationcapacityis not exceeded at anypoint beforethe assumed stateof stressis in the rest of the structure. Thus a ductility requirement is imposed,which in reached the caseof highly stressed regionsof the structure,is fulfilled by adaptingthe struts and ties of the model to the direction of the main force flow resulting from elastic analysis.In the caseof a normally or lightly stressed region, the strut/tie directions can deviate from the elastic pattern, by a limited amount, without exceedingthe structure'sductility. As a result of this adaptability,it is possibleto arrangethe ties and hence reinforcementaccording to practical detailing considerations.In this it is assumed that the designed process structurecan adaptitself to the assumed state of internal structural system,without excessdemandon the ductility of the cross section.

11

Chapter 2

The Strut & Tie Method

Typical element subjectedto normal and shearstresses

Element orientated along direction of main force flow

o f I it *" I I II % % 14 % % % % %:

if

or 't

p f

'A

it lk

.4

4r 8

. .

..

tt
compression tension

a) Deep beam

b) Elastic Principal Stresses

Compressive by stress resisted concrete strut

4,
Nodal zone " ) e n or II 0 /i f PWI III I' ll J, J,
Strut

R %V% N
V

Strut 1 %%I % I w k

reinforcement

I11

o4

.4

$1 if

fI

II% f IIt

%%% %%'

it A'4 1

"%

f of

%%

Am

Tensilestress by resisted tie reinforcement


I,

Tie -

P WD
I

c) Typical Element Principal Stresses

d) Strut andTie Model

fig. 2.3.3 General Principles of Strut-tie model

12

Chapter 2

The Strut & Tie Method

2.4

Dimensioning Struts, Ties and Nodes

2.4.1 General Procedure


Figures 2.4.1 a-c show some common stress fields occurring in structures and the from derived fields The elastic tie were resulting stress model. resulting strut and finite element analysis. Each of these strut-tie models can be analysed from statics and the resulting struts, ties and nodes can be dimensioned. The process of

dimensioning involves sizing the individual struts and ties for the forces they carry but also ensuring the load transfer between these members by checking the nodal zones. Because nodal zones concentrate the flow of forces, choice of node detail will affect the strength of the struts bearing on to them and the ties anchored in them. For this is it be to still valid after necessary check whether a chosen strut-tie model reason may detailing.

2.4.2 Strut and Tie Types


The types of strut and ties to be dimensioned can be generally categorised as one of the following:

1. Q: concrete strutsin compression


2. Tc: concrete ties in tension without reinforcement 3. T,: ties in tension with reinforcement

T, can be consideredas one dimensional elementsbetween two nodes.The Cc and Tc are two-, or three- dimensional stressfields which tend to spread in between adjacent nodes. The spreading as illustrated by the bulging of the struts in fig. 2.4.1(b) can result in transverse tensile and compressive stresses.The effect of these transverse tensile stressesmust be accountedfor by adapting the failure criterion of the strut (i. e. reduction of design compressive strength) , or by adapting the strut-tie model (i. e. introduction of additional ties)

Some of the commonly occurring compression stressfields can be defined as one of three configurations:
I

13

Chapter 2

The Strut & Tie Method

1. The fan shaped stressfield: this is an idealised stress field with negligible
curvature and where transversestresses are not developed (fig 2.4.2(a)) 2. The bottle shaped stressfield: in this type the bulging stress trajectories develop Such transversestresseswill initiate cracking and considerabletransversestresses. substantially reduce the compressive strength of the strut. Therefore, such areas need to be reinforced to take account of this.(fig 2.4.2(b)) 3. The prismatic or parallel stressfield: is the limit case of both a--Oand b/a--I, (fig 2.4.2(c)).

2.4.3 Node Types


Nodes are the regions of the model where the strut and ties meet. They are a simplified idealisation of reality. The introduction of a node implies an abrupt change in the direction of forces. In the actual reinforced concrete structure this deviation will take place over a specific length and width. The nodal types which occur in strut-tie be can categorised as either smeared or singular. Examples of each of these models in figures 2.4.1 (a-c). are categories shown node

At a nodal regionwhereoneof the strutsor ties represents a concentrated stressfield, In this case,the resulting the deviationof forceswill tend to be locally concentrated. node is referredto as a singular node. Conversely,where wide compressive stress fields meetwith other compressive fields or tensileties, the deviationof forces stress will be spread or be smeared over a particulararea.Thesetype of nodesarereferredto there are in practice, four as smearednodes.Within thesetwo types of categories, typesof nodeswhich canbe formeddepending on the combinationof adjoining struts (C) andties (T): 1. CCC-node:nodes where the intersectionof three compressivestressfields or struts occurs.Examplesof this type of node are illustrated in figure 2.4.1(b). A
schematic view of the stressfields in thesenodes is given in figure 2.4.3(a).

14

Chapter 2

The Strut & Tie Method

iiiiiiiiiiiiiIIIIIIII

smeared CCT-node

. 7T

compression tension

convergence of stresses at support, forming singular CCC-node


L-r

a)

divergence of' stressesunder point load caused transversetensile stress

sinearedCCT-node

singular CCC-node

b)

IttIIIIIIIIItIIIIIII

smearedCCC-node

sin,gular CCT-node

TZ
C)

fig. 2.4.1 Typical StressFields and Corresponding Strut-Tie Model

15

Chapter 2

The Strut & Tie Method

cr -, fd

fcd

cc

a) Fan shape

b) Bottle shape

c) Pristil shape

fig. 2.4.2 StandardCompressionFields

11)

ichor plote

cl

-1 \ I

fig. 2.4.3(a) Idealised 'hydrostatic' stress in CCC nodes, loadsfrom: transferring concentrated
i) anchor plates, ii) or bearing plate, into bottle shapedstress field

16

Chapter 2

The Strut & Tie Method

1)

ii)

(-N

I
d
onchofoge ,. j length

fig. 2.4.3(b) CCT nodes:


diagonal strut and support reaction balance with reinforcing tie anchored by: i) anchorplatebehindthe node,ii) bondwithin the node,

C
1) ii

fig. 2.4.3(c) CTT nodes: compression strut is supported by i) two bonded form bentup bar reinforcingbars,ii) radialpressure

17

Chapter 2

The Strut & Tie Method

ii)

\
fig. 2.4.3(d) TTT-nodes, compression strut replaced by bonded tie

,e zone

failure zone AA,

fig. 2.4.4.2 (a) Assumption of failure zone for check on the tensile strength of a concrete tension tie Tc.

Is

Chapter 2

The Strut & Tie Method

2. CCT-node: occurring when a tensile stressfield meets two or more compressive 2.4.3(b) in figure illustrated fields. Details and an these nodes are of stress indication of tie force anchorageand bond are given. These types of nodes occur frequently in common structures, such as at the supports of a simply supported deepbeam. 3. C77 node: compressive stress fields meeting two or more tensile stress fields. These type.of nodes will often occur in geometrical discontinuities such as comer joints. A CCT node is formed in a comer joint under a closing moment, where the diagonal strut meetsthe horizontal and vertical tie. 4. TYT node: occurring when three or more tensile stressfields meet. This can occur (fig2.4.3(d)) joint inner junction the moment. under an opening of a corner at

2.4.4 Dimensioning of Ties

2.4.4.1

Reinforcement Ties

Reinforcement is normally provided to carry the tensile forces since the tensile for Thus is deemed be the to reinforcement required negligible. strength of concrete the tie can be calculated from: (A,,) (I /y) (f T. 'a y)

fy is is is force, A, the the T, tie the the steel, of reinforcing area cross sectional where is factor. the y a material of steel and yield stress

2.4.4.2

Concrete Ties

For the case of uncracked tensile stress fields, the concrete tensile strength can be taken account of in the design. By consistently following the flow of forces within the structure to generate the strut-tie model, it may often be found that equilibrium can only be satisfied through the consideration of ties in areaswhere for practical reasons, be be hence tensile utilised. strength must cannot and concrete provided reinforcement Schlaich et al (1987) proposed some guidelines and an empirical formula based on the experimental work of Reinke (1986), for the use of concrete tensile strength. The tensile strength of concrete should only be used to achieve equilibrium in areaswhere no progressive failure is anticipated. Thus, restraint forces and micro-cracks have to

19

Chapter 2

The Strut & Tie Method

be taken into account whether the concrete is loaded or unloaded, cracked or uncracked. Redistribution of stressesin a structure which avoids progressive cracking is assumedto take place under the condition that at any point in a cracked failure zone of area Ak , strength ft (fig 2.4.4.2a). As an initial proposal, the following formula was suggested
2 4d g

the remaining increased tensile stresses do not exceed the tensile

by Schlaich et al. (1987): AA, = and ! A,:t/ 10

where A, t is the area of the tension zone and dg is the diameter of the largest

In practice,it is desirable aggregate. not to rely on the tensile strengthof concretein the designasthe only means of carryingtensileforce.

2.4.5 Dimensioning of Struts The dimensioning thanfor ties sincethe stateof of the strutsis morecomplicated Depending stress present within the strutmember canbe multi-axial. on the existing state of stresswithin thestrut,theattainable compressive strength or effectivestress f, within the concrete the will vary. For the caseof bi-axial compression-tension, increase. For compressive strength of theconcrete will decrease asthetensile stresses thecase the compressive of bi-axialcompression-compression, strength will increase increase. asthestresses
Investigations on the effective strength of concrete struts have been carried out by many researcherssuch as Nielsen et al. (1978), Vecchio and Collins (1982). At a basic level, the effective strength of a concrete strut is defined as some fraction of the concrete cylinder compressivestrength f, ' i. e.

fc= ((X)(fc')
where (x is a factor taking into accountthe effect of the given stressstate.Basedon for the effective stresslevels of concretestruts have test results,empirical equations beenderived.The following empiricalequationwasproposed by Nielsenet. al (1978) for the effectivestress of concrete strutsin beamwebs: (0.7 ')f - Tf CC,,C,

fc =

20

Chapter 2

The Strut & Tie Method

This formula was limited to cases where f, ' < 60 MPa. Ramirez and Breen (1983), proposed a value of 2.82(f, )/4(f, ) as an estimate of the maximum diagonal

(1985) Marti beam beams for type proposed an regions. and compression stress -and Bergmeister ') for level 0.6 (f, types nodes. of struts all of averageeffective stress is levels for (199 1) struts which concrete of effective stress et al. proposed an equation applicable for 20< f, <80 MPa: fe = (0.5 + 1.25/4f,')fc' For comparison, the equations described above are shown for a range of cylinder is in difference large in figure be 2.4.5(a). It that values obtained. can seen, a strengths All the equations described so far, do not take into account the individual stress-state limit for They upper as an empirical simply serve characteristics of a given strut. design strength. Depending on the choice of equation, a wide range of concrete (1987) in by Schlaich be design. In the al. and et work strengths would required MacGregor (1988), strut types were categorised according to geometry and stress A for levels Applicable category. each strut were suggested state. effective stress based is in 2.4.5. These these table on values were categories given summary of determined by (1992), Further Alshegir the effective stress work. work experimental levels of concrete struts from the analysis of experimental results from four deep beams deep beams, loads, to two three pre-stressed subjected point continuous subjected to high shear stressesand four simply supported beams with varying stirrup reinforcement. The resulting formulae are also shown in table 2.4.5.

As a means of quantifying the qualitative descriptions of the strut condition, the tensile stress ratio (al / f. ') corresponding to the compressive strength reduction is also given in table 2.4.5. The ratio was derived from the bi-axial failure envelop of Kupfer et al (1969). In general, the proposals are consistent with one another and are conservative. It can be seen from the table that the smallest reduction in compressive strength is 5%. Using the description of the strut condition corresponding to this reduction, it would be assumedthat the strut is in a state of uni-axial or bi-axial stress. Le al: 5.0.0. From the bi-axial envelop of Kupfer et al. a corresponding tensile stress ratio of 2% is associated with this reduction in strength. For the most seriously

21

Chapter 2

The Strut & Tie Method

is in 75% proposed and this strength compressive cracked strut, a reduction

ratio of 9.5%. corresponds with a tensilestress

Effective stress level 0.85 f,:' 0.68

Concrete St

t Condition

Proposed by Kupfer et.al. (1969) 5% 7.5%

0.5 1 f, ' 0.34 f, '

0.50 0.25 f, ' 0.45 f, ' 0.85 f, '

0.75 f, ' 0.50 fe' 0.95 f, '

Undisturbed and uniaxial state of compressive stress such as in a prismatic strut. Tensile strains and or reinforcement perpendicular to the axis of the strut that may cause cracking parallel to the strut with normal crack width. Tensile strains causing skew cracks and/or reinforcement at skew angles to the strut's axis. For skew cracks with extraordinary crack width. Skew cracks would be expected if modelling of the struts departs significantly from the elastic flow of internal stresses. Isolated compression struts in deep beams or Dregions Severely cracked webs of slender beams with strut angle of 30' Severely cracked webs of slender beams with strut angle of 45' Moderately confined diagonal struts going directly from point load to support with shear span-depth ratio less than 2.0. Struts fom-dng arch mechanism Arch members in pre-stressed beams and fan compression members. Undisturbed and highly stressedcompression struts

Schlaich et al. (1987) Schlaich et al. (1987) Schlaich et al. (1987) Schlaich et al. (1987)

8.5% 9%

MacGregor (1988) MacGregor (1988) MacGregor (1988) Alshegir (1992) Alshegir (1992) Alshegir Alshegir (1992

7o 8.5; 9.5% 8.7% 5%

6.5% 8.5% 2%

Table 2.4.5 Effective Stress Levels in Concrete Struts

Yun and Ramirez (1996) proposeda method of determiningthe effective stressof from finite strutsfrom the principal stress ratios.Theseratios aredetermined concrete the principal stressratios presentwithin eachelement elementanalysisby averaging of a finite elementmesh.Experimentalwork on two-dimensionalconcreteunder biaxial statesof stresswas carried out by Kupfer et al. (1969), and the relationship
between effective stressand principal stressratio derived from this work is detailed in figure 2.4.5(b). Once the principal stressratios have been found, the effective stress is

derivedfrom interpolationof figure 2.4.5(b).The use of figure 2.4.5(b) is limited to


struts inclined up to 10' from the principal compressive stress flows, after which

22

Chapter 2

The Strut & Tie Method

accountmust be made of the inclination. In caseswhere the strut angle deviates outsidethis limit, then the value obtainedfrom figure 2.4.5(a)is multiplied by cos 0, where 0 is the deviation anglebetweenthe strut and the compressive stressflow. Dependingon the level of confinementprovided by reinforcement,anchorageor bearing plates, the effective stresslevel is increased from 5-20%. This method is the advantageous sincethe useof finite elementanalysis, makesit is easyto determine stress-state characteristics of an individual strut. Using the bi-axial yield criterion described,it is then straightforwardto assign the appropriateeffective stressfor design.

For the particular caseof dimensioningthe bottle shapedcompression stressfield, Schlaichet al. (1987) proposedthe diagramshown in figure 2.4.5(b), basedon the forces are experimental work of Reinke (1986).This caseoccurswhen compressive introduced to concrete which is unreinforced in the transversedirection. The spreading of the forces in the transversedirection causes biaxial or triaxial tensionsfurther away compression underthe point of load applicationand transverse from the load. The stress field is characterised by the width of the anchorplate a, the maximum width b availablefor the stressfield in the structure,and the distanceI of becomemore uniform, i.e. the D-region ends. the anchorplate to wherethe stresses The chart showsthe permissibleratio of appliedpressure (pa)to the concretedesign field). The plot strength(fcd)(for an undisturbed compressive uni-axial compression for compression fields without transverse reinforcement(shown as the bold line) is based on elastic analysis with a concrete tensile strength ft fc'/15. = The chart also takesinto accountthe effect of transverse The amount reinforcement. of transverse reinforcement co,is measured as
())--:asfsy/t fcd

where a.,is the area of steel and f. y is the yield stressof the steel.

23

Chapter 2

The Strut & Tie Method

40

35

Nielsen et.al (1978) --D- Ramirez et.al (1983) 6 Marti (1985) --0- Bergmelster (1991) 0

30

.2 is tn
25
U 0

20

is

10 ii 10

15

20

25

30

35

40

45

so

55

60

cylinder strength fc'

fig. 2.4.5(a) Empirical Formulae for Effective Stressfc in Concrete Struts

upper limit .......................................................................................


0.9

0.8 0.7 0.6 j7 . Z:


0.5

0.4 0.3
0.2

0.1 0iiiiii
0

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

lower limit

10

20

30

40 -(ollo2)

so

60

70

80

fig.2.4.5(b)Effective Stress f. in the Strut,Yun & Ramirez(1996)

24

Chapter 2

The Strut & Tie Method

failure compression --blaxiat n bottle neck

0)

but with -crodaad. trunsverse reinfOltMent (daypew) in the bellyregion -uncracked. ptainconaPle

ItnemEnt

b)

a-r

4+P,, 0

fig. 2.4.5(c) Dimensioning of Bottle-shaped Compression Fields, Schlaich et al (1987) a) Chart showing safe bearing pressureP,,,with regard to cracking and crushing of plain unreinforced concrete stressfields, yielding of transversereinforcement and biaxial compression failure in the bottle neck region, b) geometry of stressfield, c) model and reinforcement layout of stressfield with transversereinforcement Co.

Is

a) Original nodal region intersectedby 5 one-dimensional stressfields

b)New nodal stressfield

fig. 2.4.6 Determination of nodal zone shape, Schlaich et A (1990)

25

Chapter 2

The Strut & Tie Method

2.4.6 Dimensioning of Nodes

In the strut and tie method of design, the bearing capacity of the nodal zones are of intersection by formed importance. Since the the of zone size of nodal paramount incoming stressfields can be smaller than that of the existing boundaries of the struts in for is the nodal zone. ties, there cracking and or crushing and a greater potential Thus safe nodal zone design is necessaryfor the safety of the whole structure. The strength of concrete in nodal regions is dependentupon a number of factors relating to the stress conditions present. There are three main conditions occurring which will affect concrete stressin the nodal zone:

struts,anchorage plates 9 level of confinementprovidedby reactions, compression bearingplates,reinforcement from adjoiningmembers for pre-stressing, and hoop
reinforcement

present within the nodalzonesuchaswhenties areanchored 9 straindiscontinuities


in or acrossa compressednodal zone a in bars directly from or anchorage of reinforcing splitting stresses occurring

behinda nodalzone.
As for the concrete struts, various formulae defining the design stresslimits for nodal regions have been proposed. In the case of singular nodes, which are bottlenecks of the stresses,Schlaich et al. (1987) suggestedas a general rule, that the entire D-region would be safe if the pressureunder the most heavily loaded bearing or anchor plate was less than 0.6 fcd- In this case, fcd is the concrete design strength defined as a fraction of the concrete cylinder strength,,y, is a material factor:
fcd

(0.85fc")/(yc)

This assumption was basedon the fact that all significant tensile forceswere carried by the reinforcementand that sufficient development lengthsfor the reinforcement in The 1984Canadian Codelimits the concretecompressive were achieved. stresses ') in nodal zonesboundedby compression nodal zonesto 0.850(f,, struts or bearing ') 0.750(f, in nodal zonesanchoringonly one tensiontie and 0.60(f,') in nodal areas, zonesanchoringtensionties in more than one direction, where 0 is a safety factor.

26

Chapter 2

The Strut & Tie Method

These formulae are based on experimental work and are similar to those described in

the previoussection.
A procedure for evaluating the stressesin CCC nodes with equal or unequal stress fields using Mohr's circle technique was developedby Marti (1985). In this technique, the tie forces are converted to compressive forces acting behind the nodal zone by anchoring the tie using end plates. Marti proposed that the nodal zones could be stressedup to 0.6fc' along with the concrete struts. This idealisation is close to reality since the anchorage of the tensile reinforcement will tend to generate compressive force behind the nodal zone as shown previously in figure 2.4.3(b).

Schlaich et al (1987) and MacGregor (1988) proposed values of effective stresslevels in nodal zones, taking into account the state of stress;a summary of these is presented in table 2.4.6. A general procedure for checking the nodal stressesbased on geometry was proposed by Schlaich and Anagnostou (1990). In this work, the geometry of the by is limited boundary incoming by the the and not only existing of members nodes the areaformed by the intersection of the stressfields reaching the node. In contrast to a real truss, the nodal geometry of an idealised strut-tie model is not limited. The node is surrounded by concrete whose compressive strength may be exploited. The fields consists of several triangular and rectangular areas which are separated stress by lines of stressdiscontinuity. The stress state in each of the fields is either uniform or hydrostatic, as in figure 2.4.3(a). The introduction of transition stressfields allows for zones stressfields of different intensities to be formed (seefig 2.4.6a). nodal

The effect of confinement upon the nodal zone effective stresses was studied by Bergmeister et al. (1991). He proposed effective stress equations for nodes confined by spiral reinforcement, square confined nodes with or without longitudinal reinforcement, unconfined nodes with bearing plates and triaxially confined nodes. A summary of these formulae is presented in table 2.4.6. More recently, Adebar and Zhou (1993), carried out experimental work on the compressive strength of struts confined by plain concrete. Concrete cylinders of varying diameters and heights were loaded over a constant bearing area.The travel time of an ultrasonic pulse was used to

27

Chapter 2

The Strut & Tie Mediod

indicate the level of cracking. It was found that the level of cracking dependedon the height/width by the the ratio of and concrete plain amount of confinement provided the concrete strut. From this work and analytical studies, they proposed some

designing Dbearing The for bearing stress when maximum strength. equations is regions without sufficient reinforcement limited to f 0.6fc'(1+20cp) c:

where
112_ 0.33 ( (A2/A 1) 1) :51 .0

p=0.33(h/b-l):! 1.0 be The ratio h/b is the height/widthor aspect takenas the not and should of strut ratio less than one. The parameter(x accountsfor the level of confinement and the A2 field. P A, for the the and stress compression geometryof parameter accounts limit lower A load the the of surface area respectively. areaand supporting represent for areas 0.6fc'for the bearingstress wherethereis no confinementand wassuggested ' Again, limit 1.8f,, theseproposalsare similar to those was suggested. of an upper described earlier in the sectionand given in table 2.4.6. Similarly, the corresponding The proposedvaluesare tensile stressratios from Kupfer et al. (1969) are presented. The following sectionprovidesa generalised procedure consistentand conservative. for the designof singularandsmeared nodes.

2.4.6.1

Singular nodes

In singular nodes the deviation of forces is often made more abruptly than in smeared nodes. These nodes mainly originate from single loads or support reactions, from concentrated forces introduced by the reinforcement through anchor plates, bond, or radial pressure inside bent bars such as loops. In addition to these, geometrical discontinuities such as re-entrant comers will cause stress concentrations which are representedby a singular node.

28

Chapter 2

The Strut & Tie Method

Effective stress level

Node Type

Proposed by

a, / f" Kupfer et.al. (1969) 5% 7.5%

0.85 fc' 0.68 fe'

0.85 f'

0.65 f, ' 0.5 fc' 0.8 fe, fe':5 27.6 MPa (0.9-0.25f, '/69) fc' 27.6: 5 f, ': 569 MPa 0.65 fr' f,,' '? 69 MPa k fc"(A/Abf'5+

Compression-compressioncompression Nodeswherereinforcementis anchoredin or crossingthe node Nodesboundedby compressive strutsand bearingareas Nodesanchoringone tension tie Nodesanchoringtensionties in more than one direction Unconfined nodeswithout bearingplates Unconfined nodeswithout bearingplates Unconfinednodeswithout bearingplates Confined nodes

Schlaichet al. (1 87) Schlaichet al. (1987) MacGregor (1988) MacGregor (1988) MacGregor (1988) Bergmeisteret al. (1991) Bergmeisteret al. (1991) Bergmeisteret al. (1991) Bergmeisteret al.
(1991)

5%

7.5% 8.5% 6%

7.5% -

a(Ac,om/AbAdl _S/d)2 )0*5 Unconfined nodeswith k fc"(A/Ab bearingpl es Triaxially confined nodes 2.5 f, ' Note:

Bergmeisteret al. (1991) Bergmeisteret al. (1991)

A= areaof confinedconcrete, Ab= areaof bearingplate,Ac.,,= areaof confinedstrut, "< flat= lateralpressure HyA(ds) for f, 48.3MPa ; 2fyAJ(ds)for fc' > 48.3, = d= diameter s= pitch or spacing of confinement reinforcement, of confinedcore, (4.0 for spiralconfinement, 2.0 for square cc= parameter closedhoopconfinement anchoredwith longitudinalreinforcement, 1.0for square closedhoopconfinement without longitudinal anchorage) reinforcement k=0.5 + 1.25/4f,
Table 2.4.6 Effective stress levels in nodal zones

In general,equilibrium in singularnodesis achievedby the balanceof forces in the


interior of the node through direct concrete compressive stresses. In the ideal situation, tie anchorage is, provided by an anchorage plate which transfers the load

from behind the nodethus causingcompression b(i)). Bond is in the node (fig. 2.4.3.
essentially load transfer via concrete compressive stresseswhich are supported by the

29

Chapter 2

The Strut & Tie Method

ribs of the steel bar (fig 2.4.3.b(ii)/(iii))

and by radial pressure in bent bars (fig.

involves 2.4.3. b(iv)). The dimensioning threemain steps: of a singularnode


Adapting the geometry of the node with the applied forces: In the case of CCCbe be to borderline the perpendicular to the the can assumed. node of nodes interior to field the the the node of the within of stress and state stress resultant of be plane hydrostatic. (fig 2.4.3.a(i)). In this case the resulting geometrical relation length dimension C3 be CI: C2: the to of the support or the width can used al: a2: a3-. 2 of an anchor plate.

is This limit: Checking the the condition associated concretestresses are within e automatically satisfied for the entire nodal region if the stressesalong the borderlines of the node do not exceedthose limits and if the reinforcement bonded 2.4.6.1) (fig CCT if In the with case of nodes anchorage sufficient. in it is the a2 to the a, concretestresses and sufficient check reinforcement, from it is In the geometryof the clear cases struts. most adjacentcompression is the controlling pressureand the two the struts pressures out node which of henceit is necessary to analyse only one.Figure2.4.6.1showsa numberof multiby limits proposed layeredandsingly reinforcedCCT nodesandthe dimensioning Schlaichet al (1987).
e Provision of adequateanchoragefor ties in the nodal zone: For anchor plates, this involves a check on the bending strength of the anchor plate and the welded connection with the tie. In this case,a tie having a smooth surface where it crosses the node is better than good bond quality becausestrain compatibility within the bar will tend to crack the concrete within the node. In the case of directly anchoredreinforcing bars, hoop or loop anchoragesare preferable. For straight bar is located design the that anchorage within and anchorages, engineer must ensure behind the node as shown in fig 2.4.3.b(i-ii). Anchorage begins where the transverse compression stress trajectories of the struts meet the bar and are deviated; in order to catch the outermost fibres of the deviated compression field , the bar must extend through to the opposite end of the nodal region.

30

Chapter 2

The Strut& Tie Method

2.4.6.2

Smeared nodes

It is non-nal for D-regions to contain both singular and smeared nodes. In most cases the singular nodal region is most critical and a check on concrete stresses within the smeared node is unnecessary since the applied stress levels are less than in the singular node. In addition, the geometry of the smeared node may be of a similar

magnitude to that of the singular node. This gives rise to the rule of thumb proposed by Schlaich et al; that the structure is safe, if the stress under the most heavily loaded bearing plate is less than 0.6fcd-

31

Chapter 2

The Strut & Tie Method

0.21D

a,
lb, net

2 [I cy, cF) +(a3/al)tan(p] cos (P


lb, net !

i) multi-layered tie

cd

a3-O

(7, (C OS2(p)

ii) single-layered tie

iii) as ii), with incoming compression field

fig. 2.4.6. I (a) Stressesin Typical CCT Nodes, Schlaich et a]. (1987)

32

('11LIptel 2

The strut &. Tic Method

4 -------------------------------T2

ic

C( --

--

--------

r--ol. -

i) TCT Node

T
T2 c

T, "' i) TCT Node

lb

C3

C4

a3 4

(Tc3

CU

a4

C-3

C4
cl C5

Co a,
Gc2 C2 tttttt

ao 4
C5 tttttt CFc5

a5

C,

tit) CCCCC Node

C,

CvC I

2.4.6.1 (b) Stresses in Typical Nodal Zones

33

Chapter 3

The Finite Element Method

Chapter 3 The Finite Element Method

3.1

Introduction

Sincethe work presented in this thesisinvolvesfinite elementanalysis,it is necessary to introducesomeof the relevanttheoryand a descriptionof the programusedhere. The finite elementmethodprovidesgood approximate solutionsto problemswherea closed form or exact solution is impractical.In a structuralcontinuum the actual numberof degrees of freedomis infinite. An approximate solution can be found by dividing the continuuminto a seriesof elementswith a finite numberof degrees of freedom,this processis known as discretisation.The resulting array of elementsis referredto as thefinite element mesh.In essence an approximate solution is achieved by assumingthat the behaviour of the continuum can be represented by a finite numberof unknowns.
The method has applications in many different fields, with each application coming under one of three headings: Problems: here does the system not vary with time e.g. stressanalysis Equilibrium of linear elastic systems,electrostatics,steady statethermal conduction etc. o Eigenvalue Problems: here critical values of certain parametersmust be obtained
V-

frequencycalculations e.g. stability of structures, vibrationproblems, etc.


e Propagation Problems: involves time dependent behaviour e.g. hydrodynamics and dynamic transient analysis of elastic continua.

34

Chapter 3

The Finite Element Method

The finite

element method will

inclusion the of non-linear accommodate

Finite invariably the to the element of problem. adds complexity characteristics which canbe carriedout usingthreebasicapproaches: stress analysis " initial the The displacementmethod: here the displacements as are chosen field displacement from determined the the calculated are unknownsand stresses initial The equilibrium method: here the stresses the unknowns and the are
displacementsare calculated from the resulting stresses. displacements hybrid in is The the the and the third which or mixed method o stresses are employed simultaneously as variables.

"

The displacement method is the most commonly used due to its ease of
implementation in programs. This method was implemented in the program used in

this work.

3.2

General Theory

In structural applications, the governing equilibrium equations are obtained by minimising the total potential energy 7cof the system.
L[O]TC

7C=1

L[S]Tp. dVdV _

fS[B]T

q. dS

(3.1)

where: a vector stress

e strainvector 8 displacements at any point p= body of forces per unit volume tractions q= appliedsurface V= volumeof the structure, S= loadedsurfacearea
The above equation is known as the functional. On the right hand side of the functional the first, second and third terms represent respectively; internal strain energy, work contributions of body forces and work contributions from surface loads

35

Chapter 3

The Finite Mement Method

In the finite element displacement method, the displacement is assumed to have is that the within any element variation unknown values only at nodal points, so described in terms of nodal values by means of interpolation functions or shape functions i. e. 8=N 6' (3.2)

W N functions is the and is the vector of nodal displacementsof the where set of shape in The terms of the nodal the expressed element. strains within element are displacementsvia the strain-displacementrelationship: F,= BY (3.3)

functions. B is is derivatives the the shape where strainmatrix which composed of of Likewise the stresses arerelatedto the strainsvia the elasticitymatrix D: a=De (3.4)

The total potential energy of the continuum is the sum of the energy contributions from each individual element (provided that the chosen shape functions are so as to causeno singularities in the integrandsof the functional). i. e.
7r = Dre e

(3.5)

the total potential energy of element e which on use of the where ire represents functionalcanbe written
(2 JV 13ejT [B]TDBSe. 7re =I, .

dV-j ve

[3e]T [N]Tp.

dV_

Js. WIT

[NIT q. dS)

(3.6)

where V, is the element volume and , Se the loaded element surface area.

W for Differentiation of the aboveequation with respectto the nodal displacements the elementresultsin :
3ne a8e =

(jv

([B]T DB)8e. dV -

Jv

[N]T P. dV -js. [N]T q. dS) =K e5e - F'

(3.7)

where Js. Fe =j [N]Tp.dV + [N] Tq.dS V.


(3.8)

36

Chapter 3

The Finite Element Method

for forces the element, and the are equivalent nodal

KC=L

[B]T

DBAV

(3.9)

0
is termed the element stiffness matrix. The summation of the terms in equations 3.8 in 3.9 to the and over all elements when equated zero results a system of equilibrium by be for These a standard the solved can equations complete continuum. equations technique such as Gaussian elimination in order to obtain the nodal displacements. From this the element stresses can be obtained using the aforementionedrelationships.

The following is a summaryof the basicstepsinvolved in the solution of equilibrium


problems by the finite element method:

i. finite discretisation. Creation of a e. element mesh e *


*

Evaluationof the elementstiffnessandload vector.


Assembly of the element stiffness and load vector in to an overall stiffinessmatrLx

andload vector. equationsfor the unknown nodal 9 Solution of the resulting linear simultaneous variables. 9 Evaluationof the elementstresses.

3.2.1 Isoparametric Elements


For the purpose of this study, eight node isoparametric elements were used in all the finite element analysis. An isoparametric element may be defined as an element whereby the same interpolation function is used to describe the displacement variation within the element as well as the element geometry.

The element coordinates and the displacementsare defined by functions expressedin terms of the natural coordinates of the element. The natural coordinate system is a local system which is defined by the element geometry and is independent of the element orientation in the global system. This system is normally arranged such that

37

Chapter 3

The Finite Element MethcA

the natural coords have a unit magnitude at the element comers i. e. 1 (see fig 3.2a). The main advantagesof isoparametricelementsare as follows:

e Improved accuracyover simple elements e hTiproved computational efficiency by simultaneous definition of element

definition displacement geometry and


* Can facilitate the use of curved elementswhen modelling curved boundaries.

3.2.2 Shape functions Theinterpolation function Ni hasthefundamental of havinga value or shape property functionsdefinethe equalto unity at nodei and zero at all othernodes.Shape in variationof a givenvariablee.g. length,displacement etc.,throughthe element As a resultof this, termsof the values of that variable at the nodes of the element. in functions to the shape are related numberof nodes an elementand hencethe dueto functions in theform of polynomials type. Shape areusuallychosen element in particularwhen it comesto their relativeeaseof mathematical manipulation, integration anddifferentiation.
The degreeof polynomial chosenis dependent upon the number of nodes in the element and the degreesof freedom associated with the element.The following formulae represent the shape functions for eight node isoparametric elements in termsof the naturalcoords(4, ij): expressed

comernodes:

Ni =I (I + ti)(l + lilliAti 4

+ lilli - 1)

(3.10)

midside nodes:

Ni =I

2)+j 4i (I + 44i )(I 71 22

11 i+

TlTli

42)

38

Chapter 3

The Finite Element Method

GAUSS POINT POSITION


30/ 6a

2
-401-

-4yIII' 4T i

TV-4-oL - --L , -_ 07 \07


2

--7

ORDER OF NUMBERINGJ OF ELEMENT NODAL CONNECTIONS

fig. 3.2(a) 8-node Isoparametric Element

corner

Midside fig-3.2(b) Shapefunction for 8-node isoparametric element

39

Chapter 3

The Finite Element Method

These shape functions are part of the serendipity family (Zienkiewicz 1977) and are

inside displacements definition, in fig 3.2(b). By the at any point showngraphically


the element 8 can be expressedin terms of theseshapefunctions:
8 8(4, il) = jNj(4,
i=l

Tj)5j

(3.12)

displacements 8i i is function Ni is the at the nodal of vector and of node shape where nodei.
8 u=jNj(4,7j)uj i=l
8

v=jNj(4,1j)vj i=l

(3.13)

Whereu and v are the displacements parallelto the global x and y axesrespectively. Likewise, the position of a point within the elementin global coordinatescan be definedas:
8

x=jNj(,
8

1j)xj

y=jNj(, i=l

1j)yj,

3.3

The Layer approach

is usedto takeaccount In thecase a layered approach of the of flexuraldeformation, In this system the through thethickness of an element. variation of material property is into layers divided finite to themiddle thickness of either up a number parallel plate integration through pointsareapplied planeof theplate(seefig 3.3(a))or numerical thethickness.
This scheme has been used successfully in the past by many research workers. Johnarry (1979), Hago and Bhatt (1986), employed the system using rectangular

OX, Oy).Later, EI-Hafez (1986) and W, elementswith five degreesof freedom (u,V,

40

Chapter 3

The Finite Element Method

Bensalem (1993) used an eight node isotropic element with five degreesof freedom

in this work. andthis formedthe basisfor the currentmodeladopted

3.3.1 Assumptions
In the model, each layer is assumedto be in a state of plane stresswith a linear strain variation through the depth basedon small deflection theory. The layers are allowed to (fig 3.3(b)). Variation of stressthrough the thickness of resist transverseshear stresses the layer is ignored. Since each layer can be of a different material, in reinforced concrete each constituent material can be assigned a different layer. Perfect bond between all the layers is normally assumed.The main assumptions for the case of plate bending are as follows:

Displacements are small compared with the dimensions of the plate

* *

The stress nonnal to the plateis negligible The normal to the reference surface deformation remains straight but not
necessarilynormal to the referencesurface after deformation (see fig 3.3(c))

3.3.2 DisplacementRepresentation
From the above assumptions, (u, v, w) at any point within the the displacements
structure coords (x,y, z) can be expressedas:
')

u(
v
--

(X,y) - 20. (X,y) U() vo(Xty)-ZOY(X, y) (X, w y) i 0 ,

(3.15)

where uo, vo, wo are the displacements at the plate referencesurfacein the x, y, z directionsrespectively.Ox and Oy are the rotationsof the normal in the xz and xy In this casez is the distance from the reference planesrespectively. surfaceto the midDetermination planeof the layer underconsideration. of the rotationsis asfollows:

41

Chapter 3

The Finite Element Method

IDW(X, Y) ox ax

+,, (X, Y)

Yl

DW(X, Y) +Y(X, Y) Dy

(3.16)

fig 3.3(c)) deformations (see ,, Oy the where and are shear

3.3.3 Strain Displacement Relationship


Since the nodal displacements are now defined in terms of the shape functions, the

derivatives.In in termsof the displacement strainwithin the elementcanbe expressed two dimensional analysis based on Mindlin plate bending and plane stress the straindisplacement assumptions, may be written as: relationship
l

Ex

DNi
ax

0: LN
Dy
:

00

DNi
7x

Ui Z Ni
Dy

y
YXY

0
DNi

0DNi

Vi

DNi

DNi

..... YXZ

x ax Dy ax ...................... ............ ............. ............ DNi

(3.17)
......

0
0

c
c

ax

-CNi
0

0xi yi

Tyz i

0:

DNi Dy

-CNi

where c,,, ey and y,,y are the in-plane strain components, 'Y,, and Yy, are the transverse shear components. The distance from the reference plane to the layer centre is denoted by Z (see fig. 3.3(b)). C is the strain coefficient which is dependent upon the shape of the cross section and is assumed to be equal to 1.0. The strain displacement relationship can be expressed in the simple form previously shown

[Bi I [8i

42

Chapter 3

The Finite Element Method

where [BI] is a 5x5 matrix which contains Cartesian derivatives of the shapefunctions

the formationof which aredescribed next.

3.3.4 Cartesian Shape Function Derivatives


As mentioned previously, the shape functions Ni are expressedin terms of the local natural coordinate system (k, il) of the element, it is therefore necessaryto transform in to global coordinates to obtain the strain matrix [B]. derivatives of the shapefunctions are expressedas: Using the chain rule the

DNj D4 DNj
-j 0

DNj Dx DNj Dy ax a Dy a
DNj ax ax Z, + aNi Dy ay O--, q

(3.18)

andin matrix from:


DNi' -ax

DY

DNi' ax

DNi' ax

(3.19) DNj
-l .a

ax
DTI

ay
OTl

DNj
ay J

DNj
Dy

DNi' ax DNj Dy

DNi' D4 11 DNj ch (3.20)

where [J] is the Jacobianmatrix defined as:

ax

Dy

ax

Dy

LOaq Ohl i

43

Chapter 3

The Finite Element Method

I dz T

fig. 3.3(a)Layer Idealisation

I dz T-

(TY

fig 3.3(b) Layer Plate Model, StressSign Convention

44

Chapter 3

The Finite Element Method

x
y

midplane
actual deformation

deformatioi assumed normalto midsurfac deformation I Fig 3.3(c) Cross Section Deformation of Mindlin Plate

QXZ z Qyz MXY , xy , mx I a- ',


QXZZ
-mly AA-"MY

MXY

MY

44
Qyz
MXY

x
Fig 3.3(d) A Typical Mindlin Plate (positive as shown)

45

Chapter 3

The Finite Element Method

Since an isoparametric formulation is being implemented, i. e. where x=Z y=I Ni yj then: , 8 DNj )i=-: x i
DNj 54 Yi 8 DNj

Ni xi and

ax 74

DX

Xi =Y , d-li O)TI i=l

(3.22)
Dy T4 =8
Dy

DNj I =8 Yi j arl i=l

in termsof the nodalcoordsxi andyj by the following Thus [J] canbe expressed DNj
T Xi [J]='
i=l

DNj Yi 5
DNI
Yi

DNj
-X.

(3.23)

The inverseof the Jacobian matrix is definedas:

a h-x
a Dy ax h Dyj

-Dy Dyi det J dii ax an a ax aI (3.24)

3.3.5 Stress-Strain Relationship


From the theoryof elasticitythe stress strainrelationshipfor eachlayer can be written
as:
x CTY
Icil

=- Txy -=[ DI fel ZU


LTYZ

where D is the elasticity matrix given by:

46

Chapter 3

The Finite Element Method

0
(3.25)

E
_V2

I-V 0 0 2 ...... ........... ............. 0 0 0 0 0 0

2(1.2) 0

0 I-V 2(1.2)

In the above, E and v representYoung's modulus and Poisson's ratio's respectively. The values in the top left portion refer to plane stresses.In the bottom right portion, the values refer to transverse shear stresses,the 1.2 factor is the shear deformation shapefactor.

into a thin layer of steelequivalent A reinforcingsteellayer is assumed to be smeared


to its total area.This smearedlayer of steel is assumedto have unidirectional stiffness is lie. [D] in bars In direction the this the to case matrix which actual corresponding given by

100 [D] = E., 000

(3.26) 0 0.

-0

In cases wherethe steelis positionedat an angle counterclockwisefrom the x-axis, to the global Cartesian the local modulusmatrix is transformed axis.

3.3.6 Element StiffnessMatrix & Force Vector


The informationto evaluate the elementstiffnessmatrix for elasticmaterialbehaviour has been given. As defined previously, evaluation of [K] is carried out from the following equation:
n

[K]

ff [B ]T [D] [B] dx dy ti

47

Chapter 3

TheFiniteElement Method

Where ti is the thickness of the ith layer, n is the total number of layers, [B] is the strain matrix and [D] is the material constitutive matrix depending on the material type (steel/concrete etc.) and the state of stress (elastic, cracked, plastic etc..). The constitutive material matrix will be discussed in the next chapter. Evaluation of the above equation is carried out using numerical integration and Gaussian integration rules are used to integrate over the element area as follows.

[K]

nII

ff

[B ]T [DIB] det [J] d dil t

(3.27)

andnumericallywritten as:
nmm

[K] =

Yal:

YaWjWk[B]T

[D][B] det [J] ti

(3.28)

i=l j=l k=l

where rn is the number of Gausspoints in each direction, n is the number of layers, wj

to the specifiedGausspoint with and Wkare the weight coefficientscorresponding local coords(4,

The equivalent load vector at the nodes due to the effect of uniformly distributed element surface loads is defined as:

[F]

f [N]T (q) ds s

(3.29)

or in numericalfonn:
mm

IF) --1: 1: W [N]T (q) det [J] jW k j=l i=l

(3.30)

48

Chapter 3

The Finite Element Method

3.3.7 Numerical Integration


In numericalintegration,the exact integralis replaced by evaluatingthe integrandat varioussamplingpoints and then taking a weightedsummationof thesevalues.For due to its relative easeof Legendre this work, Gauss quadrature valueswasemployed implementationand high accuracy. In this method an n-point rule integratesany 2n-1 degree polynomialof x , or less,exactly. In general, Gaussian formula takesthe form: Quadrature the one-dimensional
+1

ff (4)d4

wif (4j)

(3.31)

4i where is the coordinate of the ith integration point, wi is the weighting factor and m

is the total numberof integrationpoints.For the caseof doubleintegrationi.e. over 2 dimensions the following form is taken:

+1+1

fff

(4, il) d4dil

+1 m

lwif(4i,

il) d7l I dil

+1li=l m j lwigi(il) I
mm YYWiwjg(TIj) i=l j=l
mm IIwiw i=l j=l

(4i, TO jf

where wi, wj are the ith and jth weighting factors and 4i iIj are the coordinates of the ith integration point. The fact that limits of integration (-I / +1) coincide with the local natural coordinate system on the element boundaries makes them particularly applicable to isoparametric elements.The symmetrical positions of Gausspoints 4i and the corresponding weighting factors wi for m=1-4 are given in table 3.3.7

49

Chapter 3

The Finite Element Method

3.4

Non-linear Solution Techniques

In a non-linearproblem the relationshipbetweenthe force vector and the behaviour displacement is longer linear. in As the of material case plastic vector no displacements thecurrent atearlierstages. maydepend onthedisplacements
In finite element analysis, the displacement vector is calculated such that a state of equilibrium is achieved between the external and internal forces. Unlike linear analysis, the solution vector in a non-linear analysis cannot be found right away. In non-linear analysis, the loading of the structure is divided into a seriesof increments.

In order to achieveequilibrium at the end of each increment,an iterative solution A purely incremental algorithmis employed. unless methodcould leadto inaccuracies very small increments areused.In an iterativemethodthe occurringerrorswithin the incrementare successively Hence,most solutionsin non-linearanalysisare reduced. basedupon the incremental-iterative method.The generalprocedurein this method involvesadapting 8u until increment Au by iterativeincrements the total displacement The incrementaldisplacements tolerance. equilibrium is reached within a pre-defined from at iterationi+1 arecalculated
AU , i+i=,u, +

8ui+i

(3.32)

There a number of iteration procedures which calculate 8u in different ways. The iterative increments are calculated via the stiffness matrix K which represents a

linearized form of the relation between the force vector and the displacement vector. The stiffness matrix can change for every ith iteration. Hence the iterative increments can be found by the following:

8ui = Ki -1Ri

(3.33)

whereRi is the residualforce vector at the beginningof the ith iteration. One of the most commoniterative methodsis the Newton-Raphson method.Within the method

50

Chapter 3

The Finite Element Method

itself there are two variations, the second of which is known is the Modified Newton

Raphson method.

3.4.1 Standard Newton-Raphson


In general the stiffness matrix Ki representsthe tangential stiffness of the structure:

DR Ki = DAU

(3.34)

iteration, the stiffness relation shown above is In the normal Newton-Raphson (eq. 3.33) the predictionof the iterativeincrements evaluated everyiteration.Therefore is basedon the last-knownor predictedsituation,regardless of whethera stateof ton-Raphson is Since Ne, the normal equilibrium achieved. methodyields a quadratic The characteristic, only a few iterations are neededfor convergence. convergence of this methodis that the stiffnessmatrix hasto be set up at every main disadvantage iteration.Second to this, if the linear equations are solvedthrougha direct solver,the at every iteration as well, see fig. 3.4.1. If the matrix will have to be decomposed initial predictionis far from the final solution,the methodwill fail dueto divergence.

3.4.2 Modifled Newton-Raphson


In this method the stiffness matrix is only evaluated at the start of each increment which means the prediction is always based upon converged equilibrium state. In general the modified Newton-Raphson technique converges to equilibrium slower than the standardmethod. The advantageof this method is that for every iteration only the prediction of the incremental displacementsand the internal force vector has to be calculated, it is not necessaryto set up a new stiffness matrix, seefig. 3.4.2.

In comparisonwith the standardmethod the Modified Newton-Raphson iteration


sometimes provide convergencein problems where the standard method has failed to converge.

51

Chapter 3

The Finite Element Method

3.4.3 Incremental procedures


In the previous section the iteration techniquesused in non-linear analysis were discussed. This section deals with the incrementaltechniquesemployed in the incremental-iterativemethod. The two most common techniques are load and displacement control:
load is increased in described Load the the section, external previous 9 as control: feg. increasing force increment by the the vector external at start of each form is in load Displacement the the applied of prescribed external 9 control:

d. displacements

As can be seenfrom figure 3.4.3 the unloading branch of the load-displacement curve it displacement however be In using control. cases may be can obtained when real necessaryto obtain the unloading branch of the load/displacementcurve which results from a given design load. The Arc-length or Riks method can be used to obtain the required results in this case,(Crisfield 1991).

3.4.4 Convergencecriteria
In the numericalprocess the equilibrium conditionsareunlikely to be satisfiedexactly have to be defined. The convergence and hencecriteria to determineconvergence criterion will monitor the out-of balance residual forces until a desired level of hasbeenachieved. Convergence accuracy criterion canbe basedon the out-of balance force norms,the displacements or the internalstrain energy. The methodadoptedfor the presentwork is basedupon the out-of balanceforce norms. It is difficult and expensiveto check the decayof eachresidualforce for eachdegreeof freedomand henceforce normsareusedto achievean overall evaluationof convergence. The criterion assumes convergence whenthe following conditionis met:

ARi* Fi*

(3.35)

52

Chapter 3

The Finite Element Method

where:
JjRj F(Ri)

ARj = = norm of the residuals (Ri) = residual force vector at ith iteration V(Fi)T F, = JFjI= norm of the total applied loads I Fj )= total applied load vector 15 = specified convergencetolerance In the interest of accuracyit is desirable to set as fine a tolerance as possible, however this has to be balanced with the need to reduce computation time i. e. a finer the tolerance usually requires a higher number of iterations. The required number of iterations will often increase as more non-linear phenomena (e.g. concrete cracking, concrete crushing, steel yielding etc.) are encountered . These discontinuities in the in high laws result residuals having to be distributed which may result in material further discontinuities in other parts of the structure.

53

Chapter 3

The Finite Element Method

force (f)

f"t

fim

fe.

displacement (u) iteration fig. 3.4.1Standard Newton-Raphson

force (f)

t+At fet fim

t fext

displacement (u)
fig. 3.4.2 Modified Newton-Raphson iteration

54

Chapter 3

The Finite Element Method

41 0
1 11 +1/43

a,
2 +1

1
2

1 H 1

-1143 0 +40.6 -40.6 F --4.8! +4 7 U8 F4.8 = r,4 L 8 ,F 7 4 4:8

Id

+- _F 7

+1 8/9 5/9 5/9 430 1 2 36 1 -Nf3-O

2 36
r31 -, 0 + 2 36 430 1 + 2 36

IV

WeightingFactors Point PositionsandCorresponding Table 3.3.7Gauss

f
fext3 fext2 f,,, ti

cl u

c2 u u0

a) loadcontrol

b) displacement control

fig. 3.4.3 load& displacement control

55

Chapter 4

Model Visualisation & Direct Design

Chapter 4 Model Visualisation & Direct Design

4.1

Introduction

This chapterdetailsthe visualisationprocess usedin the formation of strut-tie models in this work. The methodis alsoappliedto slabsandis usedin conjunctionwith direct layouts.The resultsof this work aregiven in designprocedure to derivereinforcement direct design here. Details the chapters. of method are presented subsequent

4.2

The Visualisation Process

defining As a means themajorstress of automatically strut-tie model, paths andhence is proposed. is adapted from the structural This method, procedure an evolutionary by Me & Steven(1993),(1994).The general work first presented optimisation leads to isolation of themainstress andhence aid in process paths within thestructure for a givenloadcase. theidentification of suitable strut-tie models 4.2.1 Theory
The process begins with an elastic analysis of the original structure. It is often found that parts of the structure are lowly stressed and can be removed without affecting the overall strength. In the finite element mesh, a low stressed element can be removed by assigning a negligible stiffness, hence it is not necessary to redefine the mesh. As a criterion for element removal, the average vonMises stress present within the element, vM. is (Yevm, compared to the maximum vonMises Stress present in the structure, Omax If aevm is less than a certain percentage of amaxvm known as the rejection ratio (rr), , i. e. if aevM < rr (OmaxVM), it is removed

56

Chapter 4

Model Visualisation & Direct Design

The von Mises stresswas chosenas for this purpose becauseit is a measureof all the stresses present, and was calculated from the following: VM ae V(C; + 2-( Cartesian ((Yx, the +3Txy2), rxy) stresses. ; (; (Yy,, are where = xay x2 Y

Each time an element is removed, the finite element analysis is carried out using the same value of rr until a steady state is reached, Le no more elements are removed. When this stage is reached,rr can be increased.This process is repeateduntil one of the following criteria is met: " " the main stresspaths becomeclear rr becomes too great, typically a maximum value of 35% is used for rr. Any elementsremaining after this point are not consideredto be lowly stressed

becomes the structure unstable

by Oncethe main stresspathshavebeenisolated,the strut-tie modelsare generated placing strut or tie membersalong the centre lines of the main stressfields. It is degree to the model in some necessary exercise of designexperience when generating termsof the stability of the resultingtruss,and in termsof the practicalconstraints on layouts. The object of the visualisationprocessis to obtain a set of reinforcement stressdistributions along clearly defined paths which are in equilibrium with the loads. external

4.2.2 Examples
Three examples of the visualisation process are shown. The first, shown in figure 4.2(a) is a simply supported deep beam having a span/depthratio, (1/d), of 1.67, with a vertical load at the centre. The second, shown in figure 4.2(b), is the same beam cantilevered with a vertical load at the free end. The third, is similar to the second example but with a smaller span/depth ratio of 1.0. In each case, a clearly defined stress path is formed at a rr value of between 20-25%. The second and third case illustrate the effect of span-depth ratio on the load path behaviour. Strut-tie models from the visualisation processare illustrated in figure 4.2.3(d). In the case of resulting the deep beam, the resulting strut-tie model is statically determinate since the diagonal part ABCD is acting as a rigid block and hence a mechanism is not formed. The reason for the incorporation of the ties in the diagonal truss is to take account of the transversetensile stresseswhich can be clearly seenfrom the stressplot 4.2(a).

57

Chapter 4

Model Visualisation & Direct Design

#+

\'I'/
jio s \ -*, -4k, -k & 8, 0*0 .0. / //

A'

Jr xA

w e jo

IV

or fII

iT

a %wA * !. f)r iI
jr -%,, ia% x% %" ,x., % 1%
lb

II
f q, 1

'-tle.
io,'it
0, it -w

*%
% 7.1,

o,

--- i -W - 4

'di

0 01t

i) rr=O

90a
%%%
qb %

**
v p O

p rov

e,
1-19

ft .o

adr IF

compression tension

--------

t
1 .-o

.:

e
.,
t.

a -2

lb

4a

#4

..

0: 0a

40

.,

.*6

vi e0
40 0

ii) rr= 10%

tf,

iii) rr=25%

fig. 4.2(a) Deep Beam (1/d=1.67): Principal StressesDuring Stages Visualisation of

58

Chapter 4

Model Visualisation & Direct Design

i) rr=O

El compression 0 tension --p-

-.I*-

ii) i-r=10%

i1i) rr=20%

fig. 4.2(b) Cantilever, (1/d=1.67):Principal Stresses During Stagesof Visualisation

59

Chapter 4

Model Visualisation

& Direct Design

qjjj-

i NA 'I'A x 'X "A

i) rr=O
"A-A VIVA
Z4 61,WA

""A
t44

A V Ar 1 0 Ir 0 011;
A-A V. W,. wxO W00
:'

,II

, W'X OWAO

.,

El compression m tension

ii) rr=15%

-4

ni) rr=20%

fig. 4.2(c) Deep Cantilever, (1/d=1.0): Principal StressesDuring Stagesof Visualisation 60

Chapter 4

Model Visualisation & Direct Design

V
/--

': c::
///
\

___

Lz
/// ///

(/j

//

___

/// Dd /// / /

J /

B :

\\\

\\\ \ \\\

\\\

I ii
II

///

\\

iI

E T _ _
/

X\

Lx/ X

//

1 \17\ -x

/d
X

IA*x

Fi t '

strut tie

fig. 4.2(d) Strut-Tie Models Resulting from Visualisation

61

Chapter 4

Model Visualisation & Direct Design

In the case of the deep cantilevers, the strut-tie members are not always aligned

aroundthe exactlyalongthe main stress paths.This is dueto the fact that the stresses nodalzonesaresmeared.

4.2.3 Application to Slabs


In some cases, such as in simply supported slabs, the areas of low stress may occur around the supports. In this case, the visualisation process would result in instability since the elements along the supported edges are assigned negligible stiffness. For this reason, it is sometimes necessary to stipulate prior to visualisation, which elements must not be removed. Using this method, it is also possible for the designer to direct the stress paths according to a pre-determined pattern. This may be desirable in order to achieve practical reinforcement layouts.

4.3

Direct Design of Slabs

The direct design method combines analysis and design into a single operation. It is a computer orientated method enabling the structure to be designed with the minimum of designer intervention.

In this method the basic requirements of classicalplasticity i.e. equilibrium, yield


condition, mechanism and ductility demand are theoretically satisfied. Equilibrium condition: Any stress disiribution in equilibrium I'with the applied

loadscanbe usedfor design. In the proposed fields areobtained methodthe stress


using FE analysis of the unreinforced concrete structure with the uncracked

of the concrete properties so that the equilibrium conditionis satisfied. directly from the o Yieldcondition:The requiredsteelfor the structureis determined yield criteria. Therefore,the resistance providedby concreteand steel is equal to or greaterthanthe appliedstresses. * Mechanismcondition: The resistance at eachpoint in the structureis matchedas
closely as possible to the applied stresses.This means that at ultimate load, all points in the structure attain their ultimate strength with a minimum redistribution of the stresses,thus converting the structure into a mechanism

62

Chapter 4

Model Visualisation & Direct Design

9 Ductility demand: In classical plastic theory the material is assumedto possess

that yields early in the that anypart of the structure unlimited ductility. This means loading history, will continueto deform without any reductionin strength.This requirementis avoidedwhen the differencebetweenthe yield load and ultimate load aremadeassmall aspossible.As a result of this, the early yieldedpoints can deform at constantstressbefore reachingthe descending branch of the material stress-staincurve. In the direct design method this condition is satisfied automatically sincetheoreticallyall partsof the structure yield simultaneously. The first method for the provision of reinforcementfor slabs accordingto elastic by Wood theory was proposedby Hillerborg (1953).This methodwas re-examined (1968) who establishedsimple rules and equationsfor the optimum steel in slab elementssubject to a moment field (M,,, My , my) without membraneforces. Woods' equationsfor orthogonalsteel in the top and bottom face of the slab were by Armer (1968)to coverskewreinforcement. extended

4.3.1 Assumptions
The main assumptions in the direct designapproach aresummarised asfollows;
The reinforcing bars are only able to carry uni-axial stresses in their original directions, i. e. dowel action and bending of the bars is ignored

e The bars are elastic-perfectlyplastic with yield stressf, in tension and f, ' in (a) compression, seefigure 4.3.1 o The bars are taken as an areaper unit width ratherthan individual bars , because the bar spacingis smallcompared with the overall dimensions of the slab. is ignored. 9 The tensilestrengthof the concrete e The concreteis perfectly plastic, satisfying the squareyield criterion shown in figure 4.3.1 (b) by is 9 Instability or bond failure of the bars is assumed to not occur and avoided
proper choice of section and reinforcement.

63

Chapter 4

Model Visualisation & Direct Design

Response Reinforcement SteelStress-strain fig4.3.I (a) Assumed

Compressi

fig. 4.3.1(b) SquareYield Criterion For Concrete

64

Chapter 4

Model Visualisadon & Direct Design

4.3.2 Yield criteria for reinforced concrete slabs


Consideran elementof a slab subjected to bendingandtorsionmomentsas shownin in My* is M,, * figure 4.3.2(a).The flexural strength by the x and the and slab provided is directions (fig 4.3.2(b)). The a mathematical criterion yield y respectively Hence between the the the of material. strength and appliedsetof stresses relationship the yield criteria for the slabelementcanbe written as:
F(M.,, My My Mx*, My*) = 0.0

(4.1)

Now consideras shownin figure 4.3.2(c),at any point in the slabelementa line with a normaln anda tangentt. The normalappliedmomentMn mustnot exceedthe value direction. in by that the the slab generated of momentof resistance Taking the normal to the yield line at an angle0 to the x-axis, the equilibrium of the in figure in following: 4.3.2(c), the element results
Mn MxCOS -"-:
20+M 20

ysin

2M,, ysin0cosO

(4.2)

Mt = M,, sin'O + Mycos2O + 2M., ysin0cosO

(4.3)

M.

(Mx = t

My)sin0cosO

+ Mxy(COS20

- sin

20)

(4.4)

Resolving the resistance moments of the x and y bars, fig 4.3.2(d), and ignoring

torsion, the resisting normal moment at the yield line can be expressedas the following:

M,, *cos'O

+ My*sin2O

(4.5)

where the value of M. * must always be greater than Mn from equation 4.2, i. e.

M,* - M, : 0.0
Substituting equations 4.2 and 4.5 into 4.6 leads to:

(4.6)

65

Chapter 4

Model Visualisadon & Direct Design

4.3.2 Yield criteria for reinforced concrete slabs


Consideran elementof a slab subjected to bendingandtorsion momentsas shownin in My* is M,, * figure 4.3.2(a).The flexural strength by the x and the and slab provided is directions (fig 4.3.2(b)). The a mathematical criterion yield y respectively Hence between the the the of material. strength applied set of stresses and relationship the yield criteriafor the slabelement canbe written as:
F(M, My My M,, *, My*) 0.0 =

(4.1)

Now consideras shownin figure 4.3.2(c),at any point in the slabelementa line with the value a normaln anda tangentt. The normalappliedmomentM,, mustnot exceed
in direction. by that the the slab of generated of moment resistance

Taking the normalto the yield line at an angle0 to the x-axis, the equilibrium of the elementin figure 4.3.2(c),resultsin the following:
Mn MxCOS
20 + Mysin 20

2M,, ysin0cosO

(4.2)

Mt

Msin

20+M

2() YCOS

+ 2Mxysin0cosO

(4.3)

Mnt

= (Mx

My)sin0cosO

+ Mxy(cos2O

- sin

2())

(4.4)

Resolving the resistance moments of the x and y bars, fig 4.3.2(d), and ignoring

torsion, the resisting normal moment at the yield line can be expressedas the following:
2o

M,, *cos'O

+ My*sin

(4.5)

where the value of M,,* must always be greater than Mn from equation 4.2, i. e.

M. * - M, : 0.0
Substituting equations 4.2 and 4.5 into 4.6 leads to:

(4.6)

65

Chapter 4

Model Visualisation & Direct Design

(Mx*

MX)COS2()

+ (M

y*-

My)sin

2E)

+ 2Mxysin0cosO : 0.0

(4.7)

The above equation can be written in simplified from by taking:

A=M,, * - M,, B=My*-M c= MXY y

(4.8) (4.9) (4.10)

20+B 20+ 2cos0sinO t 0.0 A cos sin

(4.11)

Dividing by

COS20

equation 4.11 becomes:

20 F(O) =A +Btan + 2CtanO: 0.0

For optimumsteel,the excess strength mustbe a minimumhence:


dF(O) -

=0 dtanO

(4.13)

d2F(O) dtanO

>0

(4.14)

dF(O) BtanO+C= =B -d tanTO

O=> tan 0 =--

(4.15)

d'F(O) 2o d tan

>0 => M*>M yy

Substituting equation 4.15 in equation 4.11 results in:

A+B(-C)

2+ 2C(BB

)=O

(4.17)

66

Chapter 4

Model Visualisation & Direct Design

or AB

C2 =0 _

(4.18)

Replacing A, B and C by their values in equation 4.18 gives:


2=0,, mxy -m Y) +

(Ml*

MX) (W

(4.19)

The aboveequationis the yield criterion for reinforcedconcreteslabsknown as the Wood criterion. This is the yield criterion for positive steel (i.e. bottom steel). For
negative (i. e. top steel) a similar procedure is adoptedto give:
I- my) + mxy 2=o

(M, *, + mx) (my* -

(4.20)

Experimental work by a number of researcherssuch as Cardenas and Sozen (1973), Jain and Kennedy (1974) have confirmed the validity of the yield criterion for orthogonal steel. In the work by Hago and Bhatt (1986), elastic stress fields in conjunction with the Wood-Armer criterion was used for the design of orthogonally reinforced slabs. The method was found to be a highly practical procedure leading to lower bound solutions to slab design. Bensalem (1993) continued this work with the direct design of slabs using the non-linear stress-field. It was found that in many cases, design from the non-linear field helped to reduce steel congestion by 'smoothing' out the peak moments occurring at concentratedloads or supports.

67

Chapter 4

Model Visualisation & Direct Design

ff

--

MV

mly

fig. 4.3.2(a) +ve Moments on a slab element

M,

y my

mx

mxy

mxy

mx

MY mly

fig. 4.3.2(b) Applied and resisting moments on an element

68

Chapter 4

Model Visualisation & Direct Design

4.3.3 Design equations


The following design equations can be derived from the yield criterion described above

i)Positive Moment Field (bottom steel)


m2 mx*=. my xy *-M y + M,

(4.21)

The total amount of bottom steel providing resistanceMx* and My* is representedby the following equation:
m2 M, *+m y my "y -M y +M,, +M y

(4.22)

Which implies for minimum steel:

d(Mx *+My*) dM Y*=0,

hence My* = My + M,

or ,

since in equation 4.16, My* > My, this reduces to:

MY*= MY +1ml

(4.23)

Substitution of equation 4.23 into equation 4.19 results in :

M,, *= M, +lmyl

(4.24)

ii)Negative Moment Field (top steel) The same procedure as above is applied to the negative yield criterion (eq. 4.20) to obtain the following:

mx *t =M X -Imxyl MY *t =M y -Imxyl

(4.25) (4.26)

70

Chapter 4

Model Visualisation & Direct Design

iii)Mixed Moment Field From equations 4.21 and 4.22, if My* <- 0.0, is considered to be equal to zero and
from the yield equation (4.19), the following is obtained:

mx *=Mx+

MY

(4.27)

similarly if M, * <-0.0, from equation (4.19):

MY

My+m

xy, mx

(4.28)

The sameprocedure as above can be applied to the negative moment fields to obtain:

MxY2 Mm+ xx M-

(4.29) y

My

=MY +

mx

(4.30)

4.3.4 Procedure for Placing of Reinforcement


Given a stressfield (M,, My M, at any point in a slab the reinforcement in the x y) , , and y directions can be placed according to the following:

i)Bottom Steel
Design moments M, *and My* are calculated frorn equations 4.24 and 4.23 If M, *and My* are negative, then no bottom reinforcement is needed If M, *and My* are positive, then the calculated values are adopted as the

resistancemoments
* e If M, * < 0.0, then set M, * = 0.0 and calculate My* from equation 4.28 If My* < 0.0, then set My* = 0.0 and calculate M, * from equation 4.27

71

Chapter 4

Model Visualisation & Direct Design

fijop " " "

Steel

Design moments M, *' and My*t are calculated from equations 4.25 and 4.26 If M, *' and MY*' are positive, then no top reinforcement is needed If M, *' and MY*' are negative, then the calculated values are adopted as the

resistancemoments
*
*

If M, *'> 0.0, then set M, *'= 0.0 and calculate My*'from


If My*'> 0.0, then set My*=

equation 4.30.

0.0 and calculate M, *' from equation 4.29.

A schernatic representation of the design equations for bottom steel is given in figure 4.3.4(a).

4.3.5 Multiple Load Cases


The rules outlined so far only deal with a slab subject to a moment field resulting from a single load case. In practice, slabs such as in the case of subject to multiple bridge decks, can be is provided to

load cases. For this situation, reinforcement

accommodate the moment triad resulting from multiple load cases; (M, i, Myi , M, yi) i=l, n, where n is the number of loading cases. Here a rnethod for provision of reinforcement in multiple load case situations is presented. The procedure was used by Kernp (197 1) and was also applied to skew reinforcement.

For each load case i( i=I, n), the moments (M, i, Myj M, The are calculated yi) , . corresponding resistance moments M, *i and My*i are found using the procedure

outlined earlier.
9 At each point the maximum value of Mx*j and My*i can be found. Once the

maximum values are obtained i.e.

Wma,

and

My*max

be used for they can ,

design. The resulting design would be safe, but not necessarily econornic. Hence

an optirnisation procedure must be adopted 0 The next step is to assumethat in the x-direction, M, *,,,,, is provided for, but in
the y-direction My*i is provided for to satisfy the corresponding yield condition in
each case. The maximum value of all these My*i is found, let it be My*plll,,Ix Using .

Mx*11, in conjunction with My*pllllx design is safe produced. Ix ,a

72

Chapter 4

Model Visualisation & Direct Design

Using the same procedure as above, a corresponding value Of Wpm

can be

found for My*inax.Clearly, the optimal design would result from a set of design
moments where (M, * + My*) is the smallest.

Further optirnisation

of the design can be carried out by using a simple search

technique to examine the feasible design region as shown in figure 4.3.5. For each load case, design moments at the grid points is examined in order to determine if it is a better minimurn. If the search is positive, a check is made to ensure yield conditions design is is If then the the moment rejected. are not violated. yield condition violated, If not, then a check is made to see where on the grid the best minimum value of (M, * +My*) is obtained.

4.4

Inplane Application

In 1964,Nielsen first proposed a yield criterion for a sectionwith known orthogonal and placed symmetrically reinforcement,which can carry tension or compression, with respectto the section's middle surface.Equationswere derived to determine to resista given inplaneforce triad. In 1984,Nielsen orthogonaltensionreinforcement consideredthe case of skew tension reinforcement.His work assumesthat the concrete has sufficient compressionstrength without the need for compression When the compression the section strengthof the concreteis reached, reinforcement. thickness must be increased.In 1976, Clark proposeda series of equations for tensionand/orcompression to resista proportioningskewor orthogonal reinforcement haveusedtheseequationsto design triad of inplaneforces.A numberof researchers deepbeams.Khaskhell ( 1989)usedtheseequationsto derive reinforcementlayouts for deepbeamsfrorn the elastic stresspattern.Bensalem(1993) usedthe non-elastic layoutsfor deepbeams. to derivereinforcement patterns stress

73

Chapter 4

Model Visualisation & Direct Design

my/im,yi

M, +IM,,

/M, l

M, , = M, + IMYl NI, *- M, + IMJ

lm, ),

-----------

M, *= 0 MY*= 0

M'.

= M,

+I

M2, YIMY

MY *=0

Mmy=

My 2

-'y

fig4.3.4(a) Design Equations for Bottorn Steel

(M*)
(M, * I'llux, MX*Max)
(Willax I I W[MIX)

(my, *

pillax ,

mx* p"lax)

(Mx*max,

Mx*pmax)

(M*)

fig. 4.3.5 Simple SearchTechnique for Use in Multiple Load Case Design

74

Chapter 4

Model Visualisation & Direct Design

4.4.1 Yield Criteria


fig. 4.4.1 (a) by defined in a concrete The principal stresses a, CY2 see and are element The 0 to the the areaof reinforcement x angle stress at an with major principal -axis. by A,, Ay, is denoted directions in length the with their associated and x andy per unit From the equilibrium of the reinforcedconcrete being f,, and fy respectively. stresses by inplane being having 4.4.1(b), fig. t thickness acted upon a of and element,see for following length (n,,, forces the equations ny, n,, per unit normal and shear y), canbe derived: orthogonal reinforcement
2 0+02tsin 20

nx = Axfx + alt cos

(4.31) (4.32) (4.33)

0 0+ Ayfy tsin2 + (y2t COS2 (y, ny = nxy =-ajtcos0sinO+G2tCOSOsinO

setting:
Ayfy

nx xyY, ttttt

nn

'rxy = -ly- and axy

Axfx

*, the the r., cry, cF., and c; cT,,, are, normal shear stresses, and are resistant where y y*

in by the x and y directions respectively. the steel reinforcement stresses provided 4.31 to 4.33 canbe written as: Equations
20 20+

Ox = (71 Cos

+'92

sin

20

+(;

x
y*

(4.34) (4.35) (4.36)

ay = a, sin
'rxy = ((Y2

(Y2 COS2 0+ (y

- al)cos0sinO

If tensile steel is to be provided, then (y, = 0.0, and equations 4.34-4.36 become:

(yx (yy

(Y2sin

2 0+(y

(4.37) (4.38)

20+ (Y2COS (yy

75

Chapter 4

& DirectDesign ModelVisualisation

IrXY= (02)cos0sinO

(4.39)

hence:

ax * -Crx = (72 sin * -(T (Tyy' Irxy = 02)

2o

2o -2 (72 COS cos 0 sin 0

Eliminating'92 and 0 from the above equationsleads to:

2 * -cr (ax * -(; "XY =0.0 x)(Gy Y)

(4.40)

Equation4.40 is the yield criterion derived by Nielsen (1964) for a sectionhaving known orthogonalisotropic or orthotropicreinforcement carrying tensionforces and placed symmetrically with respect to the mid-surface of the section. From this for four differentcases designwereproposed. As of reinforcement criterion,equations that concretestrengthis not violated. Clark already stated,this criterion assumes (1976) extendedthis criterion for the provision of compressive reinforcement.The four casesoutlined by Nielsen, were extendedto nine. Table 4.4.1(a) shows the (in the table p is the reinforcement of reinforcement possiblecombinations ratio and ot is the angle of the reinforcementfrom the x axis when skew reinforcement is From this table, it can be seenthat all casescan be solved by direct considered). solution exceptcasesI and 4 whereminimisation of the total reinforcementin both directionsof the memberis necessary. The principal stressal is considered equal to is requiredand02'equalto the concrete zero whentensionreinforcement compressive strengthfc when compressive reinforcementis required. Derivation of the design is now described in the following section. equations

76

Chapter 4

Model Visualisation & Direct Design

Ax, fx

(T-)

fig. 4.4. I (a) Principal Concrete Stressesand Reinforcement Directions

ll'y

ny

11,

fig. 4.4. I (b) Sign Convention for Inplane Normal & Shear Forces per unit length

77

Chapter 4

Model Visualisation & Direct Design

Case I 2 3 4 5 6 7 8 9

Reinforcement Description Both tension No x a tension No (x x tension Both compression No x ot compression No ot x compression x tension cccompression x compression a tension No reinforcement

Known Values fx = f. = fs, al =0 f. = f" P" = 0, cr, =0 f" = fs, P. = 0, a, =0


fx = fa = fs's 02 = fcu

Method of Solution Minimisation of (p,, + py) Direct solution Direct solution Minimisation of (p,, + py)

fa= fs'qPx Os a2 fx = fs't P.

fcu

Direct solution Direct solution Direct solution Direct solution Direct solution

09G2 fcu

fx = fsq f= fs2 (; = 09 (; = fcu 2 9 1 fx = ftl f= fst C; I= P OP 02 = fcu

Px= P. =O

Table 4.4.1(a) Possible combinations of reinforcement

4.4.2 Design Equation Derivation


4.4.2.1 Caseswhere tension steel is to be provided
2

* anday* >0 I: a., * Case


4.40, a From equation yy xy CY+(Gx * -Cyx)
d daX

The total provided steel in the x-direction is minimum when


d d(YX *I 2 Txy (ax * -(y 01 ((Y rXY2 -=0 x* -(; x)2

(ax *+GY

0,

Thus:

(yx * +(Yy +

Since ax*> (y -. . x9ax*=0x+I,


P" = lffjox +jTxy I)

the steelratio in the x-directionis:


py = 1/fs(ay + ITXYI),

similarly in the y-direction:

f., is the yield strength of the steel. where

78

Chapter 4

Model Visualisation & Direct Design

o Case2: a.,* =0 anday* >0


From equation 4.40, (; y (Ty--, Ox
L2 xy CXY2

thus px = 0, and py =I/f,

((Yy -

ax

Case 3: a,,* >0 and (yy* =0 ,r2 OY

Similar to case 2, here py =0, and p., = I/ fr. (ax -

4.4.2.2

Caseswhere compression steel is needed

For this situation,the minor principal stressin the concrete reaches ultimate strength, 4.34-4.36 be i.e. cr 0.0. Equations 2= a, < and can written as -fe ,,,
20-f, ax = Gl COS

,,,

sin

20+

Cyx

(4.41) (4.42) (4.43)

ay = a, sin

20_f

0+ (T COS2 y cu

Txy = (-fcu - a, ) cos0 sin 0

thus:
ax+fc

,U

=Glcos

20+f

0+(; COS2 cu

(4.44) (4.45) (4.46)

20+f,, 2+ fcu oy + sin = a, sin u Txy = (-fcu and: cos 0 sin 0

(yy

ox

+fcu

-(TX*

=(a,

+fcu)

Cos

2o

2o fcu *= (a, fcu) + ay + sin - ay

Txy =-(fc,, +al)cos0sinO

fcu - OX *)(a + -(c; xyy

+ fc

-a

*) + TXY2 = 0.0

(4.47)

For the following sections, a.,f = cy., + fc and (Yyf = c;y +

79

Chapter 4

Model Visualisation & Direct Design

o Case4: a,,* <0 anday* <0


From equation 4.47, ay
+0 yf -T (Gxf *) (Yx XY

provided steel is minimum when:


ITxy
((Y *) (7 x xf -

d (Gx *+Cry dcTX *

0,

Thus: I-

T=O,

1r I, in the x-directionis: Since ax *< (T ax *= (7 the -. steel ratio . xf xy xf )


I)

px =I/

fs'(axf

-Irxy

in the y-direction : py =I/ similarly ,

fs'(cryf -lTxyl)

* =0 and (; <0 Case5: (Y., y*


2 xy thus: (Tyf . Gxf CXY2 Yf - (Y xf

From equation 4.47, (y y

px = 0, and py =I/

fs'((;

6: * <0 and cry* =0 Case cr., lo

similar to case5, py = 0, and p., =I/f., '(a,, f

Ir Oyf

4.4.2.3

Mixed Cases:

where a,,* and ay* are of different signs:

Case7: a.,* >0 and(YY*


0

a, =0 and a, = fr, thus equations 4.34-4.36 become: ax = fr sin 20+

.u

(y

X*=[fcu/2.0](I`

- cos20)+aX*

(4.48)

80

Chapter 4

& DirectDesip ModelVisualisation

y=

fcu Cos20+0 ,,,

y*=[fcu

/2.01(1+cos20)+

aY

(4.49)

f, Txy , =

20 / 2.0] 0= [f,,, 0 sin cos sin

(4.50)

2, r

2 1/2

From equation 4.44,

cos20 = [I C.

Eliminating 0 from equations4.48 and 4.49 gives:

0) / 2.0)(1 (f,,, *= ax cyxP) / 2.0)(1 (fcu *= + ay (Ty-

The steel ratios are:

p,, =I/f,

[c;,, - (fc,, / 2)(1- P)l

P)l / 2)(1 (fc,, fs'[ay + py =I/ -

Case 8: a,,* <0 and ay* >0

Similar to case7;

px =1/ fs'[ax - (fcu / 2)(1 + 0)] (fcu / 2)(1 fs[ay py =I/ -

Case9: No steel required, px = py =0

4.4.3 Boundary Curves


Given the stresses(a.,, (3y,Tyx)at any point, it is necessaryto find out which design horizontal be The boundary between to axis a are used. surface on each case equations of ( cr,/Irxyl ) and vertical axes of ( a,,/Irxyl ), can be constructed to define the required The figure 4.4.3. In figure 4.4.3, the the cases. see circled numbers represent case,

81

Chapter 4

Model Visualisation & Direct Design

boundary curves were derived by equating the design equations for two cases.The

2: I following describes the derivationof the curveseparating cases and

CrX PxI =-Px2 =: (Fx -ITXYI=> ITXYI = -1.0 The sameequation can be derived by equating the expressionsfor steel ratios in the ydirection.
Pyl = Py2 =:> (Ty + jTxy I =CF Y-- CFX 2 XY (TX => IrXYI = -1.0

This is the equationof a straightline and is shown as line number8 in figure 4.4.3.
Table 4.4.3 shows the boundary equationsfor the intersectionsbetween each case.

Gy

i,

Gx ITXYI

fig. 4.4.3 Boundary Curves for Orthogonal Reinforcement, f,,,= -41T. yl

82

Chapter 4

Model Visualisation & Direct Design

Curve

Equation

Curve
r

Equation

I Irxyl
Cry I =2 F'r--Y1 fU FITXY I+

cry.

7 = +*0
(Fr fc I xUy 2 -4 -1/2)

CY x1 =2 IlrXyl

fcU Jc I xy

fcU Jr 1, -4 xy

)2

-1/2)

llx llrxyl

cyY

(YYf G xf IT I -=1 IT I xy xy

fcu Gy + -XY Fr _Y1 Fr 1 x

10

x -XY I 1TxYFT

- 1/2) CY

CY

f +1

y y12

fcu Ic I xy crY llrxyi

fcU lrxyl -4 11

ITXYI ITXYI

-1/2)

--00 12

ax IT 12 xy

fcu + IT I xy

fcU oixy, -4

Table 4.4.3: Boundary Curve Equations for Orthogonal Reinforcement

83

Chapter 5

Material Behaviour & Numerical Modelling

Chapter 5 Material Behaviour & Numerical Modelling

5.1

Introduction

In the analysisof reinforced concretestructures,the developmentof an adequate difficult behaviour for the the concrete of presents most challenge. mechanical model In orderto modelthe completeresponse of concrete and steel,a numberof non-linear important be The have to most of theseeffectsare: considered. effects

" " " "

tensile cracking yielding of the steel non-linear material behaviour crushing of the concrete aggregateinterlock bond between concrete and reinforcement dowel action of reinforcing bars

The fundamental requirement is to develop a set of constitutive laws which adequately describe the multidimensional stress-strain relationships within the reinforced

concrete. These constitutive laws are mathematical expressions which approximate the constituent material behaviour. These laws are basedupon experimental data. The complexity of concrete behaviour involving some of the phenomenon outlined above, has led to difficulty in being able to fully develop accurateconstitutive models. There is at present no universally accepted constitutive law which fully describes concrete behaviour in combined stressconditions (Buyukozturk et. al (1985)). However, much

84

Chapter 5

Material Behaviour & Numerical Modelling

development in has been the the this of years carried out over work on problem material models for cracked and uncracked concrete and subsequentlyseveral
numerical models exist, (Chen 1982).

5.2

Concrete Constituent Behaviour

Concretebeing a brittle material,there exists within the results of concretetests, a marked statistical scatter.Figures 5.2(a-b) shows examplesof this for Young's in in behaviour Variation the tension. of one of stress-strain concrete modulusand threetestvariables namely: canaccountfor this scatter,

9 materialstested
9 9 test method used loading systems

The presenceof this scattersuggests that a perfect match between analytical and The following sectionsdescribe experimentaldata is neitherpossiblenor necessary. the behaviourof concreteand steel as well as detailing the model adoptedfor the bending inplane in this study. analysis of plate and problems numerical

5.2.1 Uni-axial Compression


Uni-axial compressive strength is the most widely used variable for assessingconcrete quality. In the UK, the uniaxial cube strength f,,, is determined by testing 150mm. cube strengths after say 28 days. In the U. S, the uniaxial cylinder strengthf, "is derived from testing 152005 mm. cylinders. The cylinder strength is usually around 70-90% of the cube strength. This difference can be attributed to the friction forces which are generated between the contact face of the cube specimen and the testing machine forces Such platens. result in the formation of a multi-axial stress state and result in an increase in the cube compressive strength. These multi-axial effects are reduced in the cylinder specimensdue to the increasedwidth to height ratio of the specimens.

The study of concrete under uniaxial compressionprovides a good premise from detail its behaviour under more complex stressstates. to which

85

Chapter 5

Material Behaviour & Numerical Modelling

1 66
276 7 10

4681i0
-;, 0---

V-fl-cPC IM
40 f, ,. MPG

100

50

ACI
Range formula for which ACI was derived code

318, ECz 33W.15, /Tcr Psi


11 A. 40

1.5 5 145) ECWx 166


psi 4W51s. 'q 3 0% .... JIV -,. .&E,

A
30 (40.000 (w, /145), 0 . 1e) psi

145 E C( W
M PC

x 10-3

-20

2##,

0,6

It,
& 10 &

QD psi

1000 0 ?('0

2000

3000 4000 O - ---O-,

5000 6000 iio I psi

BODO 9,100

10000 12000 14000 - -- IS 120 10

130

-0

Ile-C

fig. 5.2(a) Young's modulus E, vs. cylinder compressive strength f, ', (w= concreteunit weight, ACI-Committee-363 1984)

1.0

0.8

0.6

0.4

0.2

LtL 00

28
E, (X 1()-3)

10

12

14

fig. 5.2(b) Test data for cracked concrete in tension, (fc, = principal tensile stress,fr= stressat cracking, Vecchio & Collins 1996)

86

5, Chapter

& Numerical Modelling MaterialBehaviour

The typical stress-strain relationship for concrete under multi-axial compression is

illustrated in figure 5.2.1(a). From the experimentalresults, the following basic observations canbe made: ', f, its 30% concrete stress-strain of maximum compressivestrength 9 up until
linear is relationship roughly ' in f, increase deformation is 30% to observed up stresses above gradual at ,a 0.75-0.9f, "t bending more sharply on approachto peak strength f, ' after reaching peak strength, the stress-straincurve has a descending branch until

). This strain is normally in the crushingfailure occursat the ultimate strain (ema,, rangeof around0.003to 0.004
Figure 5.2.1(b) details the variation of uniaxial compressive stress-strainbehaviour for concrete of varying compressive strength. From the figure, it can be seen that the initial modulus of elasticity is dependentupon the specific compressive strength of the concrete. It is observed that concrete behaves in an increasingly linear fashion up to peak strength the greater the compressive strength. All peak stress points occur at around 0.002 strain. A decreasein the ductile nature of unloading is observed in the descendingbranch of the stress-strainplot as the compressive strength is increased.

Numerous formulae derived from standard mathematical functions or from curve fitting techniques have been proposed to approximate the uni-axial compressive (1995). A (1970).

stress-strain response; Saenz (1964), Mansur et al. (1995), Almusallam review of the various proposed formulae

is can be found in Popovics

Examples of various numerical uni-axial compression models are presented in figs 5.2.1 (c-e). From this work the following observations can be stated:

41 The maximum compressive strain differs from one model to the next 9 The presenceof compressionsoftening can be accountedfor. e There is no unique model for the uni-axial compressive stress-strainresponsethat

hasbeenagreedby the majority of researchers.

87

Chapter 5

Material Behaviour & Numerical Modelling

1.0
Axial strain Lateral 't strain

ProportionahtV y limit 0.3


Extension c Comg)ression

fig 5.2.1 (a) Uni-axial compressive stress-strainresponseof concrete (Chen 1982)

f', - 67 N/MM2

0246a Strain. ClU

10

12

fig. 5.2.1(b) Compressive stress-straincurves for concrete with different f, ' (Chen 1982)

88

Chapter 5

Material Behaviour & Numerical Modelling

If"C2 mat 'r-2 fr-2

0.8 - 0.34 ( /(,,

l< 1.0 2 '2) (7_ a

C max

f 2 C2 n%mx c2

'2)

T_ _

EC2 1
EO E

principal stress c) Vecchio (1989),fc2= max compressive

&1 0 I-

c cu

d) Razaqpur & Nofal (1990)

2
CL

ul Ir

STRAIN

(z

10'3 )

e) Vecchio& Nieto (1991)


fig 5.2.1(c-e) Numerical Compressive Stress-Strain Curves for Concrete

89

Chapter 5

Material Behaviour & Numerical Modelling

2A5 Curve
No, 1 2 3 4 5 Tyoe Granite Gravel Gravel Gravel Gravel Aggre-qate Size 3/8- 3/16" 3,, 8-3/16" 3/16-B S. 7 3,8-3/16" 38-3/116-

Age (monthi) 2 2 3 2 1

1,75

1 os -

Tensile sirain M

fig 5.2.2(a) Uni-axial tensile stress-straincurves for concrete (Chen 1982)

Ge
'Ip

R. q. d

Pumm

Aa

C.
v
Indirect test

Direct

test

Flexural

test

fig 5.2.2(b) Test methods used for measuring the tensile strength of concrete

90

Chapter 5

Material Behaviour & Numerical Modelling

5.2.2 Uni-axial tension is Verylow, usuallyaround Thetensilestrength 0.1 of concrete The mainreason

for thislow strength fromtheheterogeneous itself.From results of theconcrete nature interface, the outset,concrete contains manymicro-cracks at the mortar-aggregate forcewill propagate Theeffective tensile whichon application of tensile andexpand.
despiteits importance in determiningthe strength of concretefi,is difficult to evaluate behaviour of a structure. Tensile strength is the' most important parameter in determiningthe crackingbehaviourand development of constitutive'models.Three testingmethods;direct, flexural and indirect areusedto derivethe tensile strengthof The indirect test, namely the cylinder splitting test, is most concretespecimens. commonlyusedin evaluatingft. In this test,the concrete cylinder is laid horizontally betweenthe loading platens of the testing machineand compressed until it splits vertically alongthe diametricalplane. The stress-strain testedin relationshipfor a numberof different concretespecimens
uni-axial tension is shown in figure 5.2.1(a). In general, the responseis almost linear

up to a high percentage of the tensilestrengthft.

5.2.3 Bi-axial Stress


The constitutive behaviour of concrete under bi-axial loading differs form that of uniaxial loading. Many people have researchedthe behaviour of concrete under bi-axial load including with reference to the effect on strength, micro-cracking, and deformational characteristics.

The biaxial failure envelopobtained form experimental work of Kupfer et.al (1969)is shown in figure 5.2.3(a). It can be seen that the maximum compressivestrength increases A maximum compressive with the level of bi-axial compression. strength increase of roughly' 16% f, " is observed under conditions of equal bi-axial compression, and an increaseof around25% f, ' is achievedwith a stressratio of initial stiffnessin bi-axial compression is also observed (Tj/a2= 0.5. An increased which may be due to Poisson'seffect, i.e. a reduction in lateral tensile strain, see figure 5.2.3(b).

91

Chapter 5

Material Behaviour & Numerical Modelling

In the case of bi-axial tension to uni-axial tension is tensile strength similar ,a achieved. The stress-straincurves for bi-axial and uni-axial compression are similar. For the tension-compression case (fig. 5.2.3(c)), the compressive strength decreases almost linearly as the tensile stressincreases.

5.2.4 Compression Softening


As shown in the uni-axial compression curves, once the peak stress is reached, the concrete begins to unload while the strain increases.The term strain softening is used to describe the response of a material where the slope of the stress-strain curve becomesnegative. There exists considerable variation in the experimentally obtained unloading branch of the stress-strain curve. Compression softening behaviour of concrete is dependentupon the boundary conditions and the size of the specimen, see (Van Mier (1984), Vonk (1992)). Post crushing behaviour of the concrete has a definite effect on the failure mode of the structure. Work carried out by Kent & Park (1971), has shown that the level of confinement plays an important role in post crushing behaviour. As expected, the greater the level of confinement, the more ductile the post crushing response,(Issa & Tobaa (1994)). As shown in figs. 5.2.1.(ce), various models for the descending portion of the stress-strain curve have been proposed. Meyer and Bathe (1982), suggesteda straight line approximation with a free parameterto account for confinement of steel was sufficient for most analysis. A concrete post-crushing residual strength of 10-40% of f, has been proposed by many researchers.

5.3

Cracking of Concrete

The low tensile strength of concrete will result in early cracking of members in a tension zone at low load levels before the steel starts yielding. Cracking is therefore, one of the most important non-linear phenomena displayed by concrete. Three main approaches to cracking have been developed in finite element analysis; discrete cracking models, smeared cracking models, andfracture mechanics models. Each of the three methods will be briefly described in the following:

92

Chapter 5

Material Behaviour & Numerical Modelling

fig 5.2.3(a) Bi-axial strength envelope of concrete (Kupfer et al. 1969)

61
C, pp

E3

I
C2, C3 Q(i IL

F-2

6; C) AG

6'j 167 0 1/-052_

2() 5cm(2in) (79 in) Ej, E. E3 +3 tensile strain +2 +1 0-1 -2 -3 mrrttm(QOOlirViN compressive strain

fig 5.2.3(b) Concrete bi-axial behaviour: compression -compression (Kupfer et al. 1969)

93

Chapter 5

Material Behaviour & Numerical Modellin. -

"I 12
E 2'E3 E3

10 09Q8 -

6; / 6 --1 0,052 --k -1/ 0.103 -1 / 0204 --

EJAE3

00 -

/ El - --

11- OL6 1t E OL5

+6; '92 -'5cm(21n) 20 (79in)

----

OL 4- - Ei3 03 -f N

---;

--t

-1.0 tensile strain

-05

-05

-10

-15

(()Winhn) mmim -20 Compressive stratri

fig 5.2.3(c) Concrete bi-axial behaviour: tension-cornpression(Kupfer et al. 1969)

5pp pp

295 kpicmz

4 200 psi)

010 -C3
008

E,

--6', 162
1/0

'1

+6 +6 -20 5 cm(20in) (79 In) C3 Cl-C2,

CN Ow

1/1 1/Q55

-QO4 comprvssive

-002 strain

+002

CID4

006

008

alo

OL12 m-kno tensite strain

fig 5.2.3(d) Concrete bi-axial behaviour: tension-tension (Kupfer et al. 1969)

94

Chapter 5

Material Behaviour & Numerical Mcdelling

5.3.1 Discrete Crack Model


The discretecrack model was introducedby Ngo and Scordelisin 1967and was the is first model usedto represent in finite In this the method, crack cracking elements. the nodesof adjoining elementsalong the length of the modelledby disconnecting crack (fig 5.3.1). The main problem with this method lies in the redefining of the is not very efficient in the finite elementmeshcontinuallyafter cracking.This process elementmethodwhich requiresa narrowbandwidth in the structuralstiffnessmatrix. Additionally, the crack topology, though not known in advance,is dictated by the meshsize and elementtopology Le the crack must propagate along elementedges. As a result of theseproblems,the discretecrackmodelhasfound limited usein finite elementanalysis.

5.3.2 Smeared Crack Model


A practical smeared crack model was first introducedby Rashid in 1968. Further by Suidan& Schnobrich(1973) and Phillips to this model were made enhancements & Zienkiewicz (1976). In this model, the concreteis idealised as a continuum, isotropicprior to cracking.On cracking,it is possibleto describeits behaviourfrom the stress-strainrelationship. Once the concrete has cracked, the stress-strain from isotropic to orthotropic.This method is relationshipof the elementis changed computationally convenient sincethe topologyof the elementremains unchanged,
When this approach was first introduced, an orthogonal crack system was employed. In this approach, after initial cracking, subsequentcracks were only allowed to form orthogonal to the existing crack. This fixed direction approach, meant that the crack direction was governed by the direction of the first principal stress to exceedft (fig 5.3.2b). The second crack occurs when the stress parallel to the first crack becomes greater than ft (fig 5.3.2c). The fixed crack approach has been implemented in many finite element codes and is the model implemented in this work In reality, it is possible for the post cracking principal strains to deviate from the crack

due to the presence orientation of shearstrain on a crackedplane.The principal post


cracking tensile stress may reach peak on a plane other than the initial crack plane.

95

Chapter 5

Material Behaviour & Numerical Modelling

The subsequent development of the rotating crack model allowed the co-rotationof the crackwith the principal strainaxes.

5.3.3 Fracture Mechanics Model


Fracture mechanics theory has been applied successfully in the past to cracking has been involving Some and carried out ceramics rocks. research problems metals, fracture Bazant & to theory the concrete; of mechanics reinforced on applicability Cedolin (1980). It was concluded by Chen(1982), that the use of fracture mechanics in reinforced concrete is still questionable.

The smearedcrack approachwas adopted for the present study. The simplicity of this approach has made it a popular choice with many analysts. The smeared crack

approach provides a good approximation of the load-displacement relationship but is unable in many casesto realistically model exact crack patterns. In caseswhere this is desired, a fracture mechanicsbasedmodel is more suitable.

5.4

Present Concrete Model

into the applicability of variousconcretemodelshas found that the Recentresearch level of complexityof the model is not necessarily linked to the level of accuracy. It has been found that simple modelscan be just as effective, or ineffective as more complex models,Collins & Vecchio (1985).The model usedin this work although predicting the non-linear behaviour of reinforced simple, is capableof adequately been It has from modelssuccessfully structures. adapted concrete usedby Bensalem (1993) and Abdel Kader (1993) to account for compressionsoftening and nonorthogonalsteel.

5.4.1 Yield Criterion


The yield criterion definesthe combination of stresses which will initiate plasticflow to the definition of an ideal failure criterion which modeled at anypoint. With regards the exactbehaviourof concrete that such underall conditions,Chen(1982)concluded a criterion would be so complexasto makeits usein numericalanalysisimpractical.

96

Chapter 5

Material Behaviour & Numerical Modelling

Crock Wldths Not to scale

Primary

Crack

Secondary Crock

fig 5.3 Cracking of Concrete

CRACK

LINKS

Simile crack

Double crack

Ci -ack and a-grgate pl interlock

fig 5.3.1 Discrete Crack Model

97

Chapter 5

Material Behaviour & Numerical Modelling

or

figs 5.3.2 (a-c) SmearedCrack Model

98

Chapter 5

Material Behaviour & Numerical Modelling

The bi-axial yield criterion used in this work is basedupon the experimental work of Kupfer et al (1969). The yield surfacesfor concrete under bi-axial stressare shown in figure (5.4.1). The octahedral shearstress,linearised in terms of the octahedral normal stressis used to fit theseyield surfacesin the following form:

=+b

cy,,,,

where the octahedral shear stress T,,,,is defined as :

2) 1/2 -\F2((T +(Y Toct 7T x2 y2+3, xy 3

(5.2)

and octahedral normal stress (5,,, t as :


ax +0 Cy =-3 Oct

(5.3)

The following describes the derivation of constants a and b where. f. " is the uni-axial cornpression strength, previously stated as 1.16 strength (ft1f. "). is the equivalent strength under bi-axial compression

and m is the ratio of uni-axial tension to compression

5.4.1.1
a) For uni-axial

Compression- Compression Yielding


compression: (7, =-f. ", (5y = Txy = 0.0,

52 'roct = f',,' and at 33 ,,, thus on substitution into equation (5.1) = -f

33

=a-

(5.4)

b) For bi-axial compression: cy,= cyy

T,y

0.0, =

99

Chapter 5

Material Behaviour & Numerical Modelling

-5 f. ' 1.16 TOCt and (Toct= -1.16 , = 33.

2f"

into (5.1) equation substitution after ,

1.16

33

a- LI 6(bf,

(5.5)

Solving for a and b, the bi-axial compression yield equation is given by:

Toct

cyoct

0.1714fe, J'C'

0.0 0.4143 :::::::

(5.6)

5.4.1.2

Tension- Compression
(Ty 111f. following the same procedure leads to:

In this case (7,

(I CF m) -F2 Oct Oct -+. ft (I + M)

r22

m =0.0 3 (1 + m)

(5.7)

5.4.1.3

Tension-Tension

For biaxial tension the following simple circular yield criterion is implemented:

2 CY2

0.0 1.0 = -

(5.8)

where cy, and cy,are the principal stresses.

100

Chapter 5

Material Behaviour & Numerical Modelling

07

Compression
01

Fracture surface defined by stress components

Initial
yield
surface C 0 E C1.

Subsequent load6ig surfaces

fig 5.4.1 Yield Surfaces of Concrete under Bi-axial Stress (Chen 1982)

101

Chapter 5

Material Behaviour& Numerical Modelling

5.5

Concrete Non-linear Behaviour

5.5.1 Compressive Stress-Strain Relationship


For the modelling of rion-linear stress-strain behaviour of concrete in the principal stress direction, the equation describing concrete in the elastic stage ((5=DF-) first for bi-axial behaviour by Liu et al (1972) is by Seanz (1964) modified and proposed is defined This as: equation applied.

BEcE (T=-A+ (I - v(x)(I + CE+D-C2)

(5.9)

where (x is the principal stress ratio

v is Poisson's ratio. A, B,C and D are

parametersdependentupon the shape of the stress-straincurve and are calculated as follows:

I.

At initial loading F,= 0.0, (T = 0.0

da dF2.
(I

Ec
- V(X)

and (T= 0.0

On reaching the peak stress ap at strain F-p,the slope of the stress strain curve

becomeszero. Therefore at:


dcy -=0.0 dc and (y = ap

From the four conditions described above, the unknown parametersA, B, C and D of equation 5.9 can be evaluated. Solving for these constants leads to the following equation:

G --

Ec F-

(1 - V(X) 1+('

1EE

1- va E,

9p

Fp

where

102

Chapter 5

Material Behaviour & Numerical Modelling

E, is the initial modulus of elasticity of concrete for uni-axial loading E, is the secant modulus of elasticity at the peak stress ((Yp/ep) (T and c are the stress and strain in biaxial load

This equation is used to model the stress-strain relationship for concrete under biaxial cornpression up to peak strain at which point the equation ceases to become valid due to softening.

Early changes in the stiffness of the concrete are accounted for by incrementally linearizing equation 5.10 during loading. This process is carried out by assuming intermediate surfaces shown in figure 5.4.1 similar to those proposed by Bell and Elms (1971) and Chen (1982). The first loading surface corresponds to the initial discontinuity in the stress-straindiagram. Subsequentloading surfaces are assumedto have the shape of the limiting yield surface. The intermediate surfaces are

representedby equation 5.1 but with an intermediate concrete strength

repIacing

/, ". Tile following equation was proposed by Johnarry (1979): the ultimate strength.

J'(. =fCo -A+E, j, ll Ei 1

wheref, = intermediate compressive strengthj,, = 0.5f. ", f=

tensile strength, E, =

concrete elastic modulus, Ej = instantaneous elastic modulus. The instantaneous elastic modulus is calculated up to peak strain F-P using equation 5.11. For strains above this value the following expression is used until the assumedcrushing strain of 0.0035 is reached: f, '/t i

If the principal cornpressive strain exceeds0.0035 or if the failure criteria is violated, then concrete is assurnedto be crushed.

103

Chapter 5

Material Behaviour & Numerical Modelling

5.5.2 Tensile Behaviour


The tensile behaviour of concrete, most notably cracking, accounts for the main is in The concrete model able to concrete structures. present of non-linearity source generatethe main features of concrete in tension; cracking in one or two directions, for interlock to transfer aggregate and tension stiffening. account shear

5.5.2.1

Single Cracking
or tension-cornpression, are

If the yield equations 5.7 or 5.8 in tension-tension

violated then the material is said to be cracked. The direction of the cracked is taken as normal to the major principal tensile stress direction. The stiffness perpendicular to the crack is assumed equal to zero when tension stiffening is neglected. The material parallel to the crack is still able to carry stress. In addition some shear force is transmitted along the rough surfaces of the crack. The material stiffness matrix of the concrete in the local coordinate system is given below:

Ec 000 00

00

PG_

The shear modulus G is reduced (0 !! P !, 1) to account for aggregate interlock. The Poisson effect is neglected since it is assumed that there is no interaction between the two principal directions once the concrete has cracked. The following process of defining the crack direction: a: The principal stresses are calculated according to the following: describes the

1,2

x+y 22

(T

Cy x_ y2 xy

b. The principal angle 0, with respect to the x-axis from:

104

Chapter 5

Material Behaviour & Numerical Modelling

tan 20

2,r xy CY (Y x- y

It direction. lead it the 45, Since to principal major about confusion may !0:,: c. - -45
is necessary to calculate the normal stress Gn associated with angle 0 from the following
2 0+(y (y --(T COS x n y

sin

20+2,

r xy sin 0 COO

d. It is then necessaryto compare the values of (71and (T-,with the normal stress c5,1
calculated from the above equation.

0+ is inclined 90' 0 hence to the the xthen crack angle (71 cyl, is at and cy,, =
axis

if cy,= cy-), then (71 is at 0+ 90' and hence the crack angle is inclined 0 to the xaxis.

5.5.2.2

Double Cracking

Subsequentcracking and changesin crack orientation are due to the presenceof shear These tension subsequentor secondarycracks in reality may stiffening. and retention interlock implies direction that be to the since aggregate original crack orthogonal not the primary crack direction does not coincide with the principal direction. Vecchio in found from Collins (1982), that crack on shear panels, changes experiments and in do take especially unequally reinforced panels. place, orientation In the case of previously uncracked concrete, a double crack will occur when both
local fi in The tensile the strength matrix exceed material stresses stiffness principal .

is for this case given as: coordinates

000

000 0 G_ _O

(5.17)

When tension stiffening is considered in the analysis, then the first two diagonal terms be [D'],, The the matrix may updated above accordingly. matrix material stiffness of

105

Chapter 5

Material Behaviour & Numerical Modelling

it is local (x', therefore necessaryto carry the terms coordinates y'), of is expressedin [TI: into transformation the matrix transformation coordinates via global out a

[T] T [D'],, [T]

(5.18)

c2s2

cs

and

[T]

S2 -2CS

C2 2CS

-CS 2c2 -S

(5.19)

S= C= sinO cosO, where

5.5.2.3

Tension Stiffening

Cracking in a reinforced concrete member usually occurs at discrete sections. Oil cracking, the concrete tensile strength at the crack reduces to zero. However, the is between the concrete cracks still able to carry tension and thus contribute uncracked to stiffness. This phenomenonis known as tension stiffening JUT5.5.2.3a). Modelling for is important the prediction of a structure's load-defori-nation tension stiffening of characteristicsin the post cracking stage.

Numerically, there are two ways in which to model the tension stiffening effect. The first method is to modify the tensile stress-strain curve for concrete. The second method involves modifying the stress-straincurve for steel. The first method is the most popular and was introduced by Scanlon and Murray (1974). In this model, a branch is added to the stress-straincurve after the cracking strain has strain softening been exceeded.In experiments on the tensile strength of plain concrete, the presence of a softening branch of the stressstrain curve has been confirmed, Reinhardt (1985), Gopalaratnam and Shah (1985). Thus, the inclusion of this tension softening behaviour in the model is closer to reality.

Experiments by Clark and Speirs (1979) on one-way spanning slabs with different have shown that the effect of tension stiffening decreaseswith increasing ratios steel

106

Chapter 5

Material Behaviour & Numerical Modelling

be ignored They that tension could if stiffening suggested steel strains and steel ratios. the steel ratio exceeds 1.5% or the steel strains exceed0.00 16.

For the model used in this study, the tension stiffening regime shown in fig 5.5.2.3(b) linear in Tension the post as a softening was modelled stiffening was adopted. cracking regime.

when

Ei < F-cr

then

(5 ==EiFi [ (C2FFi)] cr Ecr (C2 - 1)

if

Ccr!Ci:! -C2F,,

then

a=

Cl ft

if

C2F-j> C2F-,,

then

G=O

where cyand F-iare the local stressesand strains orthogonal to the crack, the cracking is The f, /Ec the the tensile strength of concrete. F-cr andft value of coefficients strain -.,: , CI and C2 can vary between the following range: 0.5 !CI!! 1.0 and 10.0 !-C2 20.0

5.5.2.4

Shear Retention

In a cracked structure, shear can be transmitted along the cracked interface by one of two mechanisms. The first mechanism, aggregate interlock, results from the uneven fracture surface and works in combination with friction to provide resistance along the fracture plane (fig. 5.5.2.4(a)). The second, dowel action of the bars, is caused by 5.5.2.4(b)). The resistance provided by these (fig. the crossing crack reinforcement effects can be quite significant in structures under high direct shear, where the strength may be dictated by behaviour along a single dominant plane or fracture zone. Both these mechanisms are governed by the width of the crack, i. e. as crack width increases, shear resistance decreases. Aggregate interlock has been found

experimentally to provide more shear resistance than dowel action, Millar & Jonson (1985). It is not possible to implement directly the above mentioned mechanisms when using the smeared crack approach. For this purpose, the reduction in shear modulus across

107

Chapter 5

Material Behaviour & Numerical Modelling

Cracking

concrete stress

steel stress

fig. 5.5.2.3(a) Stressdistribution in a cracked reinforced concrete bearn

CY

ft

cl

Fcr

C2 c,

fig. 5.5.2.3(b) Tension stiffening idealisation

108

Chapter 5

Material BehaviOUr& Numerical Modelling

crack width -T-

sheardisplaccnient

fig. 5.5.2.4(a) Aggregate interlock

fig. 5.5.2.4(b) Dowel action of reinforcing bar

1.0

F-cr

fig. 5.5.2.4(c) shearretention factor

109

Chapter 5

Material Behaviour & Numerical Modelling

the crack is defined by the shear retention factor P. A number of researchershave in decrease for factors the to shear resistance gradual account proposedshearretention factor introduced Cedolin Deipoli (1972) increases. a variable the and crack width as A fictitious linearly to the crack. similar model normal a strain with which reduced in decrease hyperbolic Mahaidi by Al (1979) the shear stiffness using a was proposed (fig. 5.5.2.4(c)). This model was implemented in the current work, with P defined as follows.
B (F-f/ F-cr) (5.20)

defined below: fictitious is the the to as F-f strain normal crack where

cf=

2 Ocr +Ey F- sin ,

Cos

Ocr +yy

sin OcrCosocr

0, ft/Ec, is the F-y the F-x, yy is and and are strain inplane strains F, critical cracking =

the angle of the crack to the x-axis, 0 !B<1.

5.6

Modelling of Reinforcement

The modelling C, of steel bar behaviour is less complicated than that of concrete since its behaviour is largely uni-axial due to the one-dimensional nature of reinforcing The in bars typical of characteristics stress-strain reinforcing are elements. shown fig. 5.6(a). The steel exhibits initial elastic behaviour, followed by a yield plateau in which further straining occurs without significant stress increase. After this point, hardening takes place in which some stress increase is observed with strain some stram. Finally, softening occurs as the nominal stress drops with continued straining until fracture occurs. A simple bi-finear representation is sufficient to model the behaviour of the steel. and this can be modified to take account of strain elasto-plastic hardening (fig 5.6(b)).

In the elastic regime, the incremental stress-strainrelationship is:

A(T= ESAc

110

Chapter 5

Material Behaviour & Numerical Modelling

On reaching the yield stress fy, the incremental stress relationship becomes:

Acy= E, (l - E, / (E, + H))Ac

There is Young's is hardening Es H the the the modulus of steel. parameter and where for types three of models reinforcing steel; smeared, embedded and used are main discrete model (figs 5.6(c-e))

5.6.1 Smeared Model Here the reinforcements to be spreadin a steel layer over the concrete are assumed wherethereis a largenumber element.Hence,this modelis widely usedin structures in bars This first plate such as and shell structures. model closely spaced was of devisedby Wegmuller (1974) and further adaptedby Cope & Rao (1977). In this layers is divided to the and the stress-strain in structure relationshipfor each model, layeris definedas:
Jul = [D, '] Ic) (5.22)

'] for The behaviour [D, the matrix material steel. is where of the steel layer is described in the local coordinate direction of the reinforcement and hence the bars can be orientated at any angle to the global axes (x, y). The constitutive relationship can then be transformed from local to global axes

5.6.2 Discrete Model


Here, a one dimensional bar element, representing the reinforcing bar, is superimposed on the parent concrete element by assuming that the bar is pin connected (two degreesof freedom) at nodal points. This model was first introduced by Ngo and Scordelis (1967). In addition to this, beam elements can be used in place of the bars to accommodate axial forces, shear forces and bending moments. Such idealisation may be necessaryin structures where very large bars are used and hence bending becomesa significant effect.

Chapter 5

Material Behaviour & Numerical Modelling

In this method, it is possible to model the steel-concrete interaction by means of linkage elements which can account for bond slip. This model also allows the bar stiffness and strains to be calculated exactly on the bar position. The main disadvantage of this approach is that the mesh geometry is restricted by the reinforcement positions i. e. bar elements must pass through element nodes and hence rnesh sizes can often become large which in turn results in greater computation tirne. To overcome this problem, El-Mezaini and Citipitoglu (199 1) developed a technique which allows the discrete reinforcement to be modelled independent of the rnesh In this method, the desired concrete mesh is set up independent of any geometry. ,71 reinforcement. Then, the edge nodes of the concrete elements are moved to the points of intersection of the reinforcing bars. This systemcan causedistortion of the concrete element and a correction technique is necessaryto avoid this. In addition, this model does not account for non-linear behaviour such as cracking and dowel action.

5.6.3 Embedded Model


The embedded model was developed to overcome the mesh dependencyproblems of the discrete model. This method was first developed by Phillips & Zienkiewicz (1976). Bars are treated as special line elements which are positioned or ernbedded within the concrete elernent boundaries (fig. 5.6(e)). In its original formulation, the bar had to be aligned to one of the local iso-parametric axes and hence the method was only applicable to orthogonal reinforcement. Full compatibility between the steel and the concrete is assumed.The line of the bar is defined using the same shape functions as the main concrete element and becauseof compatibility, the displacements of the bar are obtainable form the displacement field of the embedding concrete element. The stiffness contribution of the bar is assumed to be only in the longitudinal direction.

Further development of the embedded method was carried out by Ranibaran ( 199 1) to allow for inclined bars. Elwi & Hrudley (1989) and Phillips and Wu (1990) developed a rnethod for embedded curved reinforcement. The main advantage of this method is

that there are no limitations to representing the reinforcement layouts. In addition, the

112

Chapter 5

Material Behaviour & Numerical Modelling

900

800
700

600

CY
500 0 400

300

200

100 01i 0

50

100

150

200

strain

x1OA3

fIg. 5.6(a) Stress-strainRelationship of Typical Reinforcing Bars

fI,

fs

A
cy C,

A
Ey F-,

i) Bi-linear (hardening)

n) Elastic-Perfectly Plastic

fig. 5.6(b) Ideallsation of Steel Stress-Strain Relationship

113

Chapter 5

Material Behaviour & Numerical Modelling

Y
CYY CY, y

01

////

//////////// 01

o' 001

fig. 5.6(c) SmearedModel

fig. 5.6(d) Discrete Model

fig. 5.6(e) Embedded formulation

114

Chapter 5

Material Behaviour & Numerical

Modelling

independently. The bar be evaluated each element can stiffness contribution of formulation for inclined reinforcement proposed by Ranjbaran (1991) was

implemented in this work and is describedby the following

5.6.3.1

Embedded Element Geometry

At a typical point P in the reinforcement (fig 5.6.(e)), the strain in the concrete with is follows: X-Y to the as axes global respect

xx Y xy/2

Y xy/2 yy

(5.23)

The strain component of the reinforcement in its local x'-y' coords, i. e. along or from is its length, the simple transformation to obtained perpendicular described

previously:

F-'=

T Rr-R

(5.24)

denoted by: direction is R the cosines matrix of where

mi Ml

direction 1 (,yy/2). Expansion of equation the are cosines, and F-,, and in where = y (5.24) leads to:

I 2E +211mle, I- xx

2+ +M]F-,

lil2F,, y

+('IMI
12F -xY, 2+

+'I'

I)F-, y
+ M2E

III 111,FIyy
2- "y

+ (1, rn, + Ion,

)F,,

+ m, mc

yy

21,m, c xy

In the reinforcing bar, only longitudinal strain is considered, hence only the following equation is required:

115

Chapter 5

Material Behaviour& Numerical Modelling

xx'-

IIF-xx +21, mlexy

+m,

F-yy

(5.25)

The stain displacementrelationship for the element is expressedas follows:

aU

n [aN,

x- Lax xx=:
aU C yy nu

0]

u Vi

ay
aV

i=11
+

aNi I O, ay

vi

i
(5.26)

aU

Exy

2 ax

ay

[aN, 1 aN, U, 1n =: -1 ay , ax Vi 2

j=I

in which N, is the shape function of the element at node i and n is the number of nodes

in the concrete element. Substituting the above into equation (5.25) leads to

Exx'=

I B'uvI B vi

U,

[B'l

U' vi (5.27)

i=l

where

Bu =1 , +IIMI I- ax aNi

aNi

aNi ay aNi

2 Bv = 1,ml + ml ax ay

With the strain matrix of the reinforcement [B'l obtained, it's stiffness contribution to the element is defined as:

Kr

ij=f,,,

iT B DBj. dQ

116

Chapter 5

Material Behaviour & Numerical Modelling

fB

u. iT J, BVDBj AQ u 21

iT DBj AQ

fB

il , iT ful B DBj AQ vv

iT DBj AQ

(5.28)

The equivalent nodal forces can be evaluatedfrom

f, B iT ar, d 2

(5.29)

The element of volume (or in this case length) is evaluated in terms of the dimensionless natural coordinates , ij of the parent element. Dimensionless

coordinate r defines the position along the reinforcement in the parent element where (-I <r< I) The direction cosines can now be expressed in terms of r: .

ax-axar ax, ar ax'

lax m C ar '

ay

ay ar

ax, ar ax'

I ay C ar

(5.30)

[( C= where

ax

2+

( ay )II ar

2 froM 12 + M2

ar

Substituting (5.30) into (5.27) leads to :

B'u = (Cl

aN' ax aN'
ax

+ C',

ay aN'

)/Cl

B'v

(C2 =

+C3

ay

2 VC

(5.31)

where
2 )( ( ax C, =, ar C2 = ax )(ay) ar ar C3 = ( ay ar 2

The Jacobian of transformation can then be evaluated from: dx jrl = =C dr (5.32)

117

Chapter 5

Material Behaviour & Numerical Modelling

from; derivatives found be that the can partial noting relevant

ax ax a ax all
ar --+-a ar ail ar

=iII-+j

a
ar

21 -ar

all

ay

ay a

a7l

a+j22

aq (5.33) ar

J12 +ay = ar aTI ar ar - a ar

and hence:
2 ir 2 (1
)(L +j2 12 11

(
+2(JIIJ,,,
--22

) a-r ar

ar

+JI'IJ22)

21 + (j +j2 2)(an

1/2 21

ar

(5.34)

J12 etc. are the elementsof the Jacobian rnatrix as defined by: where the constantsJ11,

ax ay
IIj j 21 j 12 22 aTl aTI a ax a ay

(5.35)

Since it is assumedthat the reinforcements are distributed through the thickness of the element, the element volume can be calculated from:

dKI, =A dl =A stljrldr Sr

(5.36)

where A, is the cross sectional area of the reinforcement, t is the thickness of the parent element, and S, is the spacing of the reinforcements through the element. All the equations defined are evaluated in terms of the concrete or master element Th i, coordinates. It is hence necessaryto define the relationship between r, and 17. S relationship is expressedas follows:

[TI 0 [M] I Tl*

(5.37)

118

Chapter 5

Material Behaviour & Numerical Modelling

where (*) and ITI*) are the nodal coords on the reinforcement in the parent element defined Lagrangian function, interpolation the is M as a single variable and polynomial:

)(r ri, l ) (r r,, ) ri-I ... ... mi - (ri ) (ri )(ri ) (ri r,, ri+l ri-I - r, ... ... (r-rj) (r

(5.38)

Hence for a straight reinforcing element, equation (5.36) may be written as:

22 I

r(h

+I

(b + a

(5.39) r(Tlb - 'la + (11b+ Tla

22

and

a
ar all

2 I (Ilb =2 ar

= -(b

(5.40)
- 11a

in which

) and (h, 11b) 17, are the end point coordinates of the reinforcement in the All of the necessarycalculations can now be made to

5.6(e)). (fig element parent

The the to the the contribution of reinforcement stiffness concrete element. evaluate be follows: can summarised as process calculation

0A

gauss point

coordinate

rg on the reinforcing

element

is selected and

corresponding values of

, il, a/ar, aTI/ar are calculated from eqns 5.38-5.40.

These values are evaluated once in the pre-processingstageand stored.


0 For the current values of ,, 11g,evaluation of J, ax/ar, and ay/dr is carried out

from equations (5.35) and (5.33) IJ'I, [B r, and ffir are calculated from equations (5.34), (5.31) and (5.36) respectively.

119

Chapter 5

Material Behaviour & Numerical Modelling

Kij and Fr are calculated form equations (5.28) and (5.29) respectively.

e These are then addedto the corresponding values of the concrete element.

5.7

Applications

of Numerical Model

In this section the performance of the current numerical model is assessed through comparison with experimental results on various types of structures. It is important to test the accuracy of the model in predicting the behaviour of in-plane and plate

bending structures which will be used throughout this work. A series of structures involving different modes of failure was chosen. The main areas assessed were loaddeformation response, cracking behaviour, steel yielding, ultimate load and mode of failure.

A convergence force tolerance of 4% was set for the analysis and the maximum 50 75 for iterations and was set at plate bending and plane stress analysis number of respectively. A combined algorithm was employed whereby the stiffness matrix is updated every 2nd, 5th, 10 etc. iterations until convergence or collapse is reached. An increase in the maximum amount of iterations in plane stress structures is a result of the slow rate of redistribution of residual stresses in plane stress structures

accompanied by lower deformations compared with slabs. In addition to this, small load increments were applied only to highly non-linear phases while larger increments may require more iterations. It was shown by Abdel-Hafez ( 1986) that the effect of increment size on the resulting solution is not significant.

5.7.1 Simply Supported Slab tested by Hago


This slab (model number 3) was chosen from a series of tests carried out by Hago (1982) and was used to assesthe effect of mesh size, tension stiffening and shear retention in the numerical model. Details of the slab dimensions and reinforcement layouts are given in figures 5.7.1(a-c). The slab was simply supported on four sides and was 100 nirn thick. A design load of 210 kN was applied as four point loads as figure in 5.7.1 (a). Material properties are outlined below: shown

120

Chapter 5

Material Behaviour & Numerical Modelling

Concrete Properties

Steel Properties

E,=21500

N/mm 2
-, T--

E,=214000

2 N/mm
2

f,,,=44.2 N/mm 2 f,=3.4 N/mm

fy=460 N/mm

Table 5.7.1: Hago's Slab n. 3

For numerical analysis, a symmetrical quarter of the slab was analysed. Firstly, the

four different analysed using mesh sizes; 2x2,4x4,6x6 slab was

and W

elements. and

For comparison the load-displacement curves obtained experimentally

nurnerically are displayed in figure 5.7.1(d). It can be seen that a more ductile responseoccurred as the number of elements in the mesh increased. The minirnum 2x2 element mesh was used purely for comparison. In reality, differences in dictate density layouts for a minimum may of mesh reinforcement adequate 4A A of elements was able to model the responseto sufficient mesh representation. accuracy at a reasonablecost. As the mesh density increased, the computation time increased substantially e.g the analysis with the W slower than with 4x4 elements. element niesh was four times

Generally in flexural failure of slabs, cracking is initiated at around 20-30% of the

ultimate load P, For the numerical analysis, a load increment of 20% of the design load P,j was used in the first increment, subsequent increments of 0.05 P,j were largest iterations The (26) were recorded during the 3rd increment amount of applied. where the onset of cracking took place. The numerical load-displacernent response for each mesh size was less stiff during cracking than in the experimental model. The into tension the model would help to increase the numerical of stiffening inclusion stiffness during this stage.The next parametric study focused on the effect of tension in fig. 5.5.2.3(b), whilst varying the C2 parameter the using model shown stiffening dictates descending branch the of the tensile stress-strain curve. From figure which 5.7.1(e), it can be seen that the presenceof tension stiffening in the model improves the load-deflection responseat the service load level.

121

Chapter 5

Material Behaviour & NUincrical Modclling

The ultimate load achieved numerically although very close to the experimental, increasesslightly with the value of C2.

In order to study the effect of shear transfer in the numerical model, analysis was carried out using four different values of shear retention factor (B=0.0,0.4,0.7 and

1.0). These values correspond to the transition from smooth to very rough concrete crack interfaces. The load-displacement curves resulting from each of these models are shown in figure 5.7.1(f). When shear retention is ignored, the numerical ultimate load is reduced by around 10%. When the shear retention factor B ranges from 0.4 to 1.0, no significant change in ultimate load or load-displacement characteristics is

observed. The steel-strains of the bottom reinforcement at the centre of the slab are shown in fig. 5.71(g). The numerical results shown were obtained using tension factor B=0.4. C2=10 An shear retention at and adequate correlation with the stiffening The is obtained. numerical crack pattern of the slab (C2=10.0, result experimental B=0.4), is shown compared with the experimental in figures 5.7.1 (h-i).

122

Chapter 5

Material Behaviour & Numerical Modelling

615

970

615

545

""
107( 2160

a) plan view

""
545 2100

ci.
145 145 160

b) Bottom Steel (8mni bars)


160

32 75 77 L 01I- I' 300 155 155 11-I-I bid 61, 145


014

145

-1-

c.1.

c) Top Steel (8mm 0 bars)


160

322 11
77 Id 1
1401,

300

LI -1-

155
-1

c.1.

figs. 5.7.1 (a-c) Details of Hago's Model number 3 (all sizes in rnm)

123

Chapter 5

Material Behaviour & Numerical Modelling

1.4

1.2

m1 0.8

-2 0.6 M 2
CL CL

--*-01--06

'o 0.4

experimental 4 elements 16 elements 36 elements 64 elements

0.2

-0-

0
0 10 20 30 40 50 (mm) 60 70 80 90 central deflection

fig. 5.7. I (d) Slab n.3 tested by Hago-,effcct of mesh size


1.4

1.2

Co

0.8

0.6
CL CL

0.4

-0-

experimental ts --*-without (c1=0.5, c2=20.0) --0-ts 6 ts(c1=0.5, c2=15.0) --O-ts (cl=0.5, c2=10.0)

0.2

C 0 10 20 30 40 50 (mm) 60 70 80 90 central displacement

fig. 5.7. I (e) Slab n.3 testedby Hago; effect of tension stiffening

124

Chapter 5

Material Behaviour & Numerical Modelling

1.2

0.8

c
0.6 m a,
m 0.4 rL m

0 --0-06 --0-

expehniental B=O. 0 B=O. 4 B---0.7 B--1.0

0.2

OL 0

10

20

30

40

50 (mm)

60

70

80

90

central displacement

fig. 5.7. I (f) Slab n.3 tested by Hago; Effect of shearretention


1.4

1.2

.T0.8

0.6

CL cc

0.4

* 0.2 --0-

experimental present analysis

r_ 0 0.5 1 1.5 strain/yield 2 strain 2.5 3 3.5

fig. 5.7. I (g) Slab n.3 tested by Hago; bottom steel strains at centre

125

Chapter 5

Material Behaviour & Numerical

x
x

x
X

A- A-

I'lg.5.7. I (h) Hago's Slab n.3; Underside Crack Pattern (numerical & experimental)

\ .\

\\

'

fig. 5.7.I (i) Hago's Slab n.3; Topside Crack Pattern (nunicrical & experimental)

126

5 Chapter

& NumericalModelling MaterialBehaviour

5.7.2 Corner Supported Slab tested by McNeice


(914.4014A44.45mm) This modeltestedby McNeice(1967)comprised of a square The four slab was reinforcedwith an orthotropicmeshof comers. on slabsupported
0.85% reinforcing steel and was tested under a single point load at its centre. Details in 5.7.2(a-b). The in figures the material properties are given the given are model of table below;

Concrete Properties 2 N/mm Ec = 28600 fcu= 37.92 Ntmrnn ft = 2.75 N/mm

Steel Properties E = 200000 N/mm fy = 345 N/mm 2

v -- 0.5
vame baa: mciNeice biab

II

A symmetricalquarterof the slab was analysed using a 4A elementmesh with 10 In orderto simulatea column support,threenodeswere layersthroughthe thickness. itself the the the comer slab; node and the two nodeson each comersof pinned at For comparison,the model was analysedwith and with out tension edge. adjacent
stiffening (CI=0.5, C2=10.0, B=0.4). The load-displacement curves obtained in detailed figures 5.7.2(d-e). A are numerically and satisfactory experimentally

between is the experimentaland numericalresults with tension achieved correlation


load, 80% is Up the the tension to of ultimate model around with stiffening stiffening.

Almost is load 10% the than same ultimate without. achievedwith and stiffer around
in 16 kN 12% the tension numerical model which stiffening was around or without load. the than experimental greater

127

Chapter 5

Material Behaviour & Numerical Modelling

914.4

a) Plan view

a-.

c.1.

25.4crs

b) reinforcement

25.4 crs

c.1.

figs. 5.7.2(a-b). Details of slab testedby McNeice (all sizes in mm)

c.1.

-V

fig. 5.7.2(c) Numerical Crack Pattern at failure (underside)

128

5 Chapter

& NumericalModelling MaterialBehaviour

16000 14000 12000 z 0 10000 8000

rL 6000 4000 2000 0 02468 d1splacement(mm) 10 experimental --C]-- without ts --br-- present analysis *

H
12

fig. 5.7.2(d)SlabTestedby McNeice;load-displacement at centre

18000 16000 14000 12000 z

'0 10000 cc .2 8000


CL

'o

6000 4000 2000

experimental --D- vAthoutIts --br- present analysis

0 0123456789 displacement (mm) 10

fig. 5.7.2(e)SlabTestedby McNeice,load-displacement at centreof free edge

129

Chapter 5

Material Behaviour & Numerical Modelling

5.7.3 Slabs testedby Taylor et al.


from a series The following modelswerechosen of slabstestedby Taylor et al.(1966). The original test programconsisted slabswith of 10 two-way span,simply supported
Only bottom degrees of reinforcement. and arrangements reinforcement was varying 1982.5 All in were mm. square and were simply supported slab. slabs each provided 1830mm. For S6 S 1, S2, the to a spans of analysis, give present slabs along each edge

layouts for Details S8 the of reinforcement eachof theseslabsare chosen. and were
5.7.3(a-d) below in in figures (5.7.3). the table and material properties are given given

Slab

Thickness (mm)

Concrete Properties

Steel Properties

S1 S2 S6 S8

50.8 50.8 50.8 44.45

fcu= 35.0 N/mM2 fcu= 36.3 N/mm 2 fcu=35.3N/mm2 fcu= 37.9 N/mm
Table 5.7.3: Taylor's Slabs
7-

fy = 486.1 N/mm2 fy = 486.1 N/mm fy=497. ON/mm2 FY


m2 486.1 N/mm =

All reinforcementcomprisedof 5 mm diameterbars. These slabs were chosento differing layouts the the at of program modelling performance and orientation assess of reinforcement.
During testing, loads were applied by small hydraulic jacks placed at 16 uniformly 4x4 A size mesh of elements was chosen to model each slab. positions. spaced Tension stiffening with CI=0.5 and C2=10.0 was utilised in the analysis, and shear

load-displacement The factor B=0.4. relationshipfor eachslab is displayed retention betweenexperimentaland in figures 5.7.3(e-h).It can be seenthat a good agreement numericalmodelswasachieved.

130

5 Chapter

& NumericalModelling MaterialBehaviour

C.'.

76 crs

33 crs 76 crs I

5OCTS

305 crs

305 crs 164 crs 95 crs

30 crs I. C.

c.1. a) Model S1 90crs


40 crs

b) Model S2

c.1.

105crs

105crs

90 crs

40 crs

-.

ci.

1 C.

d) Model S6

c) Model S8

figs. 5.7.3(a-d) Slabs tested by Taylor et al; reinforcement details (all sizes in mm)

131

5 Chapter

MaterialBehaviour & NumericalModelling

140

120

100
80

60

40 * 20 I present analysis

--D- experiment

00 0

10

20

30

40

50

60

70

80

90

central deflection (mm)

by Taylor et al; load-displacement fig.5.7.3(e)SlabSI tested

140

120

100 z
0 80

cc 7a 0

60

40 --$-present 20 -0-experiment analysis

00
20 40 60 central deflection (mm) 80 100 120

fig. 5.7.3(f) Slab S2 testedby Taylor et al; load-displacement

132

5 Chapter

& NumericalModelling MaterialBehaviour

120

100

80

60

40

20

--D- present analysis --*-experiment

10

20

30

40

50

60

70

80

central displacement (mm)

fig.5.7.3(g)SlabS6 Testedby Taylor et al; load-displacement

140

120

100
z .9 V cc 0
V CL cc 60

80

-a 0

40 0 20 I present analysis

--C)- experiment

0c 0

10

20

30

40

50

60

70

80

central deflection (mm)

fig. 5.7.3(h) Slab S8 testedby Taylor et al; load-displacement

133

5 Chapter

MaterialBehaviour & NumericalModelling

5.7.4 Deep Beams tested by Khaskheli


Four, two-span,continuousdeepbeamswerechosenfrom the experimental program by Khaskheli (1989). The out experimentalwork was mainly aimed at carried determining the ultimate load andserviceability performance of deepbeamsdesigned
using the direct design method. Since these beams had varying reinforcement layouts basis they ratios, provided a good span-depth on which to assessthe applicability and in the numerical model predicting the behaviour of deep beams at service current of and ultimate loads.

The first threebeamsin the serieshad the samespan-depth ratios of 1.07while beam TGRAS4 had an increased value of 1.61. Both TGRASI andTGRAS2 had the same
Beam TGRAS had I of main reinforcement. amount a greater amount of shear TGRAS2 hence failure than the and mode would be expected to be reinforcement more ductile. The shearreinforcement in TGRAS3 was the same as that for TGRAS2

but with TGRAS3 having a greateramount of main reinforcement.Details of the material propertiesand design load for each beam are given in table 5.7.4. The
layouts dimensions in figures 5.7.4 (a-d). For numerical are and given reinforcement

beam half discretised of each symmetrical was a analysis, using an 80 elementmesh. TGRAS1 19300 63.0 3.2 1.07 0.42 810

ConcreteProperties E. (N/mm
2)

TGRAS2 23200 61.0' 3.7 1.07 0.42 810

TGRAS3 20800 61.0 3.4 1.07 0.42 1100 J

TGRAS4 19200 52.0 2.6 1.61 0.69 - 810

f,,, (N/mm2) ft (N/mM2) Ratio Span/Depth Ratio Shear-Span/Depth DesignLoad Pd(kN) SteelProperties

N/mm2 6mm 0 bars,fy=513N/mm2,Es=199000 N/mm2 8mm 0 bars,fy=520N/rnrn,Es=195000


i anie . 5.7.4: Irroperties of Khaskheli's Beams

134

5 Chapter

& NumericalModelling MaterialBehaviour

In the caseof TGRAS 1, during the experiment, initial cracking occurred at the bottom first PdIn 0.3 beam the was the numerical analysis, cracking at mid-span at of in Yielding 0.4 Pd the the of the main area as experiment. same around encounteredat longitudinal steel was recorded experimentally at the mid-span at 1.3Pd.Numerical in I In this the the area at around steel was recorded experiment, main of yielding -OPdfor the this true the and stirrups was observed was also vertical no yielding of in for TGRAS I The and numerical result are given experimental numerical analysis. figures 5.7.4(e-f). From the load-displacementcurve, the onset of significant cracking is in 1.05 PdSuch be the numerical result at around a phenomenon seen can loss by The P=I. OPd the crack at sudden of stiffness. numerical patterns characterised in 5.7.4g-h) (fig 1.05Pd the shear span the of cracking confirm extensive spread and during this short load increment. This event correspondswith the opening of inclined beam during in load level the the the of shear span experiment at a of cracks shear LIP& Good correlation between the experimental and numerical crack pattern can

be seenfrom fig. 5.7.4(i). The ultimate load obtained from the numerical analysis was % 91 of that obtained experimentally. around

The effect of the shear retention factor on the numerical behaviour was also investigatedfor this beam.Figure 5.7.4(k) displaysthe load-displacement response 1.0. When B=0.1, a low ultimate load, B=0.1,0.4,0.7 and of obtainedusing values in load is This the to the cracks shear opened where original analysis, obtained. equal
1.05Pd the that cracks opening at were causedby high shear stresses.A more suggests higher the obtained using was values of B. However, no significant response accurate differences could be seenfrom the responseat B=0.4,0.7, or 1.0.

behaviour of Due to the reduction in shear reinforcement,the load-displacement beams TGRAS2 and TGRAS3 is less ductile (figs 5.7.4(1-m)).Experimentally, TGRAS2 achieved a smallerultimate load (I 216kN) than TGRAS1. For TGRAS2 the load at which inclined cracksopenedin the shearspanwas around0.925 Pd which for TGRAS 1. The lower that than numericalfailure of the beamwas reachedat was 1.375Pdwhich corresponded well with experimentalfailure load of around 1.43 PdThe onsetof inclined cracksin the shearspanof beamTGRAS3 occurredat around

135

5 Chapter

& NumericalModelling MaterialBehaviour

0.8Pd.A numerical ultimate load of 1.275Pdwas achieved which again corresponded load 1.36Pdthe of ultimate experimental well with

higher beam TGRAS4 than that of the other Sincethe span-depth was much ratio of beams,less brittle load-displacement was anticipated.In the experiment, response initial cracking occurredaroundthe lower soffit of the beam at mid span.The first first 0.12Pd in the this numerically, cracksappeared at while area crackwas recorded in the samearea at around0.3Pd.This differenceis due to the fact that the initial increment in the numerical analysis was 0.3Pd. Further increasesin load were loading In by the the towards this of crack point. propagation accompanied initiated internal beam the the top the at above support was of experiment crackingat 0.65Pdand this was recordednumerically at 0.7Pd. The sudden appearance of diagonalcracking in the shearspansof the beam was, as for the previous models, 0.7 0.75Pd between (figs 5.7.4(p-q)). The same and observed numerically failed 1.05Pd. Beam TGRAS4 at around was observedexperimentally phenomenon however, is ductile in It that clear greater was attained shear. response experimentally from the load-displacement and load-strain plots (figs 5.7.4(n-o)). Significant in (up to the the the was observed mid-span model at numerical steel main of yielding 8 times yield strain),howeverthe straingaugein the experimentfailed oncethe strain beams, As in 3.5 the times previous strain. no yielding of the vertical yield surpassed The failure experimentally or numerically. occurred of the numerical reinforcement beam was reachedat 1.15Pd which corresponded reasonablywith the experimental failure at around1.34Pd-

136

Chapter 5

Material Behaviour & Numerical Modelling

425 900

150 +

425

425
bl-

150
-lid

425

26

600

26

300

26 26 28

15Q 35 0.

YOU

0-1-d

YOU

N1

detailsfor TGRASI fig. 5.7.4(a)Dimensions andreinforcement

425

150 A J.

425

425

150 +

425 .1

900 26

600

26

300 IS 35 0
7UJ

26 26 28

you

fig. 5.7.4(b) Dimensions and reinforcement details for TGRAS2

137

5 Chapter

& NumericalModelling MaterialBehaviour

425

150 I J.
TT

425
ol 4

425

150
61bi-

425
61

900 860 26

600

26

30Q 15q 35

26 28 28

igt

YOU

boDu

NI-1

N1

detailsfor TGRAS3 fig. 5.7.4(c)Dimensions andreinforcement


675 900 850 26 700 soq 26 26 26 28 28 150 675

325 17 35. 01

fig. 5.7.4(d) Dimensions and reinforcement details for TGRAS4

138

Chapter 5

Material Behaviour & Numerical Modelling

1.6 1.4 1.2 10 10 LM . 0.8


2D 0.6
CL. m to

II

II
II

Ii I.

0.4 0.2 0 0 0.5 1 1.5 2 2.5

--*- expedmental --D- presentanalysis

3.5

4.5

central displacement (mm)

fig. 5.7.4(e) Beam TGRAS 1: Load-displacement relationship

1.6

1.4

1.2

1,:
0.8 cc 0 0.6 CL cc 0.4 0.2

?"+

0 expedmental anlaysis --13- present

0.5

1.5

2 strain/yield strain

2.5

3.5

fig. 5.7.4(f) Beam TGRAS 1: Main steel strains at mid-span

139

5 Chapter

MaterialBehaviour & NumericalModelling

_____

_____

_______

_______

______

______

II

I
___

-i
___ ___

I /I

I I
_ _

II / // 1I II

\
___ -

I I

I I

,, II

II II

I\ I

fig. 5.7.4(g)BeamTGRAS1: NumericalCrackPattem(P: I --

-OPd)

/
1
-

I I \ ll \I

1 / I

I S
__ _

\ 1

1 1 / \ \ 1 1 \ \

I / /I /I
-

\ 1

I \ \ 1 1

1 \ 1 1 1
\ \

' /
-

II II lI

I \ \ \ \ \ \ \ \ 1 \ \ \I \

I / I I

\ \ \ l

/ I I

/ I ,

II
II II

\
I I

\
\

fig. 5.7.4(h) Beam TGRAS 1: Numerical Crack Pattem (P= I

-05Pd)

140

5 Chapter

& NumericalModelling MaterialBehaviour

-.

/ I I \ \ \ \ \ \ \ I \% \I I \ \I

\
-

\ \ \ \ \

/I /
-

\ \ \ '

\ \ \

/I /I II /

/ /
_ /

/ /

/ I / I / / I
i
I

I / _

/ / / / / I I

/I /I II II II II II I I \ I \ \
\

\ \ \ \ \ I

\ \ \ \ \ I

\ \ \ \
\

\I \

I I

/ I
I

\ \ \

\ I \

fig. 5.7.4(i) Beam TGRAS 1: Numerical Crack Pattern at Ultimate Load

fig 5.7.40) Beam TGRAS 1: Experimental Crack Pattern at Ultimate Load

141

5 Chapter

& NumericalModelling MaterialBehaviour

1.6

1.4

1.2

M
II 'rI

It
0.8 V .20.6 CL
0. cc

Ii II

experimental B=O. l
6 B=0.4 --0- B=0.7 A B=1.0

0.4 0.2

0.5

1.5

2.5

3.5

4.5

central displacement (mm)

fig. 5.7.4(k)BeamTGRASI; Effect of ShearRetentionFactor

142

5 Chapter

& NumericalModelling MaterialBehaviour

1.4

1.2

c cm ,a0.8
ID V i cc 4) CL CL

II

.20.6
ca 0.4

Ii I Ii I.

' 'I,
i.

I --+0.2

expedmental --U- present analysis

OL. J 0

0.5

1.5

2.5

central d1splacement(mm)

fig. 5.7.4(l) Beam TGRAS 2: Load-displacement relationship

1.4

1.2

M1 w 0 c

0.8

II .20.6 V CL0.4 w experimental --Cl-- present analysis 0

cc

II II Ii I.

0.2

OC
0.5 1 1.5 central displacement (mm) 2 2.5 3

fig. 5.7.4(m) Beam TGRAS 3: Load-disiplacernentrelationship

143

5 Chapter

& NumericalModelling MaterialBehaviour

1.4

1.2

a m

"I
C)

0.8

II

Ii II

0.6
0.

M 0.4

0.2

-+-expehmental --G- present analysis'

0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 central displacement (mm)

fig.5.7.4(n)BeamTGRAS 4: Load-displacement relationship

1.2

o 0.8

0.6

ta
rL

L. .10.4 w 0.2 expedmental --D- present analysis 0

00 0

123456 strain/yield strain

fig. 5.7.4(o) Beam TGRAS 4: Main steel strains at mid-span

144

5 Chapter

& NumericalModelling MaterialBehaviour

II I
___________ ______ ______ ___________ ___________

II I
___ ___ _ ___ ___

II
_______

U
II I

II II II

____

_______

n
_ ___ ___

___

__

I
_______

%I %I II II II II

L'
II I
-

I I / I I I I

t II II II
_______ _______ -

fig. 5.7.4(p) Beam TGRAS4: Numerical Crack Pattem (P=0.7Pd)

1
1 %
wo

11
1%

11
11

11

11

,%

fig. 5.7.4(q)BeamTGRAS4:NumericalCrackPattem(P=0.725Pd)

145

5 Chapter

& NumericalModelling MaterialBehaviour

5.7.5 Corbel Tested by Niedenhoff


In this examplethe analysisof a one-sided corbel is carried out. The experimental to two sidedcorbels testingof one-sided corbelshasbeenrelatively scarcecompared
(e.g. Kriz & Raths 1965). This is perhaps due to greater practical difficulties

corbel resultingfrom the natureof its eccentric with testingthe one-sided associated
loading. The example chosen for this analysis was part of a series of tests carried out

by Niedenhoff (1963). The original aim of these experimentswas to obtain an


indication of the stress distribution in these types of structures. A total of 12 corbels in for M2/132 the tested the present numerical program and model was chosen were analysis. Numerical analysis of M2/B2 has also been carried out previously by van

Mier (1987) andPrasad et. al (1993).The resultsobtainedby the presentanalysisare by Details the those above. obtained with of the geometry,boundary also compared layout in figure 5.7.5(a). The material are presented and reinforcing conditions below in (5.7.5): table are shown properties

Concrete Properties E,: = 27000 N/mm 2 ft = 2.42 N/mm 2 fcu= 22.6 N/mm 2 P=0.2 v=0.2

Steel Properties Es= 2 10000 N/mm2 fy = 350 N/mm2 (70) fy = 282 N/mm2 (120) fy = 300 N/mm 2 (140) v=0.25

Table (5.7.5) Niedenhoff

Corbel

As shown from the structural system (fig 5.7.5(b)), only the corbel was loaded. Horizontal reactions were induced as a result of this loading system. During the test,

the load was applied in incrementsof 50 kN. In the original test, no displacements were recorded.A photo-elasticanalysiswas used to determinethe principal stress flows and was comparedwith the crack pattern and behaviourof the corbel. It was
found that the resultant of the principal compressive stress flows followed a diagonal loading from lower inner comer of the corbel. The occurrence of the to the point path

to this diagonal were also observed.Very little stress tensile stresses perpendicular in the lower outsidecomerof the corbel. developed

146

Chapter 5

Material Behaviour & Numerical Modelling

The first crack (numberl in figure 5.7.5(h) occurred at the second load step (P=IOOkN). This crack then developed into a vertical crack at a load level of l50kN. At P=300kN, crack number I was fully developed and due to changesin the load path, in in During 2 the this time, column cracking appeared crack number occurred. load level Failure in further the the to of corbel occurred at a corbel. cracking addition I began kN to 585 the widened. yield and crack number main reinforcement when of Also at this stage,the compression zone at the lower column-corbel junction began to 3). (crack number crush Figure 5.7.5(c) comparesthe load displacementrelationship obtained from the present Prasad (1993). by Mier (1987) In the case of that and et al. van obtained analysis with in be in Mier's to was assumed analysis, concrete elasto-plastic the model used van In Mohr-Coulomb criterion was employed. addition, a smearedcrack and compression implemented. was model It can be seen that that a reasonable agreement between

failure in Numerical the present analysis occurred at the achieved. was models each of 485 kN which is around 83% of the experimental load. As a result of the geometry of the corbel, compressive stress concentrations occurred around the lower column 5.7.5(d-e) in junction. Figures the the corbel obtained show concrete stresses corbel from the present analysis. Observation of the concrete stressesat the lower columnillustrates In junction that this the experiment, occurred crushing at point. corbel failure load. be from fig It towards this the at same point can occurred seen crushing 5.7.5(f) that yielding of the main steel in the corbel occurred at the upper in junction. This. the experiment. phenomenon was also recorded column/corbel

147

5 Chapter

& NumericalModelling MaterialBehaviour

1000

50

800

50

1000

600

500 details fig. 5.7.5(a)Corbeldimensions andreinforcement

fig. 5.7.5(b)Test set-up


148

Chapter 5

Material Behaviour & Numerical Modelling

Soo 450 400 350 z 300 C V 250 V 200 CL 150 100 50 0 0 0.5 1 1.5 displacement (mm) 2 2.5 3

present anlaysis -Cl-- van Mier 6 Prasad et al

fig. 5.7.5(c)NiedenhoffCorbelM2/B2 Load-Displacement

1 0.9 0.8 m .20.7 ,30.6 0.5 0.4 CL E 0 0.3 u 0.2 0.1 0 0.0012+00 1.OOE+02 2.0012+02 3.OOE+02 4.OOE+02 5.OOE+02 applied load (kN)
-"---

A B

-D-

A B

fig5.7.5(d)NiedenhoffCorbelM21B2Concrete Stresses

149

5 Chapter

& NumericalModelling MaterialBehaviour

0.8 0.7 0.6 0.5


(0

DV 0.4

2 0.0.3 0.2 0.1 0 O.OOE+00

1.OOE+02

2. OOE+02

3. OOE+02

4. OOE+02

5. OOE+02

appliedload(kN)

Stresses fig. 5.7.5(e)NiedenhoffCorbelConcrete


500 450 400 350 300 250 4) '& 200 150 100 50

I--o0.5 1 1.5

2.5

stralWeld strain

fig. 5.7.5(f)NiedenhoffCorbelMain SteelStrains

150

Chapter 5

Material Behaviour & Numerical Modelling

Jr )c

g)

h)

figs. 5.7.5(g-h) Niedenhoff Corbel M2/B2: Comparison between experimental and numerical crack pattern

151

Chapter 5

Material Behaviour & Numerical Modelling

5.7.6 Frame Tested by Stroband & Kolpa


This model was part of an investigation by Storband & Kolpa (1983) studying in

in frames. Likewise, beam-column behaviour a the connections portal of particular in this carried out orderto testthe numericalmodel's was structure numericalstudyof
behaviour. Previous beam-column in connection numerical predicting adequacy (1991). by (1987) Mier Ranjbaran this out van and carried was model analysis of

Work carried out by a number of researchers e.g Swann (1970), has shown that
joints in frame the the to comer strength of portal common assumption, contrary The frame less is than that the connecting members. of present model, often structures

A7, was tested under a negative or closing moment. Moment in the comer was in loads beam. The the thirds the through middle of application of point generated details in fig frame A7 dimensions together with reinforcement are given of overall in the tablebelow: 5.7.6(a). Material properties arepresented

ConcreteProperties N/mn 28000 Ec, = fcu= 26 N/mm2 ft = 2.1 N/mm2 v=0.2


Table 5.7.6: Stroband

SteelProperties E. =2 10000N/mm2 fy = 450 N/mm2 1v=0.2

& Kolpa Frame

For numerical analysis, a symmetrical half of the fame was modelled. At the comer of the frame where stress concentration occurs, the mesh was refined, elsewhere in the beams a coarser mesh was adopted. The curved reinforcement in the and columns dividing into by the three equal lengths joined by successive arc modelled was comer

bars, each turning 30' in order to give an approximateradius. Van Mier (1987) increased f,,, (37.5 N/mm2) for the elements this an using value model of analysed directly surroundingthe inner comer of the frame. It was believed that a threedimensional stress state would develop at the inner comer and hence a larger
compressive strength in this region would result. For the present analysis, this was

ignored since, although tri-axial stress may occur in very wide frames such as

152

5 Chapter

& NumericalModelling MaterialBehaviour

hence frame 70 thick the generation of the tanks was only mm. and etc, retaining walls, tri-axial stresses was believed to be negligible.

In the present analysis the frame was analysedusing displacement control (chapter 4) load-displacement be to the of relationship obtained. part which allowed unloading The experimental load-displacement curve together with those obtained numerically in the present analysis and by van Mier are shown in fig 5.7.6(b). It can be seenthat a between the experimental and the numerical response was achieved. good agreement During the analysis, first cracking was observed at a total load Pt=4kN. A significant increase in cracking was observed as Pt increased from 16-18 kN. This cracking in The the the comer. stresses principal stress plots (fig of caused redistribution 5.7.6(f-g)) during this stage show that increasedcompressive stressesform along the diagonal from the inner to outer edge. In the experiment, tensile stresses

bend Significant the to the cause at splitting. reinforcement compressive perpendicular inner in beam junction be the the comer at observed column can stress concentration from the stress plots. The concrete stresses around this point are shown in figure 5.7.6(c). The stresseswere calculated at the Gausspoints (gp), with each gp numbered be It bottom. from inner that the the the seen crushing can of concrete at clockwise load level bi-axial is 26kN. This in stress, occurs at a of a state of event comer, in face by the the steel outer of of the column near the comer at yielding accompanied 5.7.6(e)). All (fig 28kN by load these events were preceded of yielding of the main a tension steel at mid-span at a load level of 24kN (fig 5.7.6(d)).

153

5 Chapter

MaterialBehaviour & NumericalModelling

1620 14@30 crs 10@50crs r* Ai


a 11 11 1 11111111111111111 11 111i

2@ 2(6)

A
C.'.

120 5@ 104 2(6) 70 sectionA-A c C J 170 2@ 20 95 120 Luo160 sectionC-C covcr on stirrups5mm stirrups2.8 120

870

fig. 5.7.6(a) Frame testedby Stroband and Kolpa

154

5 Chapter

& NumericalModelling MaterialBehaviour

30

25

20
13 cc 2 15
C)

'1,1

10

experimental

--D- presentanalysis 6 van Mer

02468 centmi displaceffat of beam (nwn)

10

12

by Stroband & Kolpa; Load-displacement fig. 5.7.6(b)FrameA7 tested relationship

1.2

B 1 A
C 0.6
W CL

0.8
a, C)
I-

B00 A

r: 0.4 0 0 0.2

-D- B

0. 02468 central displacement of beam (mm) 10

-. -I 12

fig. 5.7.6(c) Frame A7 testedby Stroband & Kolpa; Concrete stressesat inner comer

155

Chapter 5

Material Behaviour & Numerical Modelling

30

25

-20 z he
.

*a
15

10

04
0

iiiiii
0.5 1 1.5 2 25 3 3.5 straiWeld strain

fig. 5.7.6(d) Frame A7 testedby Stroband & Kolpa: Main steel in beam at mid-span

30

25

-20 z .19 V cc .2 is

10 lo

0.2

0.4

0.6

0.8

1.2

1.4

1.6

1.8

strain/yield strain

fig. 5.7.6(e)FrameA7 testedby Stroband & Kolpa. Steelin outerface of column

156

Chapter 5

Material Behaviour & Numerical

Modelling

It o
0 00 VI * 0 '410 o.

Compression IC11"Imi

'4

..

%:

-J

fig. 5.7.6(f) Principal Stressesin Corner (P= I 6kN)

.r.

9iI

*'

fig. 5.7.6(g) Principal stressesin Corner (P= I 8kN)

157

5 Chapter

& NumericalModelling MaterialBehaviour

5.8

Conclusions

From the work carried out so far, it can be seenthat presentnumerical model is bending In for the examples the and plane stress structures. of plate analysis suitable behaviour in level discussed, the prediction of at service and accuracy of a good loads wasobtained. ultimate
In the caseof plate bending problems, it was found that a minimum 4x4 element mesh It the symmetrical modelling accuracy when quarters. was also necessary provided found that tension stiffening in the case of slabs improved the predictions of behaviour at service loads. However at loads approaching ultimate, the predictions are often stiffer. For the subsequent numerical analysis, a value of 0.5 and 10.0 was CI C2 for and respectively. constants adopted

The shear retention factor B had little influence on slabs and other flexural structures. However, in the case of shear transfer members such as deep beams and corbels, low in load. For B, under-prediction <0.1, of will result ultimate subsequent of values for B. 0.4 adopted was a value of analysis,

158

6 Chapter

SlabDesign

Chapter 6 Slab Design

6.1

Introduction

The main objectiveof the work in this chapterwas to assess the applicability of the in deriving layouts for suitable reinforcement process various types of visualisation layouts is The the reinforcing of assessedin terms of the suitability slabs. demand ductility behaviour of the slab. and serviceability This chapterdetails the designand numerical analysisof 7 different types of slabs. Two slabswere simply supported, one was simply supportedwith a central column four fourth type the on supported comersand the last type was simply was support, comerswith a columnsupportat the oppositecorner.Details of on adjacent supported the geometryand supportconditionsare given in table 6.1(a). Loading arrangements, designload Pdand materialpropertiesare given later. The thicknessof eachslab was in compliance with the limiting span-depth ratios stipulatedin BS8110Part 1. chosen
The direct design approach, described in chapter 3, was used to derive the layouts. For SMI-5, design derived from the required numerical slabs a reinforcement steel areasat (rr--O) and using the mesh evolved at the subsequent(rr) was made. The effect of using the evolved mesh on the, reinforcement layout was investigated. The designs the two of was assessedand compared in numerical analysis. performance Since direct design for the slabs from the evolved mesh involves a redistribution of the stressesfrom the elastic pattem,,it was necessaryto compare the two designs and determine the effect of this re-distribution on service behaviour. The ultimate load behaviour was assessed together with service deflections and steel strains. The mid-

159

6 Chapter

SlabDesign

span service deflection limit was taken as span/250 (BS 8110). As a means of further

yield strain of the steelwas recorded.In order serviceabilitycharacteristics, assessing to avoid large crackingstrains,a minimum load of 0.7 Pd should be reachedbefore
A the minimum steel ratio of 0.13% as stipulated in main steel occurs. of yielding BS81 10 was also adopted. In the resulting steel layouts, anchorage lengths were ignored.

Slab Model

Schematic

Support conditions

Dimensions (mm)

SMI

Square simply supported 2140 x 2140 x 100

N SM2 Rectangular simply supported N SM3 E3 777773140 x 2140 x 100

Square simply supported 2140 x 2140x 100 + centralcolumn Square simply supported 2l4Ox2l4OxlOO on adjacent sides+ columnsupportat oppositecomer Square supported on four comers 910 x 910 x 45

SM4 -77-p SM5,6 &7 0,1%

fable 6.1(a): Slab Details

6.2

Effect of Mesh Size on Visualisation

Sincethe direct designof the reinforcement is dependent upon the stressdistribution, it was necessary to assess the effect of using different mesh sizes upon the stress distribution and direct designprocess. To facilitate this study, three slabs with very

160

6 Chapter

SlabDesign

different moment distributions were chosenfrom the program; SM 1, SM3 and SM5. It

is necessary that the mesh density is fine enough to detail adequatelythe stress however density the the as mesh slab, increases,the computational variation within time increases.

The visualisation process was carried out on each slab using a symmetrical quarter

discretisedinto a meshof R5,70

and lOxlO elements.The ultimate rejectionratio

for each mesh size was comparedalong with the correspondingnumerical steel
from in The the table 6.2. The principal comparative are results study given volumes.

illustration in for vector form, for each mesh size are detailed in moments,plotted figures 6.2.2-6.2.4.In the principal moment plots, black lines indicate negative lines represent positivemoments. while grey moments, Slab SMI simply supported 5X5 70 SM3 SM5 simply supported supported on four + centralcolumn comers IOXIO 5x5 70 IOXIO 5x5 70 IOXIO

Meshsize

30% 30% 30% 30% 15% 12% 10% 40% 40% Rejection ratio (rr) 16% 24% 12% 48% 53% 66% 48% 57% 70% % area 'removed' Vol. of steel 3 1148 1151 1112 473.7 481.1 431.8 50.2 53.5 1175.9 cm I calculated 11 1 1 1 1 1
Table 6.2 Effect of Mesh Size

Firstly, it can'be observed that the rejection ratio needed to achieve a certain % area

is different for each slab. When comparingslabs SMI and SM5, for any &removal' chosenmesh size, the ultimate rejection ratio decreases while the percentagearea is due fact increases. This to the that the stresses removed are much more evenly distributedthroughoutthe simply supportedslab than in the slab supportedon four VonMises stressfor eachelementof the 70 mesh is shown in comers.The average figure 6.2.1 for eachslab type at rr--O. As a measure of the spreadof stresses within This parameter is a measure the the slab, parameter is calculated. of how the average in differs from the maximum vonMises stress, stress an element,IUVMe, vonMises (N), in the slab. all the elements, avm.,considering

161

Chapter 6

Slab Design

The following equationdefines:

N-1
The closer tends to zero, the less the variation in average vonMises stress from the for The vonMises stress average each element resulting from a 7x7 maximum. in fig. 6.2.1. detailed The values of were equal to 0.48,0.65 and element mesh are 0.70 for slabs SMI, SM3 and SM5 respectively using a 7x7 mesh, at rr--O. Hence it its lowly be that given greater number of expected stressedelements, a clearly would be lowest for Slab SM5. From this, it is clear the at would reached ff state evolved that the visualisation processis only applicable in structures where there is a relatively Values of subsequentto visualisation were calculated as wide spread of stresses. 0.40,0.54 and 0.70 for slabs SMI, SM3 and SM5 respectively. In the first two slabs become more evenly distributed, while in the third no change is observed. the stresses The lack of change in for the third slab may be attributed to the fact that while the in less becomes in the majority of the slab, a significant rise in the variation moments maximum moment at the comer support occurs, hence preserving the variation coefficient.

For each slab, the effect of increased mesh density in most casesis to increasethe is This due to the fact that the spreadof average area removed. percentage elemental over the mesh are wider than for a rough mesh.The form of the vonMisesstresses evolvedmeshfor eachmeshsize was similar. The resultingnumericalsteel areasfor For this study,a symmetricalquarterof 70 elements was similar. size eachmesh was usedfor eachmode.

162

Chapter 6

Slah Design

0,

a) Slab SM 1 (=0.48)

b) Slab SM3 (=0.65)

c) Slab SM5 ((=0.70)

fig(6.2.1) Average vonMises stress(70 elements), at rr=O

163

Chapter 6

Slah Design

i) (i-i-= 0)

(rr = 40%) fig. 6.2.2(a) Slab SM 1: Principal Moments (70 niesh)

164

Chapter 6

Slab Design

fig. 6.2.2(b) Slab SM I Principal moments, 5x5 niesh, (rr = 40%)

fig. 6.2.2(c) Slab SM I Principal moments, I Ox 10 niesh, (rr=30%)

165

Chapter 6

Slab Design

41%,
0. +

.,

Y, "O> ,',' "-, N\ Ixt


'/ "

+ J',
/I

If,
,"
* %

x//..

..

'o

...........
"". I.

/
x x

A/
iiIi

or, 'ifI

X/

. .... .....

/ /x xx xx

x,

'o

1..

-I.

iI I..

(i) rr =0

(ii) rr = 30% fig. 6.2.3(a): Slab SM3, principal moments, 70 mesh

166

Chapter 6

Slab Design

fig. 6.2.3(b): Slab SM3,5x5 mesh, (rr = 30%)

fig. 6.2.3(c): Slab SM3, I Ox10 inesh, (rr = 30%)

167

Chapter 6

Slab Design

Ii
I1

I I 1 1

f 1

f f

f7 f

tt

tt

i f 4 + + + + + ++ k*

kx--j,

6" 4 4 k0 W H t H 44 + III I I I
I I f
f
f I

I I
f f
f 1

+ f
+
+
+ #

* 4r 4 )c kk 1 +14 * it it
+ i
+ +

k yn 'A 1- -4. -+4 x V. -A

I I
I I

4r k
it x

- A*%U -0- 4.

1 1
I 4 4

+ 4
+ 4 Is

*A
-46

---

-----

-----

(i) rr =0
I;
I fI

F 1 ,, 1 1 1 .

",

W,

--, el, 00, Ool,'000"Oo

-je

-A

dp

(ii) rr = 12%

fig. 6.2.4(a) Slab SM5, principal moments, 70 niesh

168

Chapter 6

Slab Dcsign

...........

. 041

) fig. 6.2.4(b) Slab SM5, principal moments, 5x5 mesh, (rr= 151/(,

----------

Fp_-

fig. 6.2.4(c) Slab SM5, principal moments, I Ox10 mesh, (rr= 10%)

169

6 Chapter

SlabDesign

6.3

Model Design and Performance

6.3.1 Slab SM1


Loading Arrangement Design Parameters
Pd = 210 kN

635

870

635

concrete: Ec = 20.75 kN/mm2 fcu= 51.3 N/mm2 ft = 4.3 N/mm2 steel: fy = 480 N/mm 2

E. = 200 kN/mm2
635 870 635

As referred to in section6.2, the relatively evendistribution of momentsthroughout this slab meansthat only a small portion of the slab (24% of original area)has an be low low Mises The to enough assigned stress negligible von stiffness. average areasof this slab comprisethe areastowards the centre of the supporting stressed 6.2.2a). for The (fig. steel areas are shown numerical a symmetricalquarterof edges, bottom, refersto bottom steelin the x-direction, the slab in figure 6.3.1(a),whereAs,, and Asxtop,refers to top steel in the x-direction.. An overall increaseof 8% in the
for the evolved mesh was recorded. Resulting reinforcement layouts steel numerical 6.3.1 in figure (b). are shown

The numerical and provided steel areas over the 70 mesh are also shown for
6.3.1 in figure (b). In order to provide a practical steel layout, a greater comparison area of steel than theoretically required was necessary.For the models at rr--O and

increase in 20% 25% steelrespectivelywas required.The steel layout and a rr=40%, differed from the initial layout in that minimum bottom steel the mesh evolved using in 'removed' increase in bottom steel was provided the areas while an provided was
load the point. around

170

Chapter 6

Slab Design

Steel Volumes
(CM)

(rr--O)

(rr=40%)

1 Total (40%) Total (0)

A. bottom

A. top

Total

A. bottom

A. top

Total I

NurneriCal Provided L-

618.3

122.7

741

605.0

190.9

795.9 990.6 I

1.08 1.

I 799.6 1 190.8 1990.4

1749.2 1241.4

In the numerical analysis of each model, a total of 30 increments was used, with an OlPd initial increment Of O-lPd in the elastic stage,then 0.05Pdduring cracking and O, towards ultimate load. The load-displacement relationship obtained from numerical

analysisfor eachmodel is displayedin figure 6.3.1(c).The mode of failure for each be behaviour ductile. It is is that the can seen clearly of each model very model load 1.2Pd. The deflection both limit at mid-span of service attain an ultimate similar,
for both Yielding 0.65Pd slabs. around at of the steel occurred first in the was reached

bottom steelat mid-spanat a load of around I Md. (fig.6.3.I (d)). Further significant 6 in this area occurred,reaching times the yield strain at ultimate load. yielding
Similar yielding occurred in the bottom steel around the load point. Yielding of the

(fig.6.3.I (e)). top steel,aroundthe comeroccurredat a load level of around1.2Pd by the moment-curvaturerelationship is The sectional behaviour as represented
detailed at the centre and near the load point in figures 6.3.1(f-g). The moment-

in curvaturerelationshipis expressed termsof the ultimate momentof the sectionM,,,


I/Ry. Values for M,, and I/Ry were obtained from the the curvature, yield and numerical analysis of a one-way strip with equal reinforcement layers to that of the section being investigated. In this case,M, and 1/Ry at the load point were measured 0.0001mm7l 28kNnVm and respectively. At the centre of the slab, M,, and I/Ry as were l8kNm/m and 0.0001mm7l respectively. These values were the same for both

for in designs It canbe the same amount of steel was used each since case theseareas. is being beyond the that the ultimate load at thesepoints, still moment sustained seen
indicating a ductile response.In addition, it is clear that no softening occurs. For both designs, as expected, the moment curvature-relationships are similar.

171

Chapter 6

Slab Design

i) As, bottom (i-r=O)

I
, Ll

ii) A,, bottorn (rr=40%)


0

iii) A,, top (rr=O)

iv) A,, top (rr=40%)

2) fig 6.3.1(a) Slab SM 1: Symmetrical Quarter, Numerical Steel Areas (nim

172

Chapter 6

SlabDesign

I. -

150crs

75crs

150crs

75 75 !4 150crs 75crs 150crs

i) A, at bottom(rr--O)

150crs

75 75 iq 0!4 150crs

ii) A, at top (rr--o)


fig. 6.3.1(b) Slab SMI: Steel Layout, all sizes in mm, all bars 8mm diameter

173

6 Chapter

SlabDesign

150crs

75crs

150crs

75 75 150crs 75crs 150crs


T1 1

H4

>l(

iii)A, at bottom (rr--40%)

150crs

75 75 150crs

iv) A. at top (rr--40%)

fig. 6.3.I (b) SlabSMI SteelLayout, all sizesin mm, all bars8mm,diameter

174

6 Chapter

SlabDesign

9.97 (50.3) 16.86 (50.3) 24.89 (50.3) 33.90 (50.3) 41.64 (50.3) 45.89 (50.3) 46.78 (50.3)

23.23 (50.3) 28.26 (50.3) 34.42 (50.3) 42.49 (50.3) 48.52 (50.3) 49.18 (50.3) 46.08 (50.3)

37.19 (50.3) 40.24 (50.3) 45.34 (50.3) 53.76 (50.3) 55.79 (50.3) 50.65 (50.3) 42.66 (50.3)

49.98 (50.3) 56.6 (50.3) 65.43 (100.6) 73.77 (100.6) 60.51 (50.3) 47.75 (50.3) 35.98 (50.3)

55.18 (50.3) 63.87 (50.3) 73.67 (100.6) 62.18 (100.6) 50.11 (50.3) 38.45 (50.3) 26.67 (50.3)

51.22 (50.3) 53.65 (50.3) 49.87 (50.3) 40.82 (50.3) 33.79 (50.3) 26.79 (50.3) 16.93 (50.3)

45.58 (50.3) 43.69 (50.3) 38.09 (50.3) 31.17 (50.3) 24.98 (50.3) 17.82 (50.3) 8.64 (50.3)

(v) A,,, at bottom(rr--O)8mmbars


0.00 0.00 0.00 0.00 0.00 0.00

0.00

(0.00)
5.10 (0.00)
14.38

(0.00)
0.00 (0.00)
2.96

(0.00)
0.00 (0-00)
0.00

(0.00)
0.00 (0.00)
0.00

(0.00)
0.00 (0.00)
0.00

(0.00)
0.00 (0-00)
0.00

(0.00)
0.00 (0.00)
0.00

(50.3)
25.06

(0.00)
14.32

(0.00)
2.63

(0.00)
0.00

(0.00)
0.00

(0.00)
0.00

(0.00)
0.00

(50.3) 34.47 (50.3)


41.04 (50.3) 45.06 (50.3)

(50.3) 25.40 (50.3)


33.81 (50.3) 40.71 (50.3)

(0.00) 12.90 (50.3)


23.22 (50.3) 33.38 (50.3)

(0.00) 0.00 (0.00)


10.09 (0.00) 23.65 (50.3)

(0-00) 0.00 (0.00)


1.83 (0.00) 13.60 (50.3)

(0.00) 0.00 (0.00)


0.00 (0.00) 5.4i(0.00)

(0.00) 0.00 (0.00)


0.00 (0.00) 7 0.00 (0.00) I

(vi) A.,,,at top (rr--O) 8mm bars

fig. 6.3.1(b)SlabSMI: Numerical& (Provided)SteelAreasin mrn2

175

Chapter 6

Slab Design

0.00 (19.5) 0.00 (19.5) 42.55 (50.3) 51.79 (50.3) 56.34 (50.3) 59.64 (50.3) 60.44 (50.3)

0.00 (19.5) 0.00 (19.5) 57.03 (50.3) 61.35 (50.3) 62.79 (50.3) 62.44 (50.3) 59.65 (50.3)

0.00 (19.5) 0.00 (19.5) 74.58 (100.6) 67.29 (100.6) 65.71 (50.3) 61.92 (50.3) 56.17 (50.3)

19.24 (50.3) 57.53 (50.3) 80.42 (100.6) 69.68 (100.6) 59.88 (50.3) 55.21 (50.3) 49.92 (50.3)

39.43 (50.3) 52.89 (50.3) 64.69 (100.6) 59.00 (100.6) 34.03 (50.3) 35.47 (50.3) 36.69 (50.3)

43.15 (50.3) 45.10 (50.3) 51.13 (50.3) 72.91 (100.6) 0.00 (19.5) 0.00 (19.5) 0.00 (19.5)

42.83 (50.3) 43.03 (50.3) 46.80 (50.3) 57.87 (100.6) 0.00 (19.5) 0.00 (19.5) 0.00 (19.5)

(vii) As,,at bottom (ff=40%) 8mm bars


0.00

0.00

0.00

0.00

0.00

0.00

0.00

(0.00) 0.00 (0.00)


25.87 (50.3)
39.07

(0.00)
0.00

(0.00)
0.00

(0.00)
0.00

(0.00)
0.00

(0.00)
0.00

(0.00)
0.00

(0.00)
6.23 (0.00)
21.60

(0.00) '(0.00)
0.00 (0.00)
3.12

(0-00)
0.00 (0.00)
0.00

(0.00)
0.00 (0.00)
0.00

(0.00)
0.00 (0.00)
0.00

0.00 (0-00)
0.00

(50.3) 47.58 (50.3) 54.35 (50.3) 58.65 (50.3)

(50.3) 35.51 (50.3) 46.50 (50.3) 54.36 (50.3)

(0.00) 20.32 (50.3) 36.51 (50.3) 48.02 (50.3)

(0.00) 5.63 (0.00) 28.18 (50.3) 41.69 (50.3)

(0-00) 2.23 (0.00) 22.23 (50.3) 33.39 (50.3)

(0.00) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00)

(0.00) 0.00 (0.00) 0.00 (0.00) 0.00 (0-00)

(viii) Asx at top (rr=40%) 8mm bars

fig. 6.3.I (b) SlabSM I: Numerical& (Provided)SteelAreasin mm2

176

6 Chapter

SlabDesign

1.4 1.2
"0 0

c Mp 0 0.8 0.6
CLI

0- 0.4 10 0.2 0
0 10 20 30 40 central so displacement 60 (mm) 70 80 90 100

(rr=O)

--D-(rr=40%)l

fig. 6.3.1(c) Slab SMI Load-displacement relationship


1.4

1.2

m1
0.8

.00.6

0.4

0.2

1--4-- (rr- 0) --C3-(rr-40%)

0 02345 strain/yield strain

fig. 6.3. I (d) Slab SM I Bottom Steel Strains at centre


1.4 1.2

0 c

0.8

0.6

C6 0.4 cil 0.2 --*(rrm0) (rr=40%)]

0.5

1 strain/yield

1.5 strain

2.5

fig. 6.3.I (e) Slab SMI Top SteelStrainsat comer 177

6 Chapter

SlabDesign

1.2

0.8

0.6

0.4

0.2 --D- (rr--40%)

0
0 0.5 1

1.6

2.5 Ry/Ft

3.5

4.5

fig. 6.3.I (f) SlabSMI PrincipalMoment-curvature relationshipnearLoad-point

1.2

0.8

0.6

0.4

0.2

--*--U-

(rr-0) (rr-40%)

C 0 0.2 0.4 0.6 0.8 1 1.2 1.4 applied load/design load

fig. 6.3.1(g) Slab SMI Principal Moment near load point vs. Load

178

6 Chapter

SlabDesign

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1

0 (rr--O) --Cl- (rr--40%)

0.5

1.5

2 RY/R

2.5

3.5

fig. 6.3.I (h) SlabSMI: PrincipalMoment-Curvature relationshipat centre

0.9-B. O. 0.7-0.6-0.5-0.4-0.3-0.2-(rr--O) 0.1 1-. I -U(rr--40%)

0.2

0.4

0.6

0.8

1.2

1.4

applied load/design load

fig. 6.3.I (i) Slab SM I: Principal Moment at centre vs. Load

179

6 Chapter

SlabDesign

6.3.2 Slab SM2


Loading Arrangement V//Z 1070 Design Parameters Pd= 210 kN concrete: Ec = 20.75 kN/mm2 fc, = 51.3 N/mm2 ft = 4.3 N/mM2 steel: fy = 480 N/mrr Es = 200 kN/mM2

1070
P/77777777777777777 ,

1570

1570 11

The principal momentpatternof this slab(fig 6.3.2a),althoughsimilar to that of SM I, load at the centre.In is lessevenly distributeddue to the position of the concentrated lowest SMI, the the areas centre stressed of slab around occurred commonwith slab in figure is The occurring after shown path rr--30% evolved of the supportedges. 6.3.2(a).From the numericalsteel areas(fig.6.3.2b), the peak areasoccurring at the be be in the to the evolved slab can of seen smoothed comer supporting centre and increase 3% 14% in SMI, As of and numericalandprovidedsteel an model. with slab from to rr--O rr--30%. observed was respectively areas
SteelVolumes
(Cm 3)

(rr=O)

(rr--30%)

1 Total (30 %) Total (0)

A. bottom

A. top

Total

A. bottom

A. top

Total

Numerical ed

1563.5 1151.6 460.7 1612.3 1265.9 297.6 1 1 1 11 1395.3 1391.2 1 1786.5 1 1505.7 1523.2 12028.9

1.03 1.14

Resulting reinforcement layouts are given in figure 6.3.2(c). For comparison, the

in direction in the the steel areas x mesh are over also shown numericalandprovided
figure's 6.3.2(c). The main difference in the provided steel areas occurs at the

6removed'areaswhere minimum steel is placed, and the top comers, where more

180

6 Chapter

SlabDesign

in Increases for bottom is be the than the rr--O. steel around provided comers can steel

fig(6.3.2c(i)). from observed


In the numerical analysis, a total of 20 increments were used. Initial increments of OAPd.were used in the elastic stage,subsequentincrements of 0.05Pdwere used. until ultimate load. From the load-displacement relationship (fig 6.3.2d), it can be observed igns perform in a similar manner. The design at rr--O, is slightly stiffer that both des. load l. lPd an ultimate of compared to 1.05Pdat rr=30%. The service and achieves deflection limit at mid-span for both slabs was reached at around 0.7Pd. In both designs yielding of the bottom steel first occurred at a load of 0.7Pd. close to the did 6.3.2e). Yielding (fig. the top steel of not occur. centre

At the centre of the slab, M,, and I/Ry were calculated as 58kNm/m and 0.0001mnf I respectively, for each design. The moment-curvature relationship for the section at the both designs behaved that the slab shows similarly. (fig. 6.3.2g). In each of centre case, the ultimate moment was not reached until after the design load Pd was achieved.

181

Chapter 6

Skth Oesign

(1)i-r=O

(ii) rr=30% fig. 6.3.2(a) Slab SM2, principal moments

182

Chapter 6

Slab Design

(N

cc,

A,, bottom (n,=O)

ii) A,, bottom (ri-30%)

ul

l
r

iii) A,, top (n-=O)

r-_

iv) A,, top (ri-301k)

2) Steel Areas SM2, Symmetrical Slab Quarter, Numerical (mm fig. 6.3.2(b)

193

Chapter 6

Slab Design

1. . ----

40crs

75crs

75
75 14 4 150crs 014 75crs -14 4 Oc rs

(i) A, at bottom (rr=O)

I 50crs

75
75 150crs 225crs 300crs

(ii) A, at top (rr=O) fig. 6.3.2(c) Slab SM2: Steel layout, all sizes in min, all bars 8nim diameter

184

Chapter 6

Slah Dc. slll

40crs

75crs

75crs
14 -4 4

I 50ci s

40crs
1-

iii) A, at bottom (rr=30%)

I 50crs

75crs

75crs

150crs

300crs

iv) A, at top (rr=30%) fig. 6.3.2(c) Slab SM2: Steel layout, all sizes in mm, all bars 8mm diameter

185

Chapter 6

Siab Design

10.49 (50.3) 30.16 (50.3) 46.63 (50.3) 58.18 (50.3) 63.96 (50.3)

11.36 (50.3) 31.45 (50.3) 48.05 (50.3) 59.29 (50.3) 64.43 (50-3)

13.43 (50.3) 34.19 (50.3) 50.78 (50.3) 61.20 (50.3) 65.10 (50.3)

17.56 (50.3) 39.24 (50.3) 55.41 (50.3) 64.09 (50.3) 65.71 (50.3)

24.89 (50.3) 47.72 (50.3) 62.55 (50.3) 67.91 (100.6) 65.75 (50.3)

36.91 (50.3) 61.29 (100.6) 72.69 (100.6) 72.38 (100.6) 64.47 (50.3)

57.94 (100.6) 80.85 (100.6) 86.08 (100.6) 76.48 (100.6) 60.85 (50.3)

97.27 (100.6) 110.29 (100.6) 100.52 (100.6) 79.29 (100.6) 53.68 (50.3)

149.77 (150.9) 146.40 (150.9) 110.84 (100.6) 74.51 (100.6) 42.06 (50.3)

252.24 (251.5) 167.34 (150.9) 105.62 (100.6) 62.10 (100.6) 26.01 (50.3)

(v) A, at bottom (rr=O) 8rnrn bars

9.33 ).()( 29.62 (50.3) 45.93 (50.3) --577-5-7 (50.3) 63.71 (50.3)

5.50

(0.00)
28.37 (50.3) 45.36 (50.3) 57.04 (50.3) 63.55 (50.3)

0.00 (0.00) 22.38 (50.3) 44.21 (50.3) 55.98 (50.3) 63.10 (50.3)

0.00

0.00

0.00

0.00

0.00

0.00

0.00

goo)
13.27 (50.3) 40.63 (50.3) 53.58 (50.3) 61.77 (50.3)

). ()(
4.88 (0.00) 31.78 (50.3) 48.79 (50.3) 58.74 (50.3)

). ()(
0.00 ).()( 17.66 (50.3) 40.31 (50.3) 52.99 (50.3)

).()(
0.00 ).()( 3.25 ).()( 26.90 (50.3) 43.53 (50.3)

). ()(
0.00 ).()( 0.00 ).()( 8.49 (0.00) 29.82 (50.3)

). ()(
0.00 ).()( 0.00 ).()( 0.00 ).()( 13.06 (50.3)

). ()(
0.00 ).()( 0.00 ).()( 0.00 ),()( 0.00 (0.00)

(vi) A, at top (rr=O) 8mrn bars

fig. 6.3.2(c) Slab SM2: Numerical & (Provided) Steel Areas in mni 2

186

Chapter 6

Slah Design

0.00
(25.2) 0.00

0.00
(25.2) 0.00

0.00
(25.2) 0.00

9.14
(25.2) 0.00

31.70
(25.2) 0.00

24.89
(25.2) 21.61

41.07
(50.3) 36.34

87.86
(50.3) 62.65

159.01
(251.5) 185.12

277.10
(251.5) 206.57

(25.2) 73.15 (100.6) 93.57 (100.6) 86.43 (100.6)

(25.2) 73.39 (100.6) 83.36 (100.6) 86.65 (100.6)

(25.2) 73.00 (100.6) 83.07 (100.6) 87.29 (100.6)

(25.2) 71.71 (100.6) 83.38 (100.6) 88.33 (100.6)

(25.2) 68.56 (100.6) 84.39 (100.6) 89.65 (100.6)

(25.2) 60.32 (100.6) 86.62 (100.6) 89.87 (100.6)

(50.3) 59.97 (100.6) 85.06 (100.6) 89.27 (100.6)

(50.3) 49.07 (100.6) 67.89 (100.6) 72.05 (100.6)

(201.2) 0.00 (100.6) 0.00 (25.2) 0,00 (25.2)

(201.2) 91.09 (100.6) 0.00 (25.2) 0.00 (25.2)

(vil) A, at bottom (n-=30%) Snini bars

0.00 ).()( 0.00 ).()( 72.12 (50.3) 83.49 (100.6) 86.53 (100.6)

0.00 oo) (O. 0.00 ).0( 69.79 (50.3) 82.93 (100.6) 86.89 (100.6)

0.00 ).0( 0.00 oo) (O. 65.21 (50.3) 81.53 (100.6) 87.31 (100.6)

5.51 (0.()( 0.00 ).()( 56.38 (50.3) 78.76 (100.6) 87.39 (100.6)

8.61 ).()( 0.00 ).()( 38.32 (50.3) 74.10 (100.6) 86.26 (100.6)

3.99 ).()( 0.00 ).()( 20.70 (50.3) 67.54 (100.6) 84.21 (100.6)

0.00 ).()( 0.00 ).()( 11.65 (50.3) 66.31 (100.6) 81.76 (100.6)

0.00 ).()( 0.00 ).()( 11.33 (50.3) 53.30 (100.6) 69.38 (100.6)

0.00 ).()( 0.00 (0.00) 0.00 ).()( 0.00 (0.00) 0,00 (0.00)

0,00 ).()( 0.00 ).()( 0.00 ).()( 0,00 ).()( 0.00 (0.00)

(viii) A,, at top (rr=30%)

8mrn bars

fig. 6.3.2(c) Slab SM2: Numerical & (Provided) Steel Areas in ninil

187

Chapter 6

Slab Design

1.2

0.8
0.6

0.4

0.2

-0 30

10

20

30 central

40 displacement

50 (mm)

60

70

80

90

fig. 6.3.2(d) Slab SM2: Load-displacernentrelationship


1.2

0 c

0.8

0.6

rL 0.4 CL m

0.2

0 --0--

(rr=O) (rr=30%

0
0234567 strain/yield strain

fig. 6.3.2(e) Slab SM2: Bottom Steel Strains at centre


1.2

O8

0.6

04

0.2

0 0 0.1 0.2 0.3 0.4 strain/yield 0.5 strain 0.6 0.7 0.8 0.9

fig. 6.3.2(f) Stab SM2: Top Steel Strains at corner

Iss

Chlplcl 0

1.2

0.8

0.6

0.4

0.2

(rr-0)

J)

(rr=30%,

0uI
023456

7 Ry/R

fig. 6.3.2(g) Slab SM2: Principal Mornent-curvature relationship at centi-c

1.2

0.8

0.6

0.4

0.2

(rr=O) (rr=30%)

OLJ 0 0.2 0.4 applied 0.6 load/design load 0.8 1 1.2

fig. 6.3.2(h) Slab SM2: Principal Moments at centre vs. Load

199

Chapter 6

Slab Design

6.3.3 Slab SM3


Loading Arrangement Design Parameters Pj = 210 kN concrete: E, = 22.5 kN/nim 2 f, = 58.8 N/mm 3.1 Mimi 1', = steel: fy 560 N/mm 2 E, 200 Mmm" 635 870 635

635

870

635

The evolved pattern for this slab follows a clearly defined load path, from corner to in The the the support centre. column stress concentration along this crossing corner load path leads to the 'removal' of a high % area at a comparable rejection ratio to the previous slabs.

Steel

olumes

(rr=O)

(rr=30%)

Total(30%

I) (Cm A, bottom Numerical Provid d 319.6 450.2 As top 107 220.4 426.6 670.6 Total A, bottom 261.3 446.1 A, top 220.4 294.4 481.7 740.5 Total

Total(O)

1.1 1.1

From the numerical steel areas,fig. 6.3.3(a), it can be seen that the amount of bottom load load the the and along point path increasessubstantially for the design at steel at in increases Further the required top steel at the central column support can rr=30%. be seenfrom fig. 6.3.3a(i-ii). The resulting steel layout is shown in figure 6.3.3(b). The for design in these the areas at rr=30% was double that of the model at provided steel
6.3.3(b)). The fig. (see for total amount the design at rr=30% of steel provided rr=O

by 10%. increased was

190

Chapter 6

Slab Design

Because of the significant increase of provided steel in the critical areas, Le along the diagonal, for the design at rr=30%, it can be expected that the ultimate load of this

design would be greater. This assumption is confirmed by the numerical analysis. The load 1.31? 1.1 Pd of an ultimate compared model achieved with an of second ultimate d for the design at rr=O. In addition, the load-displacement response of the second model from design load to ultimate load for the The significant increase was also stiffer. model at rr=30% may in part be due to the fact that in order to make a practical steel layout, 35% increase from the total theoretical steel was required. The service deflection limit at the mid-span of each quarter was reached at 0.7Pj and 0.85P, j for the designs at rr=O and rr=30% respectively. Yielding of the bottom steel first

0.85P, for the models at 1-1-0 and and occurred at the load point at levels of 0.71? d j rr=30% respectively.

For the section at the load point, the values of 18kNm/m and 0.00008mm- I design, at rr=30%,

M, and I/Ry were calculated as

respectively, in the design at rr=o. For the second and

values of M,, and I/Ry were calculated as 22kNrn/rn

0.00008mm- 1 respectively. The increase in ultimate moment at rr=30%, is due to the increase in provided steel at this point. Inspection of the mornent-applied load

relationship at the load point shows that eventual softening occurred in the rr=30% design at an ultimate load of 1.2Pd, (fig. 6.3.3f-g). A similar increase in M,, was

observed in the section at the corner due to the provision of more top steel at rr=30%., In this case, M,, was 27kNm/m at rr=O, and 42kNm/m at rr=30%.

191

Chapter 6

Stab Design

10

ii) A,, bottom (rr=30%)

iii) A,,,top (rr=O)

iv) A,, top (rr=30%)

fig. 6.3.3(a) Slab SM3, Symmetrical Quarter, Numerical Steel Areas (MM2)

192

Chapter 6

Slab Design

75crs

150crs

150crs 1.4 Nid

75crs

150ers '

Nid
i) A, at bottom(rr--O)

NI

I-

-.

50crs 75

300crs 1 75

75 - 300crs k 014

75 50crs k 014 0

ii) A, at top (ff=O)


fig. 6.3.3(b) Slab SM3: Steel Layout, all sizes in mm, all bars 6mm,diameter

193

Chapter 6

Slab Design

I-

300crs

37.5crs

300crs

75crs

75crs 300 crs 4 014 -

37.5crs 014

300crs

iii) A, at bottom(rr--30%)

37.5 ers -Z1 75

150 crs 75 75 150 crs 75 37.5 crs 1

iv) A, at top (rr--30%) fig. 6.3.3(b) Slab SM3: Steel Layout, all sizes in mm, all bars 6mm.diameters

194

6 Chapter

SlabDesign

7.79 (28.3) 7.10 (28.3) 7.38 (28.3) 13.06 (28.3) 19.24 (28.3) 22.79 (28.3) 23.44 (28.3)

23.18 (28.3) 22.81 (28.3) 17.16 (28.3) 19.93 (28.3) 25.28 (28.3) 25.71 (28.3) 22.81 (28.3)

37.95 (56.6) 42.27 (56.6) 34.43 (56.6) 29.80 (28.3) 32.05 (28.3) 26.97 (28.3), 19.74 (28.3)

43.83 (56.6) 58.53 (56.6) 68.31 (56.6) 52.47 (56.6) 35.45 (28.3) 23.13 (28.3) 13.75 (28.3)

31.74 (56.6) 40.38 (56.6) 49.21 (56.6) 54.59 (56.6) 31.20 (28.3) 14.99 (28.3) 6.23 (28.3)

4.14 (14.2) 6.71 (14.2) 8.58 (14.2) 14.06 (14.2) 14.78 (14.2) 9.05 (14.2) 3.19 (14.2)

0.00 (14.2) 0.00 (14.2) 0.00 (14.2) 0.00 (14.2) 0.00 (14.2) 0.00 (14.2) 0.00 (14.2)

(v) A,,, at bottom (rr--O) 6mm bars

0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 4.86 (0.00) 12.52 (14.2) 18.66 (14.2) 22.06 (28.3)

0.00 (0-00) 0.00 (0.00) 0.00 (0.00) 0.00 (0-00) 4.73 (0.00) 13.24 (14.2) 18.73 (28.3)

5.32 (0.00) 0.00 (0.00) 0.00 (0-00). 0.00 (0.00) 0.00 (0-00) 6.28 (0-00) 13.24 (28.3)

14.24 (0.00) 2.72 (0.00) 0.00 (0-00) 0.00 (0.00) 0.00 (0.00) 0.00 (0-00) 5.70 (28.3)

26.28 (28.3) 6.50 (0.00) 0.00 (0-00) 0.00 (0-00) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00)

51.20 (56.6) 16.78 (28.3) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00)

86.51 (84.9) 15.43 (28.3) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00)

A.,,,at top (iL-0) 6mm bars fig. 6.3.3(b) Slab SM3: Numerical & (Provided) Steel Areas in mm2

195

Chapter 6

Slab Design

0.000 (14.2)

0.000 (14.2)

0.000 (14.2)

0.000 (14.2)

2.277 (14.2)

0.000 (14.2)

0.000 (14.2)

0.000
(14.2) 0.000 (14.2) 0.000 (14.2) 0.000 (14.2) 55.834 (56.6) 56.396 (56.6)

0.000
(14.2) 0.000 (14.2) 0.000 (14.2) 12.387 (14.2) 23.169 (56.6) 43.444 (56.6)

0.000
(14.2) 0.000 (113.1) 71.844 (113.1) 40.030 (56.6) 6.372 (14.2) 0.000 (14.2)

33.215
(28.3) 100.151 (113.1) 124.362 (113.1) 67.130 (56.6) 0.000 (14.2) 0.000 (14.2)

21.579
(28.3) 63.849 (113.1) 122.283 (113.1) 0.000 (14.2) 0.000 (14.2) 0.000 (14.2)

0.000
(14.2) 4.147 (14.2) 22.669 (14.2) 0.000 (14.2) 0.000 (14.2) 0.000 (14.2)

0.000
(14.2) 0.000 (14.2) 0.000 (14.2) 0.000 (14.2) 0.000 (14.2) 0.000 (14.2)

(vii) A.,,, at bottom(ff=30%) 6mm bars


0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 3.460 (0.000) 37.761 (56.6) 61.632 (56.6) 0.000 (0.000) 0.000 (0.000) 0.000 (0-000) 0.000 (0.000) 0.000 (0.000) 12.157 (0.000), 0.000 (14.2) 0.000 (0-000) 2.675 (0.000) 0.000 (0-000) 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (14.2) 4.922 (0.000) 4.624 (0.000) 0.000 (0-000) 1.883 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (14.2) 46.914 (150.9) 51.292 (150.9) 36.606 (100.6) 7.241 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (14.2) 152.496 (150.9) 111.046 (150.9) 85.961 (100.6) 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (14.2)

0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 55.524 (56.6) 58.099 (56.6)

(viii) A,,, at top (rr--30%) 6mm bars 2 fig. 6.3.I (b) Slab SM3: Numerical & (Provided) Steel Areas in mm.

196

Chapter 6

Slab Design

1.4

1.2

0 C Im0.8
.20.6 0.4 I --*0.2 --D(rr-0) (rr-30%)

02468 displacement at load point (mm)

10

12

fig.6.3.3(c)SlabSMI Load-displacement at load point


1.4 1.2

0.8

.20.6 V 0- 0.4 0.2 0 2345 steel strain/yield strain -"-D(rr=O) (rr=30%)

fig. 6.3.3(d) Slab SM3: Bottom Steel Strains at load point


1.4

1.2

C
ID

0.8

0.6

0.4

0.2

I --*--D-

(rr-0) (rr-30%)

0 0 0.5 1 1.5 2 2.5 3 strain 3.5 4 4.5 5 steel strain/yield

fig. 6.3.3(e) Slab SM3: Top Steel Strains at column support

197

6 Chapter

SlabDesign

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.5 1.5 RyIR 2 2.5 3 3.5

(rr=O) --D- (rr=30%) *

fig. 6.3.3(f) SlabSMI Principalmoment-curvature relationshipat load point


1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 applied load/design load

(rr=O) (rr=30%)

fig. 6.3.3(g)SlabSM3: PrincipalMomentsat load point vs. Load

198

6 Chapter

SlabDesign

1 In 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.5 1 Ry/R 1.5 2 2.5

10 (rr--O) --D- (rr=30%)]

fig. 6.3.3(h)SlabSM3: PrincipalMoment-Curvature relationshipat column support


1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1

4 (rr=O) (rr=30%) -E]-0 0.2 0.4 0.6 0.8 1.2 1.4

applied load/design load

fig. 6.3.3(i) Slab SM3: Principal Moments near column support vs. Load

199

Chapter 6

Slab Design

6.3.4 Slab SM4


Loading Arrangement DesignParameters
Pd --110 kN 713.3 713.3 713.3 concrete: Ec = 20.75 kN/mm2 fcu = 51.3 N/mm2

ft = 3.0 N/mm2 steel:

f = 480 N/mm 2 y E. = 200 kN/mm2

713.3

713.3 713.3

in this slab were concentrated The maximum stresses along the main diagonal at thenspread the comer of out moreevenlyapproaching columnsupport.Thesestresses low A the simply supported rejectionratio of 12%resultedin a high percentage edges. areshownat eachstagein figure 6.3.4(a) area'removal'. The principalmoments
L steel Volumes 'V Steel S teel ,
(CM)

(rr=O)

(rr=12%)
Total A. bottom 3826.6 3625.9 A. top 1093.4 4719.3 1264.5 6562.3 Total

Total(12%)
Total(O)

cl

A. bottom Numencal PP rovide rovided

A. top

3356.2 470.4

1.23 1.43

I 3953.6 1633.8 14587.4 15297.8

From figure 6.3.4(b), it can be seenthat some of the peak moments around the comer 'smoothed' by load the time rr--12% is reached. A 23% increase were out points and

from rr--Oto rr--12%is observed. in the total numericalsteelareas The main increases in numericalsteeloccurredaroundthis main load path and in particular,the top steel increased by 100%. Resultingreinforcement layoutsare shown the support column at in figure 6.3.4(c).The numericalandprovidedsteelareas over the slab are also shown
for comparison in figures 6.3.4(c).

200

Chapter 6

Slab Design

The load-displacement response(fig. 6.3.4d) shows that an ultimate load of 1.2Pdand be It the that 1.31? can also seen and rr--12% respectively. at rr_-O achieved was d behaviour of the model designedat rr--Owas stiffer up until around the ultimate. After increase be from is ductile first the the can seen significant as this point, model more in displacement. Both models failed in a ductile manner. The service deflection limit level load 0.75Pd 0.7Pd the at a of and at rr--O and the reached slab was of centre at first bottom Yielding the the the steel occurred centre of of at respectively. rr-_12% (fig. 6.3.4e) for both Yielding 1.2Pd level load models, of the top steel was of slab at a initiated at I Pdin the first model and at I -I Pdin the secondmodel, (fig 6.3.4f). -I For the section at the centre of the slab, values of M,, and I/Ry were calculated as l7kNm/m and 0.00005mm", at u-0. For the second design , at rr--12%, the values of M,, and I/Ry for the same section were 2lkNm/m and 0.00005mm" respectively. From the moment curvature relationship at the centre, figures 6.3.4(g-h), it can be for has in bottom increase this the steel provided at the point model that second seen led to a higher moment being sustainedbut with a much smaller curvature. Hence the design at rr--O,is more ductile in this area.

Although the overall behaviourof both designswas ductile, the increasein steel in localised loss in lead ductility. be This to a may can observed at rr--12% somezones from the moment-curvature plot at the centre section. Additionally, in this case, increasein requiredsteel createsa more complicatedsteel layout, and may lead to congestion.

201

"I'ih

Chaoter 0

(i) rr=O

(ii) rr=12% fig. 6.3.4(a) Stab SM4, principal moments 202

Chapter 6

Slab Design

) c::

-42

i) A,, bottom (rr=O)

C N

Cj ,0
-4

bottom 12%) (rr= A,,, i)

iii) A,, top (rr=O)

iv) A,, top (rr= 12%)

2) fig. 6.3.4(b) Stab SM4: Numerical Steel Areas (mm

203

Chapter 6

Slab Design

150crs

75crs

150crs

75
75 150crs i4 014 75crs 014 4 I 50crs

(i) As at bottorn (rr=o)

50crs 75

150crs

75 75
0i4i4

i4

150crs

75 50crs
014 1.

(ii) A, at top (rr=O)


fig. 6.3.4(c) Slab SM4: Steel layout, all sizes in nim, all bars 8mm diameter

204

Chapter 6

slall Design

300

37.5 crs

50 crs

75crs

75crs Oio

50 crs

37,5 crs 0! 4

300

(i i i) A, at bottom (rr= 12%)


30crs 75

L1

450

75crs

75crs

450

75 30cr-s

(i v) A, at top (rr= 12

fig. 6.3.4(c) Slab SM4: Steel Layout, all sizes in mm, all bars ginni diameter

205

Chapter 6

Slab Design

24.70 (50.3) 25.89 (50.3) 20.78 (50.3) 12 93 , (50.3) 22 05 , (50.3) 35.59 (50.3) 46.79 (50.3) 54.49 (50.3) 59.72 (50.3) 59.73 (50.3)

50.99 (50.3) 52.36 (50.3) 46.96 (50.3) 34.42 (50.3) 38.10 (50.3) 50.28 (50.3) 58.04 (50.3) 61.79 (50.3) 61.94 (50.3) 58.81 (50.3)

93.24 75.55 (100.6) (100.6) 98.67 79.07 (100.6) (100.6) 77.51 105.88 (100.6) (100.6) 64.52 113.03 (100.6) (50.3) 94.65 58.16 (50.3) (100.6) 79.90 66.28 (50.3) (100.6) 73.45 68.05 (100.6) (50.3) 66.81 68.84 (50.3) (50.3) 60.00 62.45 (50.3) (50.3) 47.76 54.88 (50.3) (50.3)

99.63 94.44 80.07 1 56.05 (50.3) (100.6) (100.6) (100.6) 94.16 79.10 58.55 102.93 (100.6) (100.6) (100.6) (50.3) 75.50 56.10 91.44 107.68 (100.6) (100.6) (100.6) (50.3) 93.15 74.13 52.46 115.28 (100.6) (100.6) (100.6) (50.3) 78.52 119.54 103.71 48.52 (100.6) (100.6) (100.6) (50.3) 87.82 109.24 94.76 47.24 (100.6) (100.6) (100.6) (50.3) 73.69 95.79 85.74 49.28 (100.6) (100.6) (100.6) (50.3) 62.70 68.00 56.02 43.17 (50.3) (50.3) (50.3) (50.3) 42.26 54.11 30.90 29.42 (50.3) (50.3) (50.3) (50.3) 23.51 37.28 11 79 1543 , (50.3) (50.3) (50.3)

17.33 (50.3) 35.16 (50.3) 35.10 (50.3) 34.01 (50.3) 30.08 (50.3) 28.82 (50.3) 29.43 (50.3) 28.35 (50.3) 23.04 (50.3) 16.43 (50.3)

0.00 (50.3) 15.52 (50.3) 18.93 (50.3) 20.95 (50.3) 18.81 (50.3) 17.49 (50.3) 17.05 (50.3) 10.37 (50.3) 14.64 (50.3) 12.59 (50.3)

(v) A, at bottom (rr=O) 8nini hars


2.47 (0.00) 0,00 ). ()( 0.00 ). ()( 0.00 ). 00) 0.00 ). 0( 0.00 ). ()( 0.00 ). ()( 0.00 (0.00) 13.09 (50.3) 140.7 (150.9)

3.39
). ()( 0.00

0.00
). ()0) 0.00

0.00
(0.00) 0.00

0.00
). ()0) 0.00

0.00
). ()( 0.00

0.00
). ()( 0,00

0.00
). ()( 0.00

0.00
). ()( 0.00

12.24
(50.3) 0.00

19.11
(50.3) 9.9-1

(0.00) O. (JO

f- o). o0 ()( ()0) ). ). 0.00 6.21


).()(
20.86

(0.00)
0.00

00) ).
0.00 0.00 0.00

(0.00) 0.00

0.00

0.00

(0.00) 0.00

). 00)
. 0.00 (0.00)
3.14

(0.00)
0.00 ).()0)
0.00

). 0(0.00 ).()0)
0.00

). ()(
0.00 ).()(
0.00

). ()(
0.00 ).()(
0.00

),()(
0.00 ).()(
0.00

). ()(
0.00 ).()(
0.00

).()(
0.00 ).()(
0.00

oo) (O.
5- 12

0.00 ).()0)

(50.3)
34.33

).()0)
19.27

).()(
0.00

).()0)
0.00

).()(
0.00

).()(
0.00

(0.00)
0.00

).()(
0.00

).()(
0.00

(50.3) 45.06 (50.3) 52.86 (50.3) 57.75 (50.3)

(50.3) 32.92 (50.3) 44.13 (50.3) 52.78 (50.3)

(0.00) 17.45 (50.3) 32.27 3) (_O. 44.69 (50.3)

).00) (0.00) ).()( 2.85 0.00 0.00 ).0( ).00) (0.00) 3.21 17.39 0.00 ).()0) (50.3) (0-00) 33.40 1 19.22 1 5.65 (503 (50.3) (50.3)

(vi) A,, at top (rr=O) 8mm bars

).()( 0.00 ).()( 0.00 (0.00) 0.00 (50.

).()( 0.00 (0.00) 0.00 ).()( 3.95 .

).()( 0.00 ).()( 0.00 ).()( 9.89 (50.3)

).()( 0.00 ).()( 3.26 (0.00) 9.54 (50.3)

fig. 6.3.4(c) Slab SM4: Numerical & (Provided) Steel Areas in mrn

206

Chapter 6

Slab Design

0.00 (25.2) 0.00 (25.2)


0.00

0.00 (25.2) 0.00 (25.2)


0.00

0.00 (25.2) 0.00 (25.2)


69.30

0.00 (25.2) 61.52 (100.6)


118.57 144.68 (150.9) 127.46 (150.9) 87,85 (150.9) 66.74 (100.6) 65.73 (100.6) 77.79 (100.6) 92.56 (100.6)

0.00 (25.2) 47.09 (100.6)


130.51

0.00 (25.2) 0.00 (100.6)


191.95

(25.2) 0.00 (100.6)


220.06

0.00 1 0.00
(25.2) 91.38 (100.6)

(25.2) 41.71 (100.6)

0.00

0.00
(25.2) 10.23 (25.2)
0.00

190.48

100.01

(25.2)
0.00 (25.2) 0.00 (25.2) 111.44 (150.9) 108.29 (100.6) 101.72 (100.6) 100.27 (100.6) 100.39 (100.6)

(25.2) ,

(201.2)
73.42 (150.9) 108.42 (150.9) 107.67 (150.9) 98.39 (100.6) 91.76 (100.6) 92.64 (100.6) 95.17 (100.6)

0,00 (25.2) 0.00 (25.2) 161.20 (150.9) 120.06 (100.6) 104.95 (100.6) 100 69 (100.6) 9896 " ( 100 6) .

(201.2) , (201.2)

(201.2)

(201.2)
158.60 (150.9) 123.45 (150.9) 114.03 (150.9) 121.18 (100.6) 108.77 (100.6) 0.00 (25.2) 0.00 (25.2)

(201.2)
160.71 (150.9) 74.25 (150.9) 41.49 (150.9) 74.43 (100.6) 65.59 000.6) 0.00 (25.2) 0.00 . (25.2)

(201.2)
0.00 (25.2) 0.00 (25.2) 3.64 (25.2) 34.09 (25.2) 0.00 (25.2) 0.00 (25.2) 0.00 (25.2)

(25.2)
0.00 (25.2) 0.00 (25.2) 0.00 (25.2) 0.00 (25.2) 0.00 (215.2) 0.00 (25.2) 0.00 (25.2)

(vii) A,, at bottom (rr= 12%) gnim bars

155.81 156.67 (150.9) (150.9) 145,88 156.38 (150.9) (150.9) 138.80 110.16 (150.9) (150.9) 100.29 58.39 (100.6) (100.6) 21.14 80.04 (100.6) (100.6) 60.41 0.00 (100.6) (25.2) 87.93 0.00 (100.6) (25.2)

0.00 ).()(
0.00

0.00 ).()(
0.00

0.00 (0.00)
0.00

0.00 ).()(
0.00

0.00 ).()(
0.00

0.00 ).()(
0.00

0,00 ).()(
0.00

0.00 (0.00)
5.15

29.93 (50.3)
10.97

293.50 (301.8)
25.98

).()(
0.00 (0.00) 0.00 ). ()(

).()0)
0.00 ). ()( 0.00 ). ()(

).()(
4.93 ). ()( 21.94 ). ()0)

).()(
0.00 ). 0( 0.00 ). ()(

(0.00)
0.00 (O.oo) 0.00 ). ()(

).()(
0.00 (0.00) 0.00 ). ()(

).()(
6.66 ). ()( 4.75 ). ()(

).()(
20.68 (0.00) 39.51 ). ()( .

).()(
3.99 ). ()( 0.00 ). ()(

(50.3)
0.00 ). ()( 0.00 ). ()(

0.00 ).()( 63.63 (100.3) 76.19 (100.3)


83.57

0.00

13.28

0.00

0.00

0.00

0.00

26.25

0.00

0.00
(0.00)

oo) (O.
11.09 (0-00) 29.91 (100.3) (100.3)
80.66 (100.3) 93.54 (100.3) 59 64 -

). ()(
0.00 (0.()( 5.22 (0.00)
1 39.89

). ()(
0.00 ).()0) 0.00 ).()(
21.40

). ()(
0.00 oo) (O. 0.00 ).()(
16.49

). ()(
0.00 ).()( 0.00 ).()(
22.75 1 1

). ()(
0.00 ).()( 0.00 ).()(
0. ()()

).()(
7.94 ).()( 4.75 ).()(
4.88

). ()(
14.15 ).()( 6.78 ).()(
0.00

0.00 ).()( 0.00 ).()(


0.00

(100.3)
92.22 (100.3) 98 15

(100.3)
71.65 (100.3) 90.11 (1

(100.3)
70.29 (100.3) 1 91.81 (100.3)

(0.00)
72.29 (100.3) 86.80 (100.3)

).0())
0 (0 o) 0.00

).()(
). ()( 1 0, oo (C

).()(
0.00 ). ()( m 0.00 (o. oo

).()(
0.00 ). ()( 1 () ( ) () .

).()(
0.00 (0.00) 1 0.00 ). ()(

(viii) A,, at top (rr=12%) Snim bars

fig. 6.3.4(c) Slab SM4: Numerical & (Provided) Steel Areas in mm2

207

Chapter 6

Slab Wsign

1.4

1.2

0.8

.20.6 A
CL CL 0.4 m

0.2 0
05 10 15 central 20 25 30 (mm) 35

1 D) 12%)l

40

45

displacement

fig. 6.3.4(d) Slab SM4: Load-displacementrelationship at centre


1.4

1.2

C,
o. 8

0.6

CL OA to

0.2 0-iiii
0 0.2 0.4 0.6 strain/yield 0.8 strain 1

(rr= 12%

1.2

1.4

fig. 6.3.4(e) Slab SM4: Bottom Steel Strain at centre


1.4

1.2

'0 1 15 0
c Im ,U; 0.8 0 0.6
0.4

0.2

01.00

0.50

1.00

1.50 strain/yield strain

2.00

2.50

3.00

fig. 6.3.4(f) Slab SM4: Top Steel Strains near colunin SLIPPOI-t 208

Chapter 6

Slab Design

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 -0(rr=O) (r 712%)j -r

0.5

1.5

2.5 Ry/R

3.5

4.5

fig. 6.3.4(g) Slab SM4: Principal Moment Curvature Relationship at centre

1 0.9 0.8 0.7 0.6

2 2

0.5
0.4 0.3 0.2 0.1

0 -EI-

(rr=0) (rr= 12%)

0.2

0.4

0.6 applied

0.8 load/design load

1.2

1.4

fig. 6.3.4(h) Slab SM4: Moments at centre vs. Load

209

Chapter 6

Slab Design

6.3.5 Slab SM5


Loading Arrangement Design Parameters Pd= 50 kN 455 concrete: Ec = 19 kN/mm2 fcu= 37.9 N/mm2 ft = 3.0 N/mm2 steel: fy = 345 N/mm2 Es = 180 kN/mm 2
455

455 I17.

455

This slab was centrally loadedand pin-supported at its four comers.Since this slab it resultedin the hasthe mostpronounced anddirect load path of all thoseconsidered, greatest% arearemoval (58%), at the lowest rejection ratio (127o)of all the slabs designed.

Steel Volumes
(CM)

(rr--O)

(rr--12%)

Total(12%) Total(O) Total

A. bottom Numerical Provided 48.2 69.1

A. top 5.4 8.4

Total

A. bottom

A. top 18.2 18.4

53.6 77.5

44.6 83.9

62.8 102.3

1.2 1.3

In terms of the numerical steel, a decrease of 8% in bottom steel is obtained at rr--12%, while an increase of 330% is observed for top steel at rr--12%. Although a

in bottom steelis observed, total decrease this is due to large % arearemoval.In the areaaroundthe main load path,much larger quantitiesof steel,an increase of around 50%, are requiredthan at rr--O,(fig.6.3.5a).This is clearly observed in fig(6.3.5a(ii)).
The resulting reinforcement layout and the element by element steel areas are displayed in figure 6.3.5(b).

210

Chapter6

SlabDesign

The numerical load-displacement responseof each model is shown in figure 6.3.5(c). An ultimate load of 1.1 and 1.35Pdwas achieved for the designs at rr--O and rr--12% load increase in is due larger The the to the the ultimate of second model respectively. fact that in order to obtain a practical layout, the total provided bottom steel was in both The the steel. stiffness numerical provided of of models was relatively excess fail in ductile Pd. Both The deflection limit a models manner. service up until similar for 0.65Pd 0.7Pd design and at was reached one and two respectively. at mid-span Yielding of the bottom steel first occurred at the centre of the slab in design one at UP& and at 1.2Pdin design two, (fig. 6.3.5c). The difference in these loads is caused by the large increase in provided bottom steel and numerical steel in the second design. Yielding of the top steel first occurred in both slabs at the comer support at (fig. 6.3.5d). O-8Pd

For the section at the centre of the slab, M,, and I/Ry were calculated as 5.2kNm/m In design, 0.0002mm" the second at rr--O. at the same section, of Mu respectively, and 1 From 0.0002mirf I/Ry 8.5kNm/m the moment curvature and respectively. were and figure 6.3.5(f), it can be seen that the second the the slab, centre of relationship at design, due to its large increase in steel area at this point, is sustaining a larger ductile first design. Softening is initiated less is than the at this moment and slightly section in the second design at around 1.35Pd.(fig. 6.3.5g). For the section at the comer support, Mu and I/Ry were 6kNm/m and 0.0002mm-1 respectively, at rr=O. In the second design at this area, M,, and I/Ry were 8.5kNm/m and 0.0002mm-1 respectively.

211

Chapter 6

Stab Design

00

i) A,, bottom (rr=O)

C'4

ii) A,, bottom (rr=12%)

iii) A,, top (rr=O)

iv) A,,,top (rr= 12%)

Areas (mmSteel Numerical Quarter, SM5, Symmetrical Slab fig6.3.5(a)

212

Chapter 6

Slab Design

------------

65crs

120crs

32.5

120crs

65crs

i) A,,at bottom(ff--O)
1

152.5

32.5 32.5 152.5 i

it NI-4

bl
(ii) Ar,at top (rr--O)

fig. 6.3.5(b) Slab SM5: Steel Layout, all sizes in mm, all bars 6mm.diameter

213

6 Chapter

SlabDesign

32.5crs

195

32.5 32.5 195 32.5crs

(iii) As at bottom (rr--12%)

130

22 crs 22 crs 130

(iv) A,, at top (rr--12%)

fig. 6.3.5(b)SlabSM5: SteelLayout, all sizesin mm, all bars6mm diameter

214

6 Chapter

SlabDesign

0.000 (3.8) 0.000 (3.8) 0.705 (3.8) 0.798 (3.8) 0.000 (3.8) 0.894 (3.8) 0.000 (3.8)

1.476 (3.8) 1.611 (3.8) 1.825 (3.8) 2.028 (3.8) 1.623 (3.8) 0.854 (3.8) 0.000 (3.8)

3.411 (3.8) 3.728 (3.8) 3.835 (3.8) 4.066 (3.8) 4.385 (3.8) 4.159 (3.8) 3.504 (3.8)

6.112 (14.2) 6.862 (14.2) 7.101 (14.2) 6.903 (14.2) 7.080 (14.2) 7.444 (14.2) 7.461 (14.2)

9.733 (14.2) 10.984 (14.2) 10.980 (14.2) 10.362 (14.2) 9.787 (14.2) 9.884 (14.2) 10.116 (14.2)

15.182 (28.3) 16.028 (28.3) 14.760 (14.2) 13.252 (14.2) 12.185 (14.2) 11.757 (14.2) 11.888 (14.2)

25.174 (28.3) 20.362 (28.3) 16.691 (14.2) 14.559 (14.2) 13.358 (14.2) 12.804 (14.2) 12.755 (14.2)

(v) A.,,, at bottom (rr=O)6mrnbars


0.000 0.000 0.000 0.000 0.000 0.000

0.000

(0.00) 0.000 (0.00)


(0.00)

(0.00)
0.000

(0.00)
0.000

(0.00)
0.000

(0.00)
0.000

(0.00)
0.000

(0.00)
0.000

(0.00)
0.000

(0.00)
0.000

(0.00)
0.000

(0.00)
0.000

(0.00)
0.000

(0.00)
0.000

0.000

(0.00) 0.000 (0.00) 0.000 (0.00)


0.758 (3.8) I 4.167 (28.3)

(0.00) 0.000 (0.00) 0.000 (0.00)


0.000 (0.00) 0.000 (0.00)

(0.00) 0.000 (0.00) 0.000 (0.00)


0.000 (0.00) 0.000 (0.00)

(0.00) 0.000 (0.00) 0.000 (0.00)


0.000 (0.00) 0.000 (0.00)

(0.00) 0.000 (0.00) 0.000 (0.00)


0.000 (0.00) 0.000 (0.00)

(0.00) 0.000 (0.00) 0.000 (0.00)


0.000 (0.00) 0.000 (0.00)

0.000 (0.00) 0.000 (0.00)


3.491 (3.8) 33.127 (28.3)

(vi) A. at top (rr--O)6mm bars


2 fig. 6.3.5(b) Slab SM5: Numerical & (Provided) Steel Areas in mm,

215

6 Chapter

SlabDesign

0.000 (3.8) 0.000 (3.8) 0.000 (3.8) 0.000 (3.8) 0.000 (3.8) 1.576 (3.8) 0.962 (3.8)

0.000 (3.8) 0.000 (3.8) 0.000 (3.8) 0.000 (3.8) 0.000 (3.8) 0.000 (3.8) 0.659 (3.8)

0.000 (3.8) 0.000 (3.8) 0.000 (3.8) 9.310 (28.3) 2.929 (3.8) 0.000 (3.8) 0.000 (3.8)

0.000 (3.8) 0.000 (3.8) 23.394 (42.4) 19.025 (28.3) 8.101 (3.8) 0.000 (3.8) 0.000 (3.8)

7.371 (42.4) 22.907 (42.4) 36.038 (42.4) 23.042 (28.3) 0.000 (3.8) 0.000 (3.8) 0.000 (3.8)

18.464 (42.4) 27.200 (42.4) 43.952 (42.4) 0.000 (3.8) 0.000 (3.8) 0.000 (3.8) 0.000 (3.8)

32.024 (42.4) 30.795 (42.4) 35.811 (42.4) 0.000 (3.8) 0.000 (3.8) 0.000 (3.8) 0.000 (3.8)

(vii) A,,,, at bottom (rr=12%)6mmbars

0.000

0.000

0.000

0.000

0.000

0.000

0.000

(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) 0.000 0.000 0.000 0.000 0.000 0.000 0.000 (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
0.000 0.000 0.000 0.000 0.000 0.000 (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) 0.000

(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)


0.000 (0.000) 4.137 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (0.000)

0.000

0.000

0.000

0.000

0.000

0.000

0.000

15.030 (14.2)
86.860 (84.8)

11.760 (14.2)
17.026 (84.8)

5.165 (0.000)
0.000 (0-000)
r ... (Vill

0.000 (0.000)
0.000 (0.000)

0.000 (0.000)
0.000 (0.000)

0.000 (0.000)
0.000 (0.000)

0.000 (0.000)
0.000 (0.000)

(rr--12%) 6mm bars top at x

fig. 6.3.5(b) Slab SM5: Numerical & (Provided) Steel Areas in mm2

216

Chapter 6

Slab Design

1.6 1.4 1.2

LM .

0.4 I --*-0.2 0 0 10 Is central displacement 20 (mm) 25 30 35 (rr-0) 2%) -0-(rr-1

fig. 6.3.5(c)SlabSM5: Load-displacement relationship


1.4

1.2

0 c Im a 0.8
0.6

40

0.4

0.2

-D-rr-1

2%J

0 0.00

0.50

1.00

1.50 $train/ylold

2.00 Wain

2.50

3.00

3.50

fig. 6.3.5(d) Slab SM5: Bottom Steel Strains at centre


1.4

1.2

0.8

0.4

0.2

0-0 0.00

ii
1.00 2.00 3.00

4.00 strain/ylold

5.00 $train

8.00

7.00

8.00

9.00

fig. 6.3.5(e) Slab SM5: Top Steel Strains near comer

217

Chapter6

SlabDesign

1.2

0.8

0.6

0.4

0.2

(rr--O) -0-- (rr=12%) 0


11iI RY/R

0 001234567

fig. 6.3.5(f)SlabSM5: PrincipalMoment-curvature relationshipat centre

1.2

0.8

0.6

0.4

0.2

(rr--O) (rr= 12%)

Li 0

0.2

0.4

0.6

0.8

1.2

1.4

1.6

applied load/design load

fig. 6.3.5(g)SlabSM5: PrincipalMoment at centrevs. Load

218

Chapter 6

Slab Design

1.2-

0.8--

0.6--

0.4--

0.2-

(rr=O) (rr=l 2%)

0 02468 Ry/R 10 12 14 16

fig. 6.3.5(h) Slab SM5: Moment-curvature relationship at comer support


1.2 11

0.8

0.6

0.4

0.2

(rr=O) --D- (rr=l 2%)l

n
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 applied load/design load

fig. 6.3.5(i)SlabSM5: Momentsat cornersupport

219

Chapter 6

Slab Design

6.3.6 Slab SM6


to the slab edgeshasbeenconsidered. For all the designsso far, only steelorthogonal
In some cases,this may not be the most efficient orientation for the steel layout. In

diagonal SM5, the the moments are orientated along main at principal maximum slab
fact, 6.2.4). In SM5 by (fig. 45' this to the view of slab was redesigned x-axis roughly from 45' horizontal. the transformed providing orthogonal steel

Steel Volumes
(CM)

(rr--O)

(rr=12%)

Total(12%) Total(O)

A. bottom Numerical Provided 48.2 69.1

A. top 5.4 8.4

Total

A, bottom

A. top 12.2 15.6

Total

53.6 77.5

41.4 64.1

53.6 179.7

1.0 1.03

In this case,the numericalsteelareas at rr--Oand 12%are equal,which is in contrasts increase The SM5 was observed a'20% where at rr--12%. resulting slab with in figure is 6.3.6(a). layout The load-displacement shown numerical reinforcement 6.3.6(b). in figure An load l-IPd is for ultimate shown of was obtained relationship
this design. The service deflection limit was reached at around 0.7Pd- Yielding of the 9Pd- Yielding bottom steel first occurred at the centre of the slab at a load level of O, began load level, (fig. 6.3.6c-d). the the top comer support at the at same steel of Inspection of the moment curvature relationship at the comer support shows that 6.3.6e-f). just 1.05Pd, (fig. This is due to the fact that once after occurs softening

top to this SM5 the commences prior steel softening, of unlike slab where yielding
in directions, is is x and y redistribution provided of stress steel more difficult equal

is less provided orthogonalto the main steel. At the centre of the steel since much beyond design is load is the the sustained and softening moment not present, slab, (fig.6.3.6g-h).

220

Chapter 6

Slab Design

32.5crs

32.5crs

195crs

195crs

(i) A, at bottom (ff--12%)

65crs

22crs
I+

22crs

65crs

(ii) A, at top (rr--12%) 6mm. in diameter bars Layout, Steel SM6: Slab all sizes 6.3.6(a) mm, all fig.

221

Chapter6

SlabDesign

0.00
(3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 1.15 (3.76) 0.00 (3.76)

0.00
(3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 1.15 (3.76)

0.00
(3.76) 0.00 (3.76) 0.00 (3.76) 9.45 (14.14) 3.32 (3.76) 0.00 (3.76) 0.00 (3.76)

0.00
(3.76) 0.00 (3.76) 24.77 (28.27) 19.52 (28.27) 9.45 (14.14) 0.00 (3.76) 0.00 (3.76)

35.76
(42.41) 43.95 (42.41) 37.03 (42.41) 24.77 (28.27) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76)

30.70
(42.41) 30.11 (42.41) 43.95 (42.41) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76)

31.70
(42.41) 30.71 (42.41) 35.77 (42.41) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76)

(iii) A,,, at bottom (rr--12%) 6mm bars

0.00 (3.76) 0 (3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 2.43 (3.76) 0.79 (3.76)

0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 2.43 (3.76)

0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 0.89 (3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76)

0.00 (3.76) 0.00 (3.76) 1.83 (3.76) 1.23 (3.76) 0.89 (3.76) 0.00 (3.76) 0.00 (3.76)

29.94 (28.27) 18.27 (28.27) 4.49 (3.76) 1.83 (3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76)

28.67 (28.27) 18.64 (28.27) 18.27 (28.27) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76)

34.30 (28.27) 28.68 (28.27) 29.95 (28.27) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76)

(iv) A, y at bottom (rr--12%) 6mm bars fig. 6.3.6(a) Slab SM6: Numerical & (Provided) Steel Areas *2 in mm

222

Chaptcr 6

Stab Design

0.00

0.00

0.00

0.00

0.00

0.00

0.00

(0.00) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00)


0.00 (0.00) 19.60
(28.27) 104.50

(0.00) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00)


5.18 (0.00) 12.05
(28.27) 19.59

00) (0., 0.00 (0-00) 0.00 (0-00) 0.00 (0.00)


0.00 (0.00) 5.17
(0.00) 0.00

(0-00) 0.00 (0-00) 0.00 (0.00) 0.00 (0.00)


0.00 (0.00) 0.00
(0.00) 0.00

(0-00) 0.00 (0.00) 0.00 (0.00) 0.00 (0-00)


0.00 (0.00) 0.00
(0.00) 0.00

(0.00) 0.00 (0-00) 0.00 (0.00) 0.00 (0.00)


0.00 (0.00) 0.00
(0.00) 0.00

(0.00) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00)


0.00 (0.00) 0.00
(0.00) 0.00

(113.08)

(28.27)

(0.00)

(0-00)

(0.00)

(0.00)

(0.00)

(v) A,,, at top (rr--12%)6mm bars


0.00 0.00 0.00 0.00 0.00

0.00

0.00

(0.00) 0.00
(0.00) 0.00 (0.00) 0.00 (0.00)
0.00 (0.00) 3.67 (14.14) 11.81
(14.14)

(0.00)
0.00

(0-00)
0.00

(0-00)
0.00

(0.00)
0.00

(0.00)
0.00

(0-00)
0.00

(0.00) 0.00 (0.00) 0.00 (0-00)


0.68 (0.00) 0.94 (0-00) 3.67
(14.14)

(0.00) 0.00 (0.00) 0.00 (0.00)


0.00 (0.00) 0.68 (0.00) 0.00
(0.00)

(0.00) 0.00 (0.00) 0.00 (0.00)


0.00 (0-00) 0.00 (0.00) 0.00
(0-00)

(0.00) 0.00 (0.00) 0.00 (0-00)


0.00 (0-00) 0.00 (0-00) 0.00
(0-00)

(0.00) 0.00 (0.00) 0.00 (0.00)


0.00 (0.00) 0.00 (0-00) 0.00
(0.00)

(0.00) 0.00 (0.00) 0.00 (0.00)


0.00 (0.00) 0.00 (0.00) 0.00
(0.00)

(vi) A, at top (rr--12%) 6mm bars y

fig. 6.3.6(a)SlabSM6: Numerical& (Provided)SteelAreasin mm

223

Chapter 6

Slab Design

1.2

0.8
0.6

0.4

0.2

20/6 1

0 0 10 20 central displacern 30 ant (rn rn) 40 so

fig.6.3.6(b)SlabSM6: Load-displacement at centre


1.2

2) .! 0.6

CL 40

0.4

0.2

-*-rr--12%

04 5 10 straln/yield 15 strain 20 25

fig. 6.3.6(c) Slab SM6 bottom steel strains at centre


1.2

0.8 c 2 0
0 13 lp 0.6

0.4

0.2

I"
04iii--------0 0.2 0.4 0.6 0.8 strain 1 1.2 strainlyield

rr=12%
1.4

1.6

fig. 6.3.6(d) Slab SM6: Top steel strains at comer support

224

Chapter 6

Slab Design

1 -0.9 0.8 0.7-0.6-0.5 0.4 0.3-0.2-I* 0.1 rr--l 2% 1

0.00

1.00

2.00

3.00 Ry/R

4.00

5.00

6.00

fig.6.3.6(e)SlabSM6: Principalmoments nearcomer


1.00 0.90 0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 0.00

I-0-

Ea, = In, =-:


1.00 1.20

0.20

0.40

0.60

0.80

applied load/. design load

fig. 6.3.6(f) SlabSM6: Principalmomentsat comervs. Load

225

Chapter 6

Slab Design

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 I 0.1 0 10 Ry/R 12 14 -0--rr=12% I

fig. 6.3.6(g)SlabSM6: PrincipalMomentsat centre


1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1

0.2

0.4

0.6

0.8

1.2

applied load/design load

fig. 6.3.6(h) Slab SM6: Principal moments at centre vs. Load

226

Chapter 6

Slab Design

6.4

Adapting Load Path for Design

load in direct in the the to In somecases, the designer main slab wish paths order may illustrated is in in This layout. the applicable case especially to make a more practical in Slab SM5 at rr--30%,wherethe natureof the evolvedmeshled to a large increase
load Given designer from the a pre-determined path, numerical. can provided steel from be 'removed' hence the to the mesh and not generate the elements stipulate load In SM7, from layout the the case path. of slab custornised using a steel required beams is to the slab edges, a system of orthogonal crossing the centre of steel which beams The the supporting onto around the edges and spanning was envisaged. slab of

in for figure (6.4.1 this are shown a). plots moment principal resulting
[ Steel Steel Volumes
(CM)

(rr--O)

(rr--15%)

Total(15%)
Total(O)

A. bottom Numerical Provided PrOvi( 48.2 69.1

A. top 5.4 8.4

Total

A. bottom

A. top 15.9 17.2

Total

53.6 77.5

46.3 66.7

62.2 839

1.1 1.1

for a quarterof the slab areshownin figure (6.4.1b).Comparisons Numerical areas of


detailed in figures the (6.4.1c). steel areas over mesh provided are and numerical Observation of the total steel volumes shows that a similar amount of steel was required at rr--O and at rr--15%.

The numerical analysis of this slab resulted in an ultimate load of around I Pd. The -I

limit deflection 0.65Pd, (fig6.4. I d). Yielding was reached at service around mid-span load bottom 0.8Pd. No the near the occurred centre at steel a of yielding was of inspection From in the top the of momentcurvaturerelationshipat the steel. observed is being be it beyond the that moment seen the design load, sustained can centre,
(fig. 6.4. lf-g).

227

Chapter 6

Slab Design

(i) rr= 10%

(il) rr=15%
fig. 6.4. I (a) Slab SM7, principal moments

228

Chapter 6

Slab Design

30

i) A,, bottom (rr=O)

ii) A,, bottom (rr= 15%)

iii) A,, top (rr=O)

ci

i v) A,, top (rr= 15%)

2) Quarter, Numerical Areas Steel (rnm Symmetrical fig. 6.4. I (b) Slab SM7,

229

Chapter 6

Slab Design

4---.

---

22crs

120

I 30crs

32.5

FT

32.5

130crs

120

22crs

(i) A, at bottom (rr= 15 %)

65

65

22crs

bb 22crs

65

65

i) A, at top (rr= 15%)


fig. 6.4. I (c) Slab SM7, Steel layout, All sizes in mrn, all bars 6mm diameter

230

Chapter 6

Slab Design

0.00 (3.76) 3.78 (3.76) 5.64 (3.76) 5.55 (3.76) 4.93 (3.76) 2.13 (3.76) 0.00 (3.76) 1

3.89 (3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 1

11.08 (28.27) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 4.19 (14.14) 1

21.16 (28.27) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 14.96 (28.27) 1

34.05 (28.27) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 26.53 (28.27) 1

57.21 (56.54) 0.00 (14.14) 0.00 (3.76) 0.00 (3.76) 0.00 (3.76) 0.00 (14.14) 37.91 (28.27)

49.88 (56.54) 22.25 (14.14) 3.09 (3.76) 1.42 (3.76) 3.11 (3.76) 15.58 (14.14) 27.68 (28.27)

(iii) A,,, at bottom(rr--15%)6 mm bars


3.17 (0.00)
1.50 (0.00)

3.96 (0.00)
0.00 (0.00)

0.00 (0.00)
0.00 (0.00)

0.00 (0.00)
0.00 (0.00)

0.00 (0.00)
0.00 (0-00)

0.00 (0.00)
0.00 (0.00)

0.00 (0.00)
0.00 (0.00)

1.53 (0.00) 3.77 (0.00) 5.79 (0.00) 7.5-5 (28.27) 70.54


(84.81) I

0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.00 (0-00) 17.05


(28.27) I

0.00 (0-00) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 7.11


(0.00) I

0.00 (0-00) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.00


(0.00)

0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.00


(0.00)

0.00 (0.00) 0.00 (0.00) 0.00 (0-00) 0.00 (0.00) 0.00


(0.00) I

0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.00


(0.00)

(iv) A,,, at top (rr--15%)6mm bars


fig. 6.4.1(c) Slab SM7: Numerical & (Provided) Steel Areas in MM2

231

Chapter 6

Slab Design

1.2

0.8
0.6

0.4

0.2

16

rr- 15
60 80

0
0 20 central 40 displacement (mm)

fig. 6.4.I (d) SlabSM7: Load-displacement relationship

1.2

0.8 .0 V
0 "a
CL m m

0.6

0.4

0.2

I --*05 steel strain/yield 10 strain 15

fig. 6.4.I (e) SlabSM7: bottom steelstrainsat centre

232

Chapter 6

Slab Design

1 -0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 IR y/R 10 12 14

fig. 6.4. I (f) Slab SM7: Principal Moment-curvature relationship at centre

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 I 0.1 0 0 0.2 0.4 0.6 0.8 load 1 1.2 applied load/design --*E/6] rr- 15-

fig. 6.4.1(g)SlabSM7: PrincipalMoment at centrevs. designload

233

Chapter 6

Slab Design

6.5

Conclusions

Through the visualisationprocess,a deviation from the original elastic pattern of i. is important in It is the a of stresses e., redistribution occurs. slab, stresses created by a loss of ductility in the resulting slab that this redistributionis not accompanied in table 6.5.1. design. The resultsfrom all the slabstestedin the seriesare presented All slab designswere able to attainthe designload readily. In most cases,an overall increasein the areaof steelresultingfrom the evolvedmeshover that resultingfrom the mesh at rr--O was observed.An increasein the provided steel areasfrom the in order to generate was often necessary areas a practical steel layout steel numerical increase This fulfil to was particularly pronounced steel requirements. minimum and in slabswherethe visualisationprocess resultedin a large % area 'removal' such as Orientation direction SM5. SM3 the to the of steel coincide with main and of slabs in in SM6. the a more efficient volume of results steel as case moments of principal leadsto in mostcases, Increasingprovidedsteelareas a greaterultimate load from the designsat rr--O. All the designsresulting from the evolved mesh maintain ductility. In all casesthe in However, increases in ductile failed in as observed some cases, manner. a slabs ductility in localised loss The of cause a areas. may occurrenceof areas steel in in but the concrete this was at loads was evident softening some cases compression in I Pd that region. in excess of and as a result yielding of steel Of -I
When considering the serviceability of the slabs resulting from the evolved meshes,an However, in achieved. was some slabs, such as SM3 and SM4, adequateperformance the steel layouts derived from the evolved mesh were often more complicated and in localised For lead it to this of steel areas. congestion reason, may concluded could that it is not desirable to use the visualisation process in every case.

The greatest potential advantage of this design process comes from being able to to the according process visualisation a pre-determined path. This allows the control designer to take account of experience and practical considerations, in order to layouts. steel generatepractical

234

Chapter 6

slah I)csign

As shown in slab SM7, this createsan efficient reinforcement pattern and is able to ductility. while maintaining of stresses re-distribution accommodate Property
(rr) 0

SMI
40% 0

SM2 30% 24c1c 1638.5


2061.7

SM3 0 425
564,8 481 740.5 MKI(

% area'removed' I Numericalsteel (CM) volume I Providedsteel 3) volume(cm PJPII Servicedeflection limit (span/250) a)First yielding of bottom steel Location of a)
b)First yielding of top steel Ecation ofb

754 980.9 1.2 0,65P, j


1 Pd -0

24% 807 990.7 1.2 0.65 P, j 1.0P', centre


1.2 P,j L cso=rner of I. ort s ort su i-

1588.9 1815.8 1.0 0.7 P, j 0.7 Pj Centre

1.05 0.71),, 0.7 P, j centre

1.1 0.7 P, j 0.6 P, j


load point
0.61),, I corner C)r1jer of of' -,

1.3
P',

0.7 P,,
10; 1(1 point
0.8 P,I ()ril,,

Centre
1.2 Pj I ri)f 11 ort o rt su

table 6.5.1: Summary

of results from test progrant

support I

Property (rr) % area'removed' I Numericalsteel (cm3) volume I Providedsteel volume(cm3) /P,, P,, Servicedeflection jimit (span/250) a)First yielding of bottom steel Location of a) b)First yielding of to stee Location of b) 0 2730

SM4 12% 38% 4719 6562.4 1.3 0.75 P(I 1.2Pj centre 1-1Pd column support 0.7 Pj 1.2P, j centre 1.0P', 53.5 77.4 1.1 0

SM5 1 571/ 0 .8 102.3 1.35 0.7 P, j 1 centre ().8 P',

SM6 121/, 571/( 53.6 79.7 11

SM7 151/, 511/( 62.1 83.9 1.1

4587.4

0.65 P, j 0.8 P', centre 0-8 Pd

5 P', Centre 0.911"1

0.81"d Centre

tame 0. -'n. I (conta): Nummary ol results from test progrant

-cor Column corner ner corner SLI2201-t support support , st port

235

Chapter 7

Experimental Program

Chapter 7 Experimental Program


7.1 Introduction

Five structuresdesignedusing the strut-tie method were tested physically in the laboratory.Two double sidedcorbels,one single sided corbel and two comer joints design from 7.1. The these table tested, of structures and results numerical see were This in describes detailed the the materials tests chapter. chapter next are andphysical in the testing. instrumentation the as well as models method of physical used and

7.2

Preparation of Models

7.2.1 Formwork
The sameformwork wasusedfor eachof the threeCorbels made.For eachdesign,the in orderto accommodate in geometry.Sincethe formwork waseasilyadapted changes identical joints geometry,only one mould was neededhere.The two comer were of 20mm, formwork For body thick the comprised of plywood of panels. stability main battens fixed 50x5Omm intervals thick timber were at regular strength, along the and length of the mould. Prior to castingof the model,the walls of the mould werecoated from in the to sticking with oil order prevent concrete

7.2.2 Concrete
For all the specimenstested, Rapid Hardening Portland Cement was used to give the fu. A 7-day 10mm maximum aggregate cube strength size of was used. The required

in batches took a number place of model of each and was properly compacted casting through vibration.

236

Chapter7

Experimental Program

Model
Pd

Schematic
Pd

50.1 CorbelC2A 300 350 thickness=250 200


Pd 250 200 Pd

1 50 10 50 CorbelC3A 150 '0 350 thickness=250 200 250 200,1 Ip , : L't 1 0 350 150 150 350 250 200 750 CornerJoints FJIA & FJ2B -; r 750 J_ L _ 150 11, 150T thickness=150 1
Pd

CorbelC4A

thickness=150

Table 7.1: Details of Experimental Program, all sizes in mm.

237

Chapter7

ExperimentdProgram

four IOOxIOOxIOOmm In addition to the main specimen, cubesand four MOW= These were cured with the main specimen control specimens cast. were cylinders half After hours. for 24 the control specimenswere this period, under polythene in with the main specimen placedunderwater the curingtank,while the restremained determine The hessian. to the cube strengthand the cylinders used were cubes under determine The tensile the to strength and modulus of elasticity. concrete used were usingthe cylinder splitting test from the following: tensile strengthft wascalculated
2P 7cDL

ft =

diameter (300mm), D=cylinder length (150mm), L--cylinder P is the and where from The load. each control specimenwere averagedto give the results ultimate in valuespresented the next chapter. experimental

7.2.3 Reinforcing Steel


high yield deformedbarswereusedfrom the samebatches Throughoutthe test series, bars. 012 10 8mrn, and of The yield stressof the bars from each batch was

Universal Olsen Testing Tinius Machine fitted the using with an S-type measured The bars the yield stress of extensiometer. wastakenasthe stressat which a electronic line startingfrom 0.2% strain,parallel to the initial slope of the curve, intersectsthe 7.2.3. figure see curve, stress-strain

7.2.4 Strain Gauges


During the experiments,strains in the bars were measuredusing 6mm electrical The fixed to the steel bars at predetermined gauges were strain gauges. resistance history. Before fixing in the to the gauges, the surfaceof the strain record order points first by filing and then smoothingwith sand bars at the appropriate areawas prepared during Care this process to avoid removing a significant areaof the taken was paper. bar which could causeweakening.The smoothedsurfacewas then treatedwith MM-prep in and neutralizer order to remove any dirt and greasein conditioner prep bond The the to a perfect with achieve steel. gaugewas then fixed to the bar order

238

Chapter 7

Experimental Program

for In operational readiness using a voltmeter checked adhesive and order to using during damage fabrication from and moisture and casting, an epoxy provide protection At terminal to the and gauge areas. each required position, coating was applied resin two gauges,each diametrically opposite one another, were fixed. The measured strain between the the gaugepairing. taken average as was position at each
600
I

500

400
04

E 300 U) 200

100

0 0

Relationship fig.7.2.3Stress-Strain for 12rnmbar

0.002

0.004,

strain (mm/mm)

0.006

0.008

0.01

7.3

Experimental Procedure

fully it When the specimen cured, was waspaintedwhite to assistin identification of Each during into it's respectivetesting rig testing. specimen was manoeuvred cracks by crane.All the strain gaugesand load cells were connectedto a 3530 Orion data
logger for automatic recording. Each connection was checked prior to testing. Strain defective disconnected. At each load increment, the checked and ones were gauges

for later disc The to processing. stored were specimen results was illuminated using a in identification light to source order ease of cracks. The crack widths were powerful
increment load each using a micro-crack reader. The crack development at measured ink load increment. traced This procedure was repeated at an marker at each with was load increment the until ultimate was achieved. Details of the test rig for each each described. are now model

239

Chapter 7

Experimental Program

7.3.1 Double Sided Corbels,


in the past on double sidedcorbels,seeKriz Experimentaltestshavebeenconducted & Raths(1965), Somerville(1974),Mattock et al. (1976).In thesetests, the corbels load the the corbel ends, and at was applied supported were orientatedupside-down, in This testing this work, see figures the of used the method was to column end. Losenhausen UniversalTesting 7.3.1(a-b).The loadingwas appliedusing a 100OOkN Machine.In order to ensureevendistributionof the load, eachside of the corbel was At 500 bearing kN load by the supports, on rollers. plates cells were steel supported layouts Details the the of steel to and strain gaugepositions reactions. placed measure in the next chapter. given are

7.3.2 Single Sided Corbel Testingof a singlesidedcorbelpresents moredifficultiesthandoublesidedcorbels from rotating due to the because the ends of the corbel have to be prevented C4A in Tinius Olsen load. In tested this the the corbel was program, of eccentricity The test set-upis shownin figure 7.3.2(a). The testing UniversalTestingMachine. from by the the to ends of column prevent rotation as providinga such methodwas bottom. The block 10kN the tested top the at tie model was using a steel and at steel Machine fixed head Theloading increments. of theTiniusOlsen was anda hydraulic between thehead jack and500kNloadcell wereplaced andthecorbel.

7.3.3 Corner Joints


Comer joint MA for designed a closing moment and FJ2B was designed for an was identical joints Both Details for the test were of geometry. moment. of set-up opening in 7.3.3(a-d). figures The leg joint vertical are given of each comer was fixed to each diameter in 30mm floor lab the figure. For FJIA, the steel rod as shown via a the in hydraulic jack the the via comer was generated at the end of the closing moment horizontal. The moment was then calculated as the product of the load and the lever For FJ2B, the distance the to the opening moment was of vertical. centre arm jacking by in horizontal the the in figure upwards against corbel as shown generated 7.3.3(c). In order to monitor rotation at the base of the comers, displacement both In cases, negligible rotations were observed during used. were transducers

testing.

240

Chapter 7

Experimental Program

2P
1'Top loadliq-, Platcli ()I Losenhausen Machinc

bottom platen of Losenhausen Machine, fixed in horizontal and directions vertical

75x25Ox25mm steel baseplates on rollers 50OkN Load Cell!

N
I1

Pl

fig. 7.3. I (a) Testing Method for Corbels C2A and C3 A

fig. 7.3. I (b) Corbel C3A: Experimental Set-up

241

Chapter 7

Expenniental Program

It'

hloA It) picvctil Iiding ,

fig. 7.3.2(a) Corbel C4A: Testing Method

fig. 7.3.2(b) Corbel C4A: Experimental Set-up

242

Chapter 7

Experimental Program

i 9 -E

bolted to tab floor 625 IOOXIOOXIO hollow sections hydraulic. jack & load cell --.,k 50

1080

steel -10111111 rod

41 <1

bolted to lab floor

III]

11511

fixed to lah floor i) Plan view i) Section A-A

fig-7.3.3(a) Comer Joint FJ I A: Testing Method, al I sizes in nini

fig. 7.3.3(b) Corner HIA:

Expei-miental Set-up

243

Chapter 7

Expeninclital Program

bolted to lab floor


625 50

IOOXIOOXIO hollow sections A


1080

50x 150 steel hearing plate on roller \

loadcell hYdraulic jack

bolted to lab floor

i) Plan view

i) Section A-A

fig. 7.3.3(c) Corner FJ2B: Testing Method, all sizes in mIll.

't,

/I

im

fig. 7.3.3(d) Corner FJ2B: Experimental Set-Lip

44

Chaptcr 8

Strut & Tic Design

Chapter 8 Strut and Tie Design


8.1 Introduction

This chapter details the design by strut and tie models of a number oftypical D-region type structures. The main objective of this work is to assess(lie SUItabIlItyOfthe Struttie method in achieving the required performance from a designed Sti-LICtUrC.

The visualisation process described previously is used to generate tile Outline 01'tile designs, deep I beam, The 3 2 test comprises of series six models. corbels and strut-tie frame corner joints. All of these designs were assessed through non-linear analysis. in joints laboratory. tested the to this, tile were physically and corner corbels in addition Details of the experimental set-up were given in chapter 7.

As in the case of slabs, the service and ultimate load characteristics of- the Sti-LICIIII-CS, is important It from investigated. that the structures resulting the strut-tie desjgn were load. This ductile by at ultimate Is achieved response a ensuring that crushing possess of the concrete prior to yielding of the main steel is avoided at the design loads. In the displacements beams deep serviceability corbels, and of are not a serious Issue case low the span-depth ratios of these structures results it, very sniall very since deflections. However during the experimental test series, the ina oi- crack widths were service crack width limit of' 0.3nini

rnomtored and compared with the maximum stipulated

by BS81 10. In all designs, adequacy of anchorage and 1)()11(1 (11C ()1,

BS81 10. to checked according was reinforcement

245

Chapter 8

Sout & Tie I)Lsign

8.2

Deep Beam BI

This model was a typical simply supported deep beam with an effective span-depth in below, Dimensions the table 1.7. are properties given material where and ratio of
Pd, f,,,, ft, E,, E, and fy are the design load, uni-axial cube crushing strength of

concrete, uni-axial tensile strength of concrete, elastic modulus of concrete, elastic modulus of steel and yield strength of steel respectively.
Schematic Design Material P,1=250kN concrete: 2 F,,, =35N/m I,, 2 ON/mm 1', =3. E, =21.5kN/mm T 525 525 steel: 1'=460N/mm 2 y 2 Es=200kN/m III thickiiess= I 00nim Properti

500

The initial elastic principal stressesand resulting strut-tie model are given in figure 8.2(a). From figure 8.2a(ii), the main load paths in the structure can be clearly seen its from load the diagonal running point to tile support, all(I tile strut compression the
horizontal tension tie at the bottom. The presence of transverse tensile stress along the

length of the strut can also be observed. These transverse tensile stresses havc it detrimental effect on the strength of the strut and are accounted for in tile j-c.,,,, jltjjjg by introducing (fig8.2a(iii)) ties the along strut at the third points. model strut-tie Using this model, the member stresseswere evaluated and the resulting reinforccinent is in A below. the For diniens,(),11jig table of analysis given summary was calculated. t',,, the design cube crushing strength (lie zones, was nodal used and as struts of factors be fC,,. Similarly, fcl. In to would applied material practice, no matchal strength factors were applied to the yield strength of the steel. The tensile strength ()f tile concrete was ignored.

246

Chapter8

Strut& Tic Design

Member 1

Force (kN)

A. required
(MM2)

A, provided
(MM2)

No. of Bars

A, provided A, required

4 7

125 53.52 53.07

271.74 133.73 133.73

314.16 157.08 157.08

4(010) 2(010) 2(010)

1.15 1.17 1.17

Nodal zones are the critical areasof the model and stressesmust be checked to ensure is Figure details 8.2(b) the not exceeded. gives of stress state given that capacity under 6, fig (see 8.2a), load is At dimensions. the where applied, node of the nodal zone

in bi-axial in this is the capacity of concrete zone must not compression, concrete in 5, biIn the concrete chapter strength under mentioned as reality, exceed f,,,, 1.16 Kupfer (1969). The increase to al. et stressat around can axial compression bearing dimensions) is hence / At (Pd N/mm2 12.5 is to and plate safe. this zone equal horizontal diagonal tie the 1, and the vertical strength reaction meet strut, where node in due is the tie, this the to the of case presence stresses must not the reduced node of 2. is in 6, Since N/mm, 26.25 this the the the 0.75fc,, node same as node stress = exceed I 6 is in figure 8.2(g). A is and of given nodes representation schematic node safe. is in figure layout 8.2(c). The designed given reinforcement in figures 8.2(d-h). From the the are shown The results of non-linear analysis
figure 8.2(d), it be load-displacement that the can relationship seen numerical 1.32Pd. load In to this, of addition a certain amount of an ultimate attains structure ductility is observed before eventual collapse. If yielding of the steel is assumed to load 1.15Pd. tie failure, the then and model would strut predict an ultimate of govern for the tie. 15% main used was steel extra since As shown in figure 8.2(c), yielding of the main steel occurred at the mid-span at

1.25Pd,which is close to the load predictedby the strut-tie model for the onset of began, Once I the the steel main crushing of of concrete yielding at node yielding. in is in Rd. The bi-axial I. this initiated tensionconcrete zone a state of at was
is 0.8f,, hence The idealisation stress of an ultimate reached. strut-tie of compression, LOPd, 12.5N/mm2, OAPd. that of a stress at would be reached, see this node suggested it finite From 8.2(g). the analysis, was observed that a higher stress was fig element due the bearing to which the concentrations occur stress the around edge of reached

247

Chapter8

Strut& Tie Design

in from figures be 8.2(i-j), However, the the stresses stress plots seen can as plates. by diagonal to those the strut-tie model. the similar predicted more were much along Details of the Gausspoint positions at which the numerical stresseswere obtained are in figure 8.2(h) given

In this example, a more detailed analysis of nodal zone I was carried out. Using the displacements from the original analysis, a displacement controlled analysis using a implemented. Details of the principal stressesand stress states are mesh was refined be form It in figures 8.2(i-j). the principal stress plot that the largest can seen given are concentratedalong the outer edge of the bearing plate. compressive stresses

248

Chapter 8

Strut & Tic IX-sign

1) (1-1-0)

compre "

ii)

(rl)

El

000,

F0 ol 40
n5

Strut &TicModcl III)


5 (x=26.56" 18.44"

...

n,

tic

fig. 8.2(a) Bearn B 1: Elastic Stress Patterns and Strut Tie modcl

249

Chapter 8

Slitil & Tie I)Lsign

2()()x 1()()X-' hcll llll, 111,11c


(T 1.1

120mm (Tjj,, (jcj<-0.75


I 00x I OOx25 bearing plate

fig. 8.2(b) Beam B 1: Dimensioning of Nodal Zones

web reinforcement in the form oflinks to provide 30 70

, q)K

Igo

120
1010

60 '30
I-III0d IT -11-II-, bid 4+04

2(1()

40

80

90

235

80

SO

235

90

10 .

fig. 8.2(c) Beam BI Reinforcement Details, al sizes in nirn, all NtirrUI)S06

250

Chapter8

Strut& Tie Design

1.4 1.2
o 19 i

to a 0.6

0.4 0.2 0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 central displacement (mm)

fig. 8.2(d)BeamB 1: Load vs CentralDisplacement


1.4 1.2

.2 0.8 2 0.6 (P
Ck. ck.

0.4 0.2 0 0 0.2 0.4 0.6 0.8 strain 1 1.2 1.4 1.6 strai*ield

L:
1--o-numed7cal I

fig. 8.2(e) Beam B 1: Main Steel Strains


1.4 1.2

0.8 2 0.6
cL 0.4 a
CL

0.2 num 0 0.05 0.1 straiWeld strain 0.15 0.2

fig. 8.2(f) Beam B 1: Web Steel Strains

251

Chapter 8

Stim &, 'Fie DLsign

2 1 1AN/rum (5=

\1551.6ninri

i) Node I

tie force
converted to compressive stress via 45"

ahcorage
1112 IOAN/rn cy=

-. 0
--p
--f

i2mflm
4F

12.5N/nim cF=

(T=I 2.5N/mm2

Nodc 6 ii)

45"

4 50

(F=I 1AN/nim

cr= I IAN/nilll

fig. 8.2(g) Beam B 1: Stresses in Nodes I and 6

I 4, -F D (303.1,343.7)

(490.6,492.9)

500 B (94.4,242.9)

E (415.8,305.4)

C (221.8,117.9) 1 A(109.2,7.0)

525 for Numerical

f ig. 8.2(h) Beam B 1: Gauss Point Positions

Stresses

252

Chapter 8

s(rul & TIC Dclgll

0.8

0.7

Al

:CB
0.6

0.5
U

0.4 U) 0.3

"

("

"

0.2

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


0.1

Li

0.2

0.4

0.6

0.8 load

1.2

1.4

applied load/design

fig. 8.2(i) Beam B 1: Numerical Compressive Stresses


1.2

1
-O-E LT
0.8

0.6
C) U)

0.4

D
0.2
_____

E
0.6 0.8 load 1 1.2 1.4

0
0 0.2 0.4 applied load/design

fig. 8.20) Beam B 1: Numerical Compressive Stresses

253

Chapter 8

Strut & Tic Ocsipii

IK - /v''4
-, '\I\iI 1 //i -' '

-7 ---

500 111111

150
It )5Uiu tu

compression tension

100111111 fig. 8.2(k) Beam BI Nodal Zone 1, Principal Stressesat Lilt"Ilate load

Stress-Statc
tens i on-com pressi on approaching peak tension-compresmon approaching crushing

fig. 8.2(i) Beam BI Nodal Zone 1, Principal Stresses at ultinijite load

254

Chapter 9

Strut & Tie Dcij! n

8.3

Corbel C2A

This model was a symmetrical two sided squarecorbel. Both numerical and physical dimensions The C2A and material properties are presente(I testing of was carried out. in the table below.
[[ r

em :S-c Schematic h: at ic
Pd

Pd

Material Properties (Design) P,1=350kN concrete: 2 f,,,=50N/mi-n 2 f,=3. ON/i-n111 E, =21.5kN/m 1112 steel: fy=50ON/nim 2 E, =200kN/mn 12

Material Properties (Experimental)

II 5; 0) ( 0 3(0)(0 concrete: I'C, =49.8N/mm2 I't=3. I N/mm 2 F, '"1.61AN/mm2 =steel: 1', 12) (0 =504N/mm" E, =201.6kN/mIII-' t' =509N/mm 2 (08) y Es=214.5kN/mm 2

35() ! f250 200 200 1( AI( 1 9: r

thickness=250mm

8.3.1 Strut-Tie Model and Design


The elastic principal stress fields and resulting strut-tie niodel arc pi-cwnted In hgUrC 8.3(a). The strut-tie model which results from Visualisatloll, consists of' a horizontal first by is by Niedenhoff diagonal that to ( 1901) similar proposed and strut, tie and a
The (1983). by Hagberg developed later critical areas with rcgard to concrete and

in I load f1cre the the node and tile stressesmust strut point, Itself. at are stresses less than or equal to 0.75fcdbecauseof the presenceof the tic. ']'he geometry of' the
diagonal strut can be determined fi-om the geornetry of' the baseplate. In tills case, tile 2p)t], Pd/[(W. is in bclrillg tO the the COS where is equal vi, width strut ol'the pjjtc, stress

P is the angle of the diagonal strut to the vertical, t is the thickticss of the strut. III tills leads P 250mm, is 30' is 75mm, t to is which and a maxinium stre.,, w s ()t, case, 0.75f,,,. At node 2, where the diagonal strut Meek IIIC COILIIIIII, a state _ hence is in the bi-axial this and created stresses compression node IIILIStnot CXCcc(j of If,,,. The column is designed to carry load in excess of 5 times the corhel loadim, 24.9N/m M
2

to checkconcrete capacityin nodalzone2 clueto the level ()tHenceit is not necessary


reinforcement provided.

255

Chapter 8

strill & TIC Design

compression tension

i) rr=0

IOOXI hcarlll plat

75 1_1

n2 ja amo
II

III) Strut-tic model

i\)

fig. 8.3(a) Corbel C2A: Elastic Stress Paths and Strut-tie illo(tel

256

Chapter8

Strut& Tic Design

As can be seenfrom the elasticstress are spreadover a certain plot, the tensilestresses depth at the top of the corbel. It is thereforesensiblefrom a serviceability point of depth. In distribute for the the to provision a certain of over reinforcement steel view the tie forces, the steel will be madeup of the main bars and lower reinforcement distributed throughoutthe top half of the corbel. Each of these contribute to the Both the primary and secondary tensile reinforcementcontribute to strength. overall tensile strength by each carrying a proportion of the tie force. As a means of deriving force for primary and the the tie corresponding proportion of empirically following is the procedure reinforcement, adopted. secondary The resultant of the main steel acts at the top of the corbel and the resultantof the by lower reinforcementacts at the end of the top third of the corbel, as suggested Hagberg,seefigure 8.3(b).This resultsin the following designequationderivedfrom the staticsof the truss: Pd = tanp, + tanP2

Where Ady and Ady are the tensile force capacity of the primary and secondary P2 P, 40' 30' For 312 bars the are and respectively. and main steel, reinforcements, form in bars 28 the of stirrups were provided as the lower and used were leads This load 353.73kN to theoretical a ultimate of > 350 kN, Pd. reinforcement. the reinforcement layout in the designedstructure is shown in figure 8.3(d).
Pd A,, fy
Jy

411

F,1717
* 0-2-"; Fc2

jo

fig. 8.3(b) Dimensioning of Ties

257

Chapter 8

Strut & Tic Design

8.3.2 Numerical and Physical Testing


The experimentalultimate During testing, the load was appliedin 25kN increments. which comparedwell with total load of the corbel was recordedas 827kN, (1.18Pd) SORN, (1.15Pd). As load total of with all subsequent the numerical ultimate The first the material strengths were used. cracks experimental numerical analysis, in junction, (0.18Pd) URN, the tie-column the corbel, at upper see at occurred fig. 8.3(e). The largest of these initial cracks was 0.05mm wide. These cracks load At 375kN, (0.53 Pd), the towards column. a of up propagated new gradually bearing diagonally formed the toward the centre running plate, upwards around cracks 0.15mm The limit largest The these was cracks wide. service the of crack column. of in junction first 0.3mm the the tie-column cracks around at a reached was width of load of MOM (0-78Pd). The steelstrainsobtainedfrom the gauges areplotted togetherwith the corresponding initiated by in Failure fig. 8.3(g-i). the corbel was yielding of the see numerical values in lower (I. IPd). The began 770kN, the tension reinforcement zone also at steel main A fig. 8.3(i). this was achievedwith the numerical good agreement point, to yield at Subsequent led the tension to straining of main strains. steel steel and experimental be from Demec diagonal in figure the the can seen cracks, as readings of widening 8.3(f). Figures8.30-k) showthe crackpatterns at ultimate load. distributionin nodalzone 1 is represented in figure 8.3(n) and the The idealisedstress for highest the in are shown stresses areas of compressive compression numerical figure 8.3(o-p). As in the strut-tie model, the effective strength of the concrete is 0.95fu, diagonal The the occurred stress, compressive at greatest exceeded. nowhere ,
load. This due junction to the stress concentration was ultimate at strut-column in by the change geometry, i. e. right angle comer. In practice, it is brought about sharp lower for to This the has the approach corbel edge column at an angle. the common junction the the concentration at stress column corbel of reducing and was effect designs C3A and C4A. Along the diagonal strut, the in corbel subsequent utilised 0.55fu, is than greater are no stresses compressive which similar to the strut numerical

tie model predictionof

2 0.5fcu, 24.9N/mm

258

Chapter 8

sil tit & 'I Ic Design

180 50 3012

L 300

--0-

strain gatioc

08 links (a)80 crs 8 links @ 125 crs 350


2025

125 i) elevation

200

2SO li) side clc\,; Ill()Il

fig. 8.3(c) Corbel C2A, all sizes in mni, cover 20nim

"Is'

",-

Wwwwow

fig. 8.3(d) Corbel C2A: Reinforcernctit A: Fornmork

259

cluipicr 8

Strut & Tic I)Lsign

fig. 8.3(e) Corbel C2A Main Crack Patterns

1)CIllecs 100111111 placed al IlliddIc thild, "

0
rcn 55 CI)

tu IIJL

-0.001

0.001

0.002

0.003

0.004

0.005

0.006

strain

fig. 8.3(f) Corbel C2A Demec Readings, +ve. tensile, -vc. compressive

260

Chapter 8

Suill & I'le I )c%lgll

1.2

I 130.8 0 c
10

O6

CL 04 c to

0.2

0.2

OA

0.6

08 strain/yield

11214162 strain

fig. 9.3(g) Corbel C2A: Main Steel Strams


1.4

1.2

0.8

-2 0.6 '0 g 0.4

02

0 0.2 04 06 straln/Vield sirsin

IIj

fig. 9.301) Corbel C2A: Uppcr Stirrup Stccl Strains

V 0

e za 0 '0
CL m

0.1

00102030400H straintyleld Mrsin

09

fig. 8.3(i) Corbel C2A: Lower Stirrup Sicel Strains

261

Chapter 8

siful & VICI)CSIgil

fig. 8.3(j) Corbel C2A at ultimate load (I X)P,j -I

7 /_.

fj,

T
I
Ig-8.3(k) ,iij1.1

6 22

Chapter 9

Strut & 'I'le I

2 N/mm 18.7 (T=

I N/nun

I (H(nun

t 1) N/mm

65m

I'Igg.3(l) Corbel C2A-. Stresscs in Node I

700
650 I (2.19.5,6 'H (, 1 0 . 1) (240.5.5S c (I K5.0,5 (X)

350
el

B( 145.0,40 k 2)

1 120.3.143.2)

(X)

125 fig. 8.3(m) Corbel ('2A: (Yauss 1)()Iilt POsItiolls for Numerical

325 Compresove Stresscs

263

Chapter 8

slitil & I'le Deslyn

1.2

1 08
0.6

I
OA

02

0,1

020304508 displacomOnt (MM)

fig. 8.3(ii) Coi-belC2A: Nuinerical


1 0.9 0.8 0.7 0.6 05 U) 04 0.3 0.2 0.1 00 0
02 04 applied Of, load/design load () H

fig. 8.3(ci) Corbel C2A: Numerical


0.45 0.4

F)
0.35 03 025 02 0.15 0.1 005 0 0 0 04 06 2PPII*d load/design load 08 1

fig. 9.3(p) Corhd C2A Numel-icill Compressive Strcsscs 264

Chapter 8

Strut & Tic Desiyri

8.3.3 Comparison of Strut-Tie Design with Direct Design


As outlined in chapter 3, the direct design procedure can be applied to inplane for Equations proportioning orthogonal and/or skew reinforcement to resist situations. inplane forces were first proposed by Clark (1976). Using these equations, many design deep direct beams investigated have the of and other members. researchers Khaskheli (1989) used elastic stress fields for direct design, while Bensalem (1993) fields. deep beams Bensalern design non-elastic stress using of concluded out carried that the use of elastic stressfields was sufficient for the design of deep beams.

In this section, the direct design of corbel C2A from the elastic stress pattern, (Le described in 3 The chapter was used to derive tile procedure out. rr=O) is carried layout. reinforcernent The performance of the design was then assessed in

design. the to strut-tie comparison

The required steel areas resulting from the direct design procedure are presented in figure 8.3(q). For comparison, the numerical and provided steel areas are shown in figure 8.3(r). As expected, the greatest quantity of steel occurs at the area of highest tensile stress, i. e. the upper corbel-column junction. The quantity of required horizontal steel in the tension zone of the corbel is similar to that provided in the struttie design, seefig. 8.3(s). It can be seenthat in the case of horizontal steel, the direct design method results in a 10% increase in provided steel from the strut-tie design. Secondly, direct design also results in the provision of some vertical steel.

The results from the numerical test are shown in figures 8.3(t-v). Figure 8.3(t) shows that as a result of the increase in provided steel, the direct design corbel attains a higher ultimate load of 1.25Pd. Yielding of the main steel takes place at the upper

column-corbel junction at 1.2PI. The lower horizontal reinforcement in the tension zone approached yield at the ultimate load. No yielding of the vertical reinforcement was observed. In general the behaviour of the direct design corbel was similar to that design the corbel since the quantities of main tension steel were similar. strut-tie of

265

Chapter 8

Strut & Tie Design

rD

r )n considered y

LX
i) A,, (rr=O)

n) A, y (ri-=O)

fig. 8.3(q) Corbel C2A: Numerical Steel Areas Resulting from Direct Design, (11IM2)

266

Chapter 9

Strut & Tie Design

313 (339.3) 128 (157.1) 80 (100.5) 32.5 (100.5) 1.23

240 (339.3) 145 (157.1) 100 (100.5) 46.7 (100.5) 0.98

191 (339.3) 154 (157.1) 106 (100.5) 48.4 (100.5) 0.47

110 (339.3) 96.2 (157.1) 75.7 (100.5) 43.6 (100.5) 3.36

50.2 (339.3) 21.6 (157.1) 14.7 (100.5) 13.4 (100.5) 3.14


region considered

(0.0)

(0.0)

(0.0)
(i) A, (rr=O)

w)

w)

193.5 (157.1) 114.4 (157.1) 59.9 (157.1) 21.9 (157.1) 26.4 (0.0)

91.1 (100.5) 133.5 (100.5) 102 (100.5) 82.2 (100.5) 38.4 (0.0)

59.8 (100.5) 69.7 (100.5) 73.9 (100.5) 78.45 (100.5) 33.8 (0.0)

1.9 (0.0)

1.1

0 (0.0) 7.9 (0-0) 26.5 (0.0) 16.2 (0.0) 0 (0.0) 0.2 (0.0) 2.71 (0.0)

(ii) A,

(rr=O)

fig. 8.3(r) Corbel C2A Numerical and (Provided) Steel Areas (MM2)

267

chaptel

TIC I)CIvll

40crs

50

3012

2010 60 crs 208 208

300

35C

F014 1-1 10

125 i) elevation

200

250 ii) side elevation

fig. 8.3(s) Corbel C2A, Direct Design: Steel LayOLIt


1.

1.2

0.8

-2 0.6 m T 'a
CL

m 0.4

* 0.2 --DC 0 0.1 0.2 0.3 0.4 0.5 0.6

direct design strut-tie design

0.7

0.8

0.9

displacement (mm)

fig. 8.3(t) Corbel C2A: Comparison of Direct Design and Strut-Tie Design

268

Chaplei 8

Sll [I( &, *1 Ic

1.4

1.2

"a

tu 0 cn *ui 0.8 CD 13
m

-2 0.6
12.

m 0.4

Lii
0 numerical I
0 0.5 1 strain/yield strain 1.5

0.2

fig. 8.3(u) Corbel C2A, Direct Design: Main Steel Strains

1.4

1.2

'a m 0
r_

'0

cn U; 0.8 (D

0.6
'a 0.

2)

10 0.4

0.2

numerical

04
0.2 0.4 strain/yield 0.6 strain 0.8

fig. 8.3(v) Corbel C2A, Direct Design: Lower Steel

269

Chapter 8

Strut & Tic Design

8.4

Corbel C3A

Like C2A, this corbel was double sided and symmetrical. The same, thickness, width in lower half for C2A. The depth this of corbel angled was of corbel was used as and junction, the the which to at corbel-column stress concentrations reduce order in Geometric C2A. the given model and material properties are previous occurred in the table below. Schematic
P', Pd

Material Properties (Design)


P,1=250kN

Material Properties (Experimental)

50 150 150

concrete: f,,,=35N/mm 2 fi=ION/im-n 1)


E, =21.5kNhm-n 2 2

concrete: M2 f,,,=37.ONhi1
fi=33Nhimi" E, = I 9.95kN/inin2 steel: I' =508Nhiun 2 (010) y E, =210.35kNhn 1112 2( 08 I'y=509Nhimi E, =214.5kN/iii 1112

350
12-

steel:
fy=50ONhnin

E, =200kNh-nin 2
200 k 1 4 r19 -I250 1,200,1 -I-

thickiiess=250iiim

8.4.1 Strut-Tie Model and Design


Since the corbel height, depth and thickness of the model are similar to that of C2A, the same strut and tie model can be used. Once again, the stressesin the nodal zones Using bearing for C2A, the the maximum stress in the same plates as are checked. diagonal strut is calculated as 17.8 N/mm
Using the equation described
22

is less than 0.75f,,,, (26.25N/mrn which ,


the reinforcement and 406mm arrangement was bars as lower

previously,

calculated

as 3010mm

bars main

reinforcement

reinforcement.

Since no high yield 6mm bars were available during Fabrication,

408mm bars were used as the auxiliary reinforcement. With the new reinforcement layout, the calculated maximum load, assuming yielding of the steel governs failure, is 325kN, ( 1.3Pd).Details of the reinforcement layout are given in figures 8.4(a-b).

270

Chapter 8

Strut & Tie Design

8.4.2 Numerical and Physical Testing


The corbel 25kN C2A, the tested manner as same in was The experimental ultimate ultimate load total with load the load being applied was recorded (13P, I). The The in increments. as 710kN, first cracks

(1.42Pd) occurred

and the numerical at 0.2Pd, around

was 650kN,

the upper corbel-column On subsequent

junction. loading,

largest of these to strut see

0.05mm. as measured was cracks propagate region downwards

these cracks continued in the diagonal wide,

into the column,

see fig. 8.4(c). Cracking

did not occur until Further cracking

0.7Pd. The largest of these cracks was 0.1 illin in the diagonal strut region was observed

fig. 8.4(d). increments.

on successive

The service crack limit junction.

width of 0.3rnrn was reached at I -OPd, in cracks lit Failure of the corbel was initiated junction. The numerical by yielding ofthe steel led to

the upper column-corbel

main steel at the upper column-corbel strains show a good comparison significant widening

and experimental

figures 8.4(e-g). Tills yielding are shown and in

of the cracks both at this point and in the diagonal fig. 8.4(h). The corbel at ultimate

strut, as call be in

seen frorn the Dernec readings, figure 8.4(i).

load is shown

The idealised stressesin node I are shown in figure 8.40). It can be seen that the fc, 0.55 The in less than this numerical concrete stresses zone are all stresses applied load, in figures 8.4(m-n) displayed that crushing of tile and show at ultimate are concrete was initiated at the lower diagonal strut. This crushing was present in tile but did happen load had the not until ultimate was steel yielded and experiment be in diagonal It that the the to the can seen closer column, stresses are attained. by fact This is due those than the the that as the to envisaged strut-tie model. greater bottle it be However, the column, a neck effect occurs. can also approaches strut level J1.11101011 the that the of stress concentration IS around colurnn-corbel observed in 0.8fc,,, The C2A. I. Ifc,, to to about comparison compressive reached in reduced diagonal the the upper part of along strut are of a similar magnitude to those stresses by fig. 8.4(n). finite beam detailed As BI, the model, strut-tie more in a predicted elernent analysis of this nodal region was carried out using a displacement control displayed The in figures 8.4(o-p). The funnelling of the this results of are analysis. be from the principal stressplot. can clearly stresses seen compressive

271

X,

Strut & Tic Design

180 50

150

150

350

bi 4

0-

125
i) elevation

200

250
ii ) Side
c1cvatiOll

20mm fl(,. 9.4(a) Corbel C3A, all sizes In 111111, COVel. 11

V
A
Formwork

I'lg. 8.4(b) Corbel C3A Reinforcement

272

Chapter 8

Sum &, Tie Design

fig. 8.4(c) Corbel C3A: Initial Crack Pattern (0.4PI)

fig. 8.4(d) Corbel C3A: Crack Pattern (0.7Pd)

273

chaptel

8SI

ILI I&IICI

)C I '[ I

1.6 1.4 1.2

a Co 0

c1 AM 0
0.8 m
' CL 0.6

'

0.4 0.2

numerical

0.2

0.4

0.6 strainlyield strain

0.8

1,2

fig. 8.4(e) Corbel C3A: Main Steel Strains


1.6

1.4

1.2

2 C 0,

08

0.6 CL
a

0.4

tnu
0.2

-ericalr mf expe_e rlrrlenjal i al

0 0 0.1 0.2 0.3 strain/yield strain 0.4 0.5 0.6

fig. 8.4(f) Corbel C3A: Upper Stirrup Steel Strains

1.4

1.2

0
r U)

0.8

0
0.6
CL 0

OA 0 numedcal

0.2

--Cl- expenmenial

-0.2

0.2

0.4

0.6 strainlyield strain

0.8

1.2

1.4

Ij
274

fig. 8.4(g) Corbel C3A: Lower Stirrup Reinforcement

Chapter 8

Sti tit &Tic Design

1.2

100mm Demecs placed at middle thirds

Co
C

a) 7A

0.8 (12
0.6 1

CL

CL cc

0.4

0 -0-d2

dl

-0.0005

0.0005

0.001

0.0015 strain

0.002

0.0025

0.003

0.0035

fig. 8.4(h) Corbel C3A: Dernec Readings,+ve tensile, -ve compressive

t'lg. 8.4(i) Corbel C3A lt Ultimate load ( 1.4211,1)

275

Chapter 8

Strut & Tie Design

2 N/mm 13.4 (Y=

--7 =

(j=5.8

100111111

7.8 N/nim 2

65ninli

fig. 8.40) Corbel C3A: Stressesin Node I

700 650 E (265.0,645.5) D (225.0,565.5) C (2(X). 5,523.2)

500

350
*I

B (129.5,391.1)

A (120.3,342.1 )

125

325

fig. 8.4(k) Corbel C3A: GuassPoint Positions for Numerical Stresses

276

Strut & Tic Design

14
12

la
ob

08

06

04

0,2

o 0 040.5 displacement (mm) 0.6 0.7 0.8 09

fig'.8.4(l) Corbel C3A: Numerical Load-displacement


1.2

B
1

08 LL 0.6

0.4 A 0.2

Bj L77077
04 0.6 0.8 load 1 1.2

00 0 02 1.4 applied load/design

Corbel C3A: Numerical compressive stresses


07 0.6 0.5 0.4 0.3 0.2 0.1 01 0 02 04 0.6 0.8 load 1 1.2 1.4 applied load/design

Z;

fig. 8.4(n) Corbel C3A: Numerical Compressive stresses

277

( '11kj)kl

Yt III &II"I),.

I,,II

11 1

120nim

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

70nim

compression tension

93.3mm

80111111

fig. 8.4(o) Corbel C3A: Nodal Zone 2 at ultimate load

's-SLItc

ion-conipression -oachingpeak x1alcompression -oaching crushint-,

El
m

fig. 8.4(p) Corbel C3A: Nodal Zone 2 at LdtirllatCload

278

Chapter 8

Sirut & Tie Design

8.5

Corbel C4A

C4A was a one-sided corbel with the same width and depth as C2A and C3A. The lower for designed hence less C4A the a than and was other corbels thickness of was load. The geometric details and material properties are given in the table below.
Schematic Material Properties (Design) Pd=200kN Pd concrete: f,, =50N/nu-n 2 f, =3. ON/min 2 E, =21.5kNh-nni') steel: fy=50ON/mm 2 2 E, =200kN/nii-n thickness= I 50nini Material Properties (E xperimental)

350 150,, 150 350

1,250j00j F,

concrete: f,,,=49.6N/n-u-n2 2 1', =33N/aun E, =2 1.77kN/nun steel: 2 (01()) f =508N/nini y E, =210.35kN/nun 2 f =509N/ni 1112(0g) y Es=214.5kN/unni 2

8.5.1 Design
Using the same strut-tie model as for C2A and C3A, the maximurn stress ill 111C 2, 23.7N/mm diagonal strut was calculated as which Is less, thall the Pernilucd

bars layout 2xOlO N/rnM2) 0.75f,,,, (37.5 reinforcement min of' as maximum of .A main steel and 4x06 rnm bars as lower steel gave a maximum load of 203kN. As *in
the case with corbel C3A, only 08mm high yield bars were available, and these were

6mrn in bars. Using the this increased value of steel, structure place of actual in used
the theoretical ultimate load was 255kN, (l. 3P,j). The reinforcement layout is shown

in figures 8.5(a-b)

8.5.2 Numerical and Physical Testing


The model was tested using lOkN increments. A total experimental load of 240kN, (1.2P,I) was achieved. In the numerical model, an ultimate load of 250kN, (l. 25Pj), During the experiment, the first cracks, its with C2A and C3A, achieved. was appearedat the upper column-corbel junction at the point of highest tension, at 0.2P,I,
largest fig. 8.5(c). The of these cracks was measured as 0.05nini. see On further

loading, these cracks propagateddownwards towards the column. At 0.7P,I, a diagonal

279

Chapter 8

Strut & Tie Design

crack appearedin the compressivezone inside the column, see figure 8.5(d). A crack along the main diagonal strut, running from the baseplate along the line of the diagonal strut occurred at 0.8Pd. see fig. 8.5(e). This crack was 0.1mill wide. Subsequent loading led to further propagation and widening of the rnam diagonal crack and the service limit width of 0.3mrn was reachedat a load of 1.15P, I. The experimental steel strains are shown in figures 8.5(g-i). A relatively good agreementwith the numerical strains is observed. In both the numerical and physical model, the main steel began to yield just prior to the ultimate load. Both lower bars were close to yield at this point, reaching around 95'Y(, yield strain. This yielding led to opening tip or the main diagonal crack and led to crushing of the concrete around tile column-corbel junction, see figure 8.5(l). The Deinec readings across (lie diagonal track confirm this trend, seefigures 8.5(l).
No crushing of the concrete was observed nuincrically, or during tile experiment. Thc be in It figures 9.5(ni-n). that observed can shown are stresses numerical compressive 0.7f,, C3A, in tile order of largest the as was of area its sarne the stress occurred Elsewhere, the stresses along the diagonal did not cxceed tile theoretical maximum of' 0.5fc,,, see figure 8.5(m). Figure 8.5(n) shows that the stress along tile diagonal at 0.15fc,, 0.3fc, frorn This 0.8P, to C at Sudden around suddenly increases I. point in load level. due this taking the place to region at same cracking increase is

280

Chapter 8

Strut &, Tle Deil, ll

lln

150

150
0

crs

350

00 cl-S

1.4

250

1014 200

ig

lb

150 ii) mdc

i) elevation

fig. 8.5(a) Corbel C4A, all sizes in mm, cover 20nim

.. ___________________ -____

t'lg.8.5(b) Corbel C4A: Reinforcement Layout & Formwork

281

Chapter 8

Strut & Tic Design

Initial Crack,Pattcrn (0.21), 117,. Z:, 8.5(c) Coi-belC4A: 1)

figy.8.5(d) Corbel C4A (0.7P, I)

282

Chapter 8

sti kit & Tic Design

fig. 8.5(e) Corbel C4A (LOP,I)

fi g 8.5(1') Corbel ('4A at Ultimate load (1.213,1) .,.

283

Chapter 8

Strut & Tic Design

1.4

1.2

1
C

0.8

O6

0.4 0 0.2 numerical

0.2

0.4

0.6 strain/yield

0.8 strain

1.2

1.4

fig. 8.5(g) Corbel C4A: Main Reinforcement


1.4

12

0.8

0.6

F1

0.4

0',

expew... W]

0010.2

0.3

0.4

0.5 strain

0.6

0.7

0.8

strain/yield

fig. 8.5(h) Corbel C4A: Upper stirrup reinforcement

V 0 C a, 0, 0, V 0, 0 V 4, 0.

-0.1

0.1

0.2

0.3 strain/yield

0.4 strain

0.5

0.6

0.7

0.8

fig. 8.5(i) Corbel C4A: Lower Stirrup reinforcement

284

Chapter 8

S(rut & Tic Design

2 N/nim 17.9 (Y=

--77-

(T=7.7 N/nun2

100111111

3.7 N/mm 2

65ni

fig. 8.50) Corbel C4A: Stressesin Nocle I

1000

---------650
1) (390.9,645.6) 0

500

C (285.5,4 16.8) 0

350

"B

(254.5,359.8)

A (240.6,336.9)

________

250

450

fig. 8.5(k) Corbel C4A: Gausspoint Positions for Numerical Stresses

285

Clllplcl

SI1,1),

100mm Demec

gauges at middle thirds . ........... (12

X]

dI

0 --0-

dl d2

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

-0.0005

0.0005

0.001 strain

0.0015

0.002

0.0025

0.003

fig. 8.5(l) Corbel C4A: Dernec Readings


0.80

0.70

0.60

0.50

A
0.40

0.30

0.20

0A
0.10

-0-B

0.00 0.00 0.20 0.40 0.60 applied 0.80 load/design load 1.00 1.20 1.40

fig. 8.5(m) Corbel C4A: Numerical Compressive Stresses


0.5 D 0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1
0.05 0r 0 0.2 OA 0.6 applied load/design 0.8 load 1 1.2 1.4

0C
--0DI

fig. 8.5(n) Corbel C4A: Numerical Compressive

286

Chapter 8

Strut & Tie Design

8.6

Corner Joint FJIA

In the test series,two designs for a corner joint of a typical frame structure were made. FJ IA described here, was designed for a closing moment as shown in figure 8.6(a). Previous experiments have shown that current design procedures in corner joints are design is M, joints This inadequate the to attain moments, problern unable with often I.
in Swann( 1969), Mayfield the of opening moments; case et. al pronounced more even

(1971), Nillson et.al (1976), Jackson (1995), Abdul-Wahab et al (1999). Details ot


below. design material experimental properties are the corner geometry and and given Schematic 750 Material Properties (Design) M, 1=12kNin concrete: f = 40Nh nm ,,, 2 f, =3. ONh-ni-n
E, =21.5N/iii 112 1

Material Properties (Experimental)

concrete: I'

150T 750 j 150

steel: fv=50ON/inin 2 E, =200kN/nun tlilckness=150nmi

=38.9Nhmii c,, f, =12N/nun 2 E, =21.76kN/niiii 2 steel: l'y=504N/imn 2 (012) E, =201.6kN/imn 2 G=509N/inin 2 (08) E, =214.5kNhimi 2

8.6.1 Strut-Tie

Model and Design

The results of the elastic analysis and corresponding strut-tie model are given in fig 8.6(a). Since a cover of 15nim will be used in the designed structure, the effective depth is 120mm. This results in an application of lOOkN horizontal loads to the strut-

tie model as shown in fig8.6a(iii) to create the design moment of 12kNrn. The tic forces and resulting reinforcement provisions are given in the table below.
Member 1 5 6 9 Force (kN) 100 103.1 103.2 100 A, required
(MM2)

A, provided (mm') 226 226 226 226

No. of Bars 2012 2012 2012 2012

200 206.2 206.2 200

A, provided A, required 1.1 1.1 1.1 1.1

287

Chapter 8

Strut & Tie Design

At the inner corner of the structure, i. e. node 2, the concrete is in a state of bi-axial is limited Llf,, hence to the and stress compression The depth of struts 2 and 8 is

block in depth the the to stress of of compressive reinforced member equivalent bending, in this case 30mm. This value was calculated from simple bending theory. The critical area in this model is the TCT node 3. Here the tie forces cause inner the circumference of the reinforcement. along compression

From the geometry of the bar, the effective width of the strut at this point is equal to 50rnm. Due to the presence of the two ties, the concrete stress in this zone must be less than 0.6f,,,. From the strut-tie model, the force in this strut is equal to 107kN, 2<0.6f,,,. A in 14.3N/i-nm of schematic of this node is shown *111 a stress which results figure 8.60). The design reinforcement layout is given in figure 8.6(b). Where tile fixed floor, in form leg the to the the was reinforcement corner of' a spiral of vertical for extra strength. provided was

288

Chapter 9

Strut & Tie Design

i) (rr=O)

n) (rr=30%)

I( 1k N

allode (). 6 fcd

()

1-21

ir

I WkN

I OOkN

I NAN

iii) Strut-tie Model

iv) Dimensioning

fig. 8.6(a) Corner Joint FJ I A, Elastic stressesand Strut-tie mode

289

Chapter 8

Strut & Tic I)csign

750 100 7 (a)100 crs *A-I II

100

Strain (Illutye C7 C7 10,


E

17-

A
I Z-150

750

100
d

06 50mm spiral cc 150

150 section A-A cova on stirrups 15min stirrLIPSX


Section

20)
100

fig. 8.6b(i) Corner FJ IA Reinforcement Layout, al I sizes in inni

1,4 ,

:
fig. 8.6b(ii) Corner FJ IA Reinforcement & Formwork

290

Chapter 8

Strut & Tic Desigii

8.6.2 Numerical and Physical Testing


Details of the experimental set-up are given in chapter 7. As seen from fig. 7.3.3(a), a closing moment was generated in the corner via the hydraulic Jack at the end of the horizontal. The moment was calculated as the load times the lever arm to the centre of the vertical, in this case 625mm. Strain gauges were placed at the points maximurn tension in the main reinforcement.

During the experiment, the load was applied in RN increments until failure occurred. The experimental ultimate load was recorded as 20.7kN, which corresponds to a moment of 12.96kNi-n, 0 -08Md). The numerical ultimate moment was recorded its 14.5kNi-n, (1.2 Md). In the experiment, the first cracks occurred at a load of 4kN, (02MA around the tension face of the vertical and horizontal, fig. 8.6(c). The largest

of these initial cracks was measured as 0.02mm. Under increasing moment, the cracks widened and propagated toward the centre of the members. New cracks appeared in the corner at the seventh increment, (0.73M, j), along the direction of the main tension reinforcement. These cracks correspond to increased strain in the steel at these points, fig. 8.6(d). The largest crack, at the beam column junction, reached the service crack limit width of 0.3mm at a load of l6kN, (0.8 Mj). At the ultimate moment, the main horizontal and vertical steel yielded, and this was accompanied by widening of tile cracks around the tension zone, fig. 8.6(e-f). The numerical and experimental strains presented in figs. 8.6(g-h) show a reasonable correlation. steel

The numerical compressive stresses show that stress concentrations occurred in tile inner corner, which as previously stated is under bi-axial compression, fig. 8.6(k). The limit stress of 1. If,,, was not reached at this point. Similarly, as predicted in the struttie model, the hrnit stress of 0.6f,,, was not reached in the diagonal strut, fig. 8.6(i)

291

Chapter 8

Strut & Tic Design

fig. 8.6(c) Corner FJ IA

(0.3Md)

fig. 8.6(d) Corner FJ IA (0.73M, I)

292

Chapter 8

Stl ul & Tic I

fig. 8.6(e) Corner FJ IA at ultinialc (1.08M,I)

I'lg.8.6(f) Corner FJ IA at Ultimate (1.08M,I) 293

Sll w 'K, VIC I

1.4

1.2

E 0 E 0.8

E 0.6 0 E 13 (D '& 0.4 CL (ts Ital


0.2

--11- ex

JE
1.2

0
0 0.2 0.4 0.6 strairNeld strain 0.8 1

figg. 6(g) Corner FJ I A: Main Steel Strains

1.4

12

m1 tu
'70 (D

0.8

-2 0.6 'a
CL

m 0.4

0.2

0 rmancl --D- e)pennvrtA

0.2

0.4

0.6 s"r4idd stWn

0.8

1.2

fig. 8.6(h) Corner FJ I A: Main Steel Strains

294

Chapter 8

Strut & Tic Design

50nim

inner radius ofbend = 36mm -mm- I WkN

14.3Nj caF=

/MM2\\\

-------------------------------I OOkN

fig. 8.6(l) Corner FJ I A: Node 3 Stresses

750

0D

(56.25,693.7)

0C 600

(93.7,656.2) -B( 161.0-6062)

A( 131.2,568.7)

450

150

300

fig. 8.60) Corner Joint FJ I A: GaussPoint Positions for Numerical Stresses

295

Cilaplel 8

S11111 &. I lk. I)c, wil

0.9

B
0.8

0.7

0.6

0.5 (0)

0.4 0.3 0.2 0.1

A -------------

10 --0-

A B

0.2

0.4

0.6

0.8

1.2

1.4

applied load/clesign load

fig. 8.6(k) Corner FJ I A: Numerical Compressive Stresses


0.6

C
0.5

D
0.4

0.3 Z;

v'

0.2

----------

0.1

6c

1--0-D

ou 0

0.2

0.4

0.6 applied

0.8 load/design load

1.2

1.4

fig. 7.6(l) Corner FJ I A: Numerical Compressive Stresses

296

Chapter 8

Strut & Tic Design

8.7

Corner FJ2B

As previously stated, this model was the second of two typical frame corner joints designed. FJ2B was designed for the opening moment which has been proved by a be load. to the to worst scenario with regard achieving ultimate of researchers number
In practice, corners subject to opening moments occur in retaining walls under active earth pressure, water storage tanks under hydrostatic pressure etc. The material

below below in table the are given properties etc. Schematic


750

Material Properties (Design)


M, I= I 2kNi-n

Material Properties (Experimental)

concrete:
150T 750
L
Ik-

concrete:
1 ' =41.3N ,,,

f = 40N/ nu n ,,, f, =3. ON/i-nM2 E, =21.5kN/nim steel: f, =50ONh-nni


E, =20UN/nun

h-n 111 f',=3.5N/iw-n 2 2 E, =23.73kN/nini


y E, =201.6kN/ni

150

steel: f =504N/nini

2 (012)
III 2

X-) thickness= I 50nu-n

fy=509N/n-u-n 2 (08) Es=214.5kN/nini 2

8.7.1 Strut-Tie Model and Design


The sarnestrut and tie model can be used in the design for the opening nlonlent with the reversal of the member force directions, so that struts now become ties and vice
figure 8.7(a). The forces tie and provided steel are shown the table below. see versa,

Figure 8.7(b) shows the designedreinforcement layout.


Member 2 3 4 7 8 Force (kN) 100 0 106.4 0 100 A, required
(MM2)

A, provided
(MM2)

No. of Bars 2012 408 2012

200 200 200

226 201.1 226

A, provided A, required 1.1 H) 1.1

297

Chapter 8

Strut & Tie Design

I OMN

()

-imp I WkN

I MkN

I OOkN

i) Strut-tie Model

5 0.7 t,,, <

ii) Dimensioning

fig. 8.7(a) Corner FJ2B Strut-Tie Model

298

Chaptei 8

StILIt&'.TIC DC-Sll)

750 7 (cil100 c rs 08 stirrLIPS

straIn ILI LYC 0 A 12) 108) 150


750 5 (a-' 100

2(012)
150 spiral section A-A Cover oil stirrups StI H-LIPS0 section C-C 15111111

06 50mm

150

12

(a' I 100

fig-8.7b(l) Corner FJ2B Reinforcement

Detills

. ih I I) Corl IcI I Io. ',

PCIII force I] ICII

299

Chapter 8

Strul & Tie Design

8.7.2 Numerical and Physical Testing


In the experiment, the load was applied in increments of RN. The ultimate mornent In the experiment was recorded as 12.48kNm, (1.04MA and the numerical ultimate The first 05Mj). load 12.6kNm, (I. of cracks appeared at a recorded as was moment 4kN, (0.2M,j), occurring around the tension face of the inner corner, see fig. 8.7(d). The largest of these cracks was measuredas 0.05mm. As the moment increased, the cracks propagated toward the centre of the members. The widest crack, at the inner 0.3mm Ili-nit the of on the seventh increment, (0.73M,j), service width reached corner, see fig. 8.7(e). As the main tension steel approached yield, widening of the existing by further in cracking accompanied a diagonal along the outer face of the cracks was dead Yielding known the zone. of the main vertical steel occurred at around as corner 0.9M,l in the inner corner. At ultirnate load, the largest crack at the inner corner was measuredas 0.6rnm, seefigures 8.7(f-g). The experimental and numerical steel strains are shown in figures 8.7(h-j).

As envisaged in the design, crushing of concrete in the compression zone did not occur. This is due firstly to the fact that the calculated design stress levels were within those permitted, and secondly due to the additional contribution to strength from tile additional steel in the compressive zone. The numerical cornpressive stresses in the corner are shown in figures 8.7(1-m). It can be seen that the largest compressive stress in the concrete is around 0.32fcu. The sudden increases in stress observed at points C, F and E correspond to the formation of cracks oil the tension side of the beam and column, close to the inner corner.

300

Ic

Comci H2B (0.4-'N],, )

HI

72

//'
I

I.

fig. 8.7(e) Corner FJ2B (0.73M, i)

301

,, I Ic I ), "Wil

Ei"

c'

Ii

/
Corner FJ2B at ultimatc ( hg,. t7l 8.7(J')

, i "I

,I

-"

fig. 8.7(g) Coriler l, J2B at Uitimate (1.04M, I)

302

Chaptel

sullt "Q Ile I

1.2

E 0.8

0,6

*F,
0.4
CL

0.2

[ emcal -I ; ; experimental

0.5

1 strain/yield strain

1.5

2,5

fig. 8.7(h) Corner FJ2B Main Steel Strains


1.2

E 0.8 0 E c P 0.6 E 0 E

1 =

0.4

----IF,
0.2 0 -0-Aenta numeric al expe rim

on 0

0.1

0.2

0.3

0.4

0.5 strain/yield strain

0.6

M080.9

t.ig8.7(i) Corner FJ213Main Steel Strains


t2

0 0.8

r 0.6

0.4

0.2

numerical

-0--expenmenlal

F' ..........
0,9

0.1

0.2

0.3

0.4 strain/yield

0.5 strain

0.6

0.7

0.8

fig8.7 ) Corner FJ2B Diagonal Steel Strains

303

Chapter S

'Strul

&, TIC I)CIgll

75
BCF A 60 F) A B C D ------E 0 fig. 8.7(k) Corner Joint FJ2B: 15 Gauss Point Positions (33.3,694.5) (33.3,716.7) (56.2,716.7) (33.3,619.5) (131.2,716.7) 30 for Numerical Stresses

45

U.Z 0.18 0.16 0.14 0.12

00

0.1 0.08 0.06 0.04 0.02

0.2

0.4

0.6

0.8

1.2

aMiled momeWdesign mom"

Numerical compressive stresses fig. 8.7(l) Corner FJ213


0.25

0.2

0.15

0.1

0.05

D
01 0

............
1.2

0.2

0.4

0.6

0.8

apMied moniw*dwJgn nimat

fig. 8.7(m) Corner FJ2B Numerical compressive stresses

304

Chapter 8

Sti-tit & Tic Dcsign

8.8

Conclusions

The results from the physical models are summarised in the table below, where As, is the total area of tension steel. Model
P/Pd

C2A
1.18

C3A
1.42

C4A
1.20

FJIA
1.08

FJ211
1.04

Astprovided
Load at first fielding of ma n steel Load at f rst reaching limit crack service

1.0
I-I Pd 0.78Pd

1.3
1.3Pd I. OPj

1.3
1.2P,j 1.15Pj
Program

1.1
1.1 Ri (). 8p',

1.1
(). gpl 0.7P,,

A required st

Table 8.8(a) Summary

of Experimental

As shown in the table above, every model was able to achieve its design load. In all C, the models tested, failure was initiated by yielding of the main steel. It can be seen from the table that the main reason for extra strength in the model is clUe to the increasein provided steel. A comparison of the ultimate load predictions obtained by the strut-tie models, assuming that steel yielding governs failure, finite elements and experimental ultimate loads is presented in table 8.8(b). In each case, the strut-tie model provided a good prediction of the ultimate load behaviour of the structure, comparable with that of the finite element model. Model (strut-tie) _P,, P,,(finite element) P,,(experimental) BI 1.15 1.32 1 C2A 1 1.15 1.18 C3A 1.3 1.3 1.42 C4A 1.3 1.25 1.2 F.IlA 1.1 1.2 1.08 F.1211 1.1 1.05 1.04

Table 8.8(b) Comparison of ultimate loads

For every model, the nodal zone capacities were assessed according to the given state of stress. The stressesin the struts and nodes resulting form the design load were checked to be within the permitted levels. During the physical testing, no concrete crushing was observed until after yielding and hence ultimate loading had occurred. From this it may be concluded that the factors used in the design process were adequate.

305

Chapter 8

Strut & Tie Design

It was also shown that the strut and tie method as a design tool, can produce designs with comparable performance to the direct design procedure. The advantage that the is it designer direct design has that to gain the the method allows over method strut-tie be It behaviour into load insight the the possible to use of structure. would carrying an the direct design procedure in combination with the visualisation process for plane for 6. However, difficulties done in chapter as was slabs sorne applications stress in

this application may arise due to the fact that the steel must be orientated to the principal stress directions in each element.

306

Chapter 9

Conclusions

Chapter 9 Conclusions

9.1

Summary

The main stress paths in a structure were Isolated using an eVOILItionary procedure, termed as visualisation. Using these stresspaths, strut and tie models were developed and reinforcement layouts designed accordingly. These stress fields, were also used for the direct design of reinforced concrete slabs. All the designed structures were tested numerically using non-linear finite elements, and a number of structures were testedphysically in the laboratory.

9.2

Slab Design
Design using the visualisation process results in satisfactory behaviour both at service and ultimate loads The rnethod does not always result in a practical reinforcement layout and it is often found that a greater quantity of steel than the numerical amount is required for practical considerations. This increase in provided steel often leads to higher ultimate loads being achieved.

The visualisation process is not always applicable in stabs where the stressesare evenly distributed, such as in a 2-way simply supported slab subject to it uniformly distributed load. It is necessaryfor there to be a good spread of initial elastic stressesbefore visualisation can be effective.
The degree of mesh refinement does not have on effect upon the direction of the evolved stress paths. However, it is necessary to use a rnesh fine enough to model

the stressvariation adequately.

307

Chapter 9

Conclusions

In areas where a reinforcement layout is anticipated by the designer, e.g. in be by dictated the considerations, visualisation process can practical situations very useful.

9.3
*

Strut and Tie Design


in is developing useful process strut-tie models. Tile re-

The visualisation

distribution of stress caused by the evolutionary process is similar to the redistribution taking place in the actual reinforced concrete structure.
0 Design from strut and tie models can produce satisfactory behaviour both at

ultimate and service loads. 0 As a design tool, the strut and tie method can produce designs comparable to the direct design method. In addition, the strut and tie method helps the designer to understandthe load carrying mechanism of the structure whereas the direct design rnethod may often be treated as a 'black box' type systern. 0 As an analytical tool, strut and tie models can lead to ultimate load predictions with comparable accuracy to non-linear finite elements. 0 Nodal areas are the critical areas in the structure and the strut strength used for design must take into account the stressstate of the node. Sometimes, compressive stressesin the nodal areas are not evenly spread over the width of' tile node, but concentrated at a localised point. This may lead to compressive stressesin excess of the design strength.

9.4
0

Suggestions for Further Work


All the structures designed in this work were subjected to a single load case. The method can be extended to multiple load cases. In this scenario, it may be necessaryto use more refined meshesin anticipation of more complicated stress paths. Similarly more complex structurescould be examined.

The use of designer intervention

in the stress path evolution should be further

developed both in slab design and in strut-tie model development.

Further investigation in nodal zone behaviour is necessary.In particular the nature


of compressive stress concentrations such as the funnelling effect observed In the

corbels.

308

References
Abdel-Hafez, L. M., 'Direct Design of Reinforced Concrete Skew Slabs', Ph.D. Thesis, University of Glasgow, Civil Engineering Dept., (1986) 2. Abdel-Kader, M., 'Prediction of ShearStrength of Reinforced and Prestressed Concrete Beams by Finite Element Analysis', Ph.D. Thesis, University of Glasgow, Civil Engineering Dept., (1993)
Abdul-Wahab, H. M. S. & Salman, S.A. R., 'Effect of Corner Angle on Efficiency of Concrete Joints under Opening Bending Moment' ACI Structural Journal, Vol. 96, No. 1, pp. 115-12 1, (Jan/Feb 1999)

3.

4.

Adebar, P., Kuchma, D., & Collins, M. P., 'Strut and Tie Models for the Design Study' Experimental Caps: An Pile of ACI Structural Journal, Vol. 87, No. 1, (Jan[Feb 1990) Adebar, P., & Zhou, L. Z., 'Bearing Strength of Compressive Struts Confined by Plain Concrete' ACI Structural Jounial, Vol. 90, No. 5, (Sep/Oct 1993) Adebar, P. & Zhou L. Z., 'Design of Deep Pile caps by Strut and Tie Modcls' ACI Structural Journal, Vol. 93, No.4, (Jul/AUg 1996) Ahrnad, S.H., & Mallare M. P., 'A Comparative Study ofModels for Confinement of Concrete by Spirals' Magazine of Concrete Research,Vol 46, No. 166, pp.49-56, (Mar 1994)
Al-Mahaidi, R. S.H., 'Non-linear Finite Element Analysis of Concrete Deep Members' Report No. 79- 1, Cornell University, (Jan 1979)

5.

6.

7.

9.

Almusallam, T. H., & Alsayed S.H., 'Stress-Strain Relationship of Normal, High Strength and Lightweight Concrete' Magazine offoncrete Research, Vol 47, No. 170, pp. 39-44, (May 1995)

10.

Alshegeir, A., & Remirez, J.A., 'Strut-tie Approach in PretensionedDeep Beams' ACI Structural Journal, Vol. 89, pp.296-304 (1992) Alshegeir, A., & Remirez, J.A., 'Computer graphics in Detailing Strut-tie models' ASCE Journal offomputing in Civil Engineering, No.6, pp.220-232 (1992)

12.

Armer, G. S.T., Discussion of Ref. 129. Concrete, Vol. 2, pp. 319-320, (Aug. 1968)

309

13.

Bazant, Z. P., & Cedolin L., 'Fracture Mechanics of Reinforced Concrete', , ASCE Journal of'Meclianical Division, Vol. 106, pp. 1287-1306, (1980)

14.

Bell, J.C., & Elms, D. G., 'Finite Element Approach to Post-elastic Slab Behaviour' ACI Special Publications, SP30-15,pp.325-344, (Mar 1971)
Bensalem, A., 'Direct Design of RC Structures using Non-elastic Stress Fields' Ph.D. Thesis, University of Glasgow, Civil Engineering Dept. (1993)

15.

16.

Bergmeister, K., Breen, J.E., & Jirsa, J.0., 'Dimensioning of the Nodes in and Development of Reinforcement' IABSE Colloqiunz, Stuttgart, Germany, pp.55 1-556, (199 1) BS81 10 Part I&2, 'Structural Use of Concrete' British Standards Inst., ( 1985) Buyukozturk, 0., Nilson, A. H., Slate, F.O., 'Stress-Strain Responseand Fracture of a Concrete Model in Biaxial Loading' A CI Journal, Vol. 68, pp. 590-599, (Aug 1971) Cedolin, L. & Deipoli, S., 'Finite Element Studies of Shear Critical Reinforced Concrete Bearns' ASCE Journal of* Mech. Div., Vol. 103, No. EM3, pp.359-4 10, (Jun 1972) Cervenka, V., Gerstle, K., 'Inelastic Analysis of Reinforced Concrete Panels Part 1: Theory' IABSE Publications, Vol. 31-11,(197 1) Cervenka, V., Gerstle, K., 'Inelastic Analysis ofReinforced Concrete Panels Part 2: Experimental Verification and Application' IABSE Publications, Vol. 31-11,(197 1) Chen, W. F., 'Plasticity in Reinforced Concrete', McGraw Hill, New York, ISBN 0-07-010687-8, (1982)
Clark, L. A., 'The Provision of Tension and Compression Reinforcement to Resist In-plane Forces' Magazine of Concrete Research, Vol 28, No. 94, (Mar 1976)

17.

18.

19.

20.

21.

22.

23.

24.

Clark, L. A., & Speirs, D. M., 'Tension Stiffening in RC Beams and Slabs under Short Term Load' Cement & Concrete Association, Technical Report 42.521, (1979)
Collins, M. P., Vecchio, F. J. & Mehlhorn, G., 'An International Competition to Predict the Response of Reinforced Concrete Panels' Canadian Journal oj'Civil Engineering, Vol. 12, pp. 624-644, (1985)

25.

310

26.

Collins, M. P., & Mitchell, D., 'A Rational Approach to Shear Design - The 1994 Canadian Code Provisions' ACI Journal, (Nov/Dec 1986)

27.

Collins, M. P., Mitchell, D., Adebar, P., & Vecchio, FT, 'A General Shear Design Method' ACI Structural Journal, Vol. 93, No. 1, (Jan[Feb 1996)
Cook, W. D., & Mitchell, D., 'Studies of Disturbed Regions near Discontinuities in Reinforced Concrete Members' A CI Structural Jounial, (Mar/Apr 1988) Cope, R. J., & Vasudeva Rao P., 'Non-linear Finite Element Analysis of Concrete Slab Structures' Proceedings oj'ICE, Vol 63, Part 2, pp. 159-179, (Mar 1977)

28.

29.

30.

Crisfield, M. A., 'Non-Linear Finite Element Analysis of Solids and Structures, vol. 1: Essentials' John Wiley & Sons, (199 1) Desayl, P., & Krishnan, S., 'Equation for the Stress-StrainCurve of Concrete' Journal of the American Concrete Institute, (Mar 1964) Eibl, J., Akkermann, J., Idda, K., Lucero-Cirnas H, 'Rotational Behaviour of Reinforced Concrete Corners and bond under Lateral Tension' CEB Bulletin d7i! lormation, No. 242, 'Ductility of Reinforced Concrete Structures', (May 1998) EI-Mezaini, N., & Citipitioglu, E., 'Finite Element Analysis of Prestressed and Reinforced Concrete Structures' ASCE Journal of Structural Engineering, Vol. 117, No. 10, (Oct. 1991)
Elwi, A. E., & Hrudey, M. T., 'Finite Element Model for Curved Embedded Reinforcement' ASCE Journal of'Mech. Div., Vol. 115, No. 4, pp. 740-754 (Apr 1989) Foster, ST, & Gilbert, R. I., 'Experimental Studies on High-Strength Concrete Deep Beams' ACI Structural Journal, Vol. 95, No. 4, (Jul/Aug 1998) Gerstle, K, H, 'Simple Formulation of Biaxial Concrete Behaviour' A C1 Journal, (Jan/Feb 199 1)

31.

32.

33.

34.

35.

36.

37.

Gopalaratnam,V. S., & Shah, S.P., 'Softening Responseof Plain Concrete in Direct Tension', A CI Journal, Vol. 82, pp. 310-322, (1985)

311

38.

Gupta, A. K., Akbar, H., 'Cracking in Reinforced Concrete Analysis' ASCE Journal qf Structural Eng., Vol. I 10, No. 8, pp. 1735-1747, (Aug. 1984)

39.

Hagberg, T., 'Design of Concrete Brackets: On the Application of the Truss Analogy' ACI Journal, Vol. 80, pp. 3-12, (Jan/Feb 1983) Hago, W. A., 'Direct Design of Reinforced Concrete Slabs' PhD Thesis, University of Glasgow, Civil Engineering Dept., (1982)
Hago, W. A., & Bhatt, P., 'Tests on Reinforced Concrete Slabs Designed by Direct Design Procedure' ACI Journal, (Nov[Dec 1986)

40.

41.

42.

Hermansen,B. J. & Cowan, J., 'Modified Shear-friction Theory for Bracket Design' ACI Journal, Vol. 7 1, No.2, pp.3-12, (Feb 1974) Hillerborg, A., 'Reinforcement of Slabs and Shells Designed According to the Theory of Elasticity' Betong, 38(2), pp. 101-109, (1953) Hong, S.G., 'Truss Model for Tension Bars in Reinforced Concrete Bearns: Tension-Tension-CornpressionRegions' ACI Structural Journal, Vol. 93, No.6, (Nov/Dec 1996) Huang, F.C., Lee, I. S.,& Mo, Y. L., 'Designing Pier Caps with Strut and Tie Models' Concrete International, 43-47, (Jan 1998)
Issa, M. A., & Tobaa, H., 'Strength and Ductility Enhancement in High Strength Confined Concrete' Magazine of Concrete Research, Vol 46, No. 168, pp. 177-189, (Sep 1994)

43.

44.

45.

46.

47.

Jackson,N., 'Design of Reinforced Concrete Opening Corners' The Structural Engineer, Vol. 73, No. 13, (Jul. 1995) Jain, S.C., & Kennedy, J.B., 'Yield Criterion for Reinforced Concrete Slabs' Proceedings oj'ASCE, Journal of'the Structural Division, Vol. 100, (Mar 1974)
Jofriet, J.C., & McNeice, G. M., 'Finite Element Analysis of Reinforced Concrete Slabs' Proceedings (#'ASCE, Journal oj'the Structural Division, (Mar 197 1)

48.

49.

50.

Johnarry, T., 'Elasto-plastic Analysis of Concrete Structures using Finite Elements', Ph.D. Thesis, University of Strathclyde, Dept. of Civil Eng., (May 1979)

312

51.

Kernp, K. O., 'Optimum Reinforcement in a Concrete Slab Subjected to Multiple Loading' IABSE Publication, Vol. 3 1, pp. 93-105, (197 1)

52.

Kent, D.C., & Park, R, 'Flexural Members With Confined Concrete', Proceedings oj'ASCE, Journal oj'the Structural Division, Vol. 6 1, (Jul 1971)
Khaskheh, G. B., 'Direct Design of Reinforced Concrete Transfer Girders'. Ph.D. Thesis, University of Glasgow, Civil Engineering Dept., (1989)

53.

54.

Kong, F.K., Robins, P.J., & Cole, D. F., 'Web Reinforcement Effects oil Deep Beams' ACI Journal, (Dec 1970) Kong, F.K., Robins, P.J., Kirby, D. P. & Short, D. R., 'Deep Beanis with Inclined Web Reinforcement' ACI Jou"ial, Vol. 69, pp. 172-176, (Mar. 1972) Kong, F.K., Robins, P.J., Singh, A., & Sharp, G.R., 'Shear Analysis and Design of Reinforced Concrete Deep Beams' The Structural Engineer, Vol. 50, No. 10, (Oct. 1972) Kong, F.K., & Sharp, G.R., 'Structural Idealization for Deep Bearns with Web Openings' Magazine of Concrete Research,Vol. 29, No. 99, pp.81-9 1, (JU11,1977) Kong, F.K., Sharp, G.R., Appleton, S.C., Beaumont, C.J, & Kubik, L. A., 'Structural idealization for Deep Beams with Web Openings: Further Evidence' Magazine qf Concrete Research,Vol. 30, No. 103, (Jun. 1978)
Kong, F. K., 'Reinforced Concrete Deep Beams', Blackie and Son Ltd, Glasgow, ISBN 0-216-92695-5, (1990) Kotsovos, M. D., & Pavlovic, M. N., 'Structural Concrete Finite Element Analysis for Limit State Design' Thomas Telford Publications, London, ISBN 0-7277-2027-9, (1995)

55.

56.

57.

58.

59.

60.

61.

Kriz, L. B. & Raths C.H., 'Connections in PrecastConcrete Structures: Strength of Corbels' Journal oj' The PrestressedConcrete Inst., Vol. 10, No. 1, pp. 16-61, (Feb 1965) Kupfer, H., Hilsdorf, H. K. & Rusch, H., 'Behaviour of Concrete Under Bi-axIal Stresses' ACI Journal, Vol. 66, pp. 656-667, (Aug 1969)

62.

313

63.

Leonhardt, F., 'Reducing the Shearreinforcement in RC beams and Slabs' Magazine of Concrete Research 17, No. 53 (1965) Liu, T. C.Y., Nilson, A. H., & Slate, F.O., 'Stress-Strain Responseand Fracture Compression' in Uniaxial Biaxial Concrete and of ACI Journal, Vol. 69, pp. 291-295, (May. 1972)
MacGregor, J.G., 'Reinforced Concrete Mechanics and Design' Prentice-Hall, Englewood Cliffs, N. J., (1988)

64.

65.

66.

Manickarajah, D., Xie, Y. M., & Steven,G.P., 'An Evolutionary Method for Optimization of Plate Buckling Resistance' Finite Elements in Analvsis & Design, Vol. 29, pp.205-230, (1998)
Mansur, A., Wee, T. H., Chin, M. S., 'Derivation of the Complete Stress-Strain Curves for Concrete in Compression' Magazine of Concrete Research, Vol 47, No. 173, pp. 285-290, (May 1994)

67.

68.

Marti, P., 'Basic Tools of RC Beam Design' Journal oj'the American Concrete Inst. 82, pp.46-56, (1985) Marti, P., 'Truss Models in Detailing' Concrete International, (Dec 1985) Marti, P., 'Dimensioning and Detailing' IABSE Colloqiwn, Structural Concrete, Stuttgart, ( 1991 Mast, R.F., 'Auxiliary Reinforcement in Concrete Connections' Proceedings of*ASCE, Journal oj'the Structural Division, (Jun 1968) Mattock, A. H., Chen, K. C., Soongswang,K., 'The Behaviour of Reinforced Concrete Corbels' Journal of The PrestressedConcrete Inst, Vol. 2 1, No.2, pp.52-77, (Mar/Apr 1976) Mayfield, B., Kong, F.K., Bennison, A., & Davies, J.C.D.T, 'Corner Joint Details in Structural Lightweight Concrete' A CI Journal, Vol. 68, pp. 366-372, (May 1971) Mayfield, B., Kong, F, & Bennison, A., 'Strength and Stiffness of Lightweight Concrete Corners' A CI Journal, Vol. 69, pp. 420-427, (Jul. 1972)
Meyer, C. & Bathe, K. J., 'Non-I i near Analysis of R. C Structures in Practice', Proceedings ol'ASCE, Journal ofthe Structural Divisimi, Vol. 108, No. 7, pp. 1605-1622, (Jul 1982)

69.

70.

71.

72.

73.

74.

75.

314

76.

McNeice, A. M., 'Elastic-Plastic Bending of Plates and Slabs by the Finite Element Method' PhD. Thesis, London University, (1967)

77.

Millar, S.G. & Jonson, R.P., 'Shear Transfer in Cracked Reinforced Concrete' Magazine of'Concrete Research,Vol. 37, No. 130, pp.3-15, (1985) Morsch, E., 'Concrete-steel Construction',
English Translation by E. P. Goodrich, MacGraw-Hill, New York, (1909)

78.

79.

Muller, P., 'Failure Mechanisms for RC Beams in Torsion & Bending' International Association fiv Britqe & Structural Engineering Publications 36-11,147-163 (1976) Ngo, D. & Scordelis, A. C., 'Finite Element Analysis of R. C. Beams', ACI Journal, Vol. 64, No. 3, pp. 152-163, (1967)

80.

81.

Nielsen, M. P., 'Yield Condition for Reinforced Concrete Shells in the Membrane State, Non Classical Shell Problems' IASS Symposium, Warsaw 1963, Ed. W. Olsak, Amsterdam, North Holland Publishing Co., pp. 1030-1038,(1964) Nielsen, M. P., Braestrup, M. W., Jensen,B.C., Bach. F., 'Concrete Plasticity, Bearn Shear-Shearin Joints-Punching Shear' Spec. Publ., Danish Soc. for Stuct. Sci. & Engrg., Tech. University of' Denmark, Lyngby, (1978) Nielsen, M. P., 'Limit Analysis and Concrete Plasticity', Prentice Hall, New Jersey,(1984)
Nilsson, I. H., & Losberg, A., Discussion of ACI Paper 'Opportunities in Bond Research' ACI Journal, Vol. 67, pp. 393-396, (May 197 1) Nilsson, I. H., & Losberg A., 'Reinforced Concrete Corners and Joints Subjected to Bending Moments' Proceedings oJ'ASCE, Journal of the Structural Division, Vol. 102, pp. 12291254, (Jun 1976)

82.

83.

84.

85.

86.

Noor, F.A., 'Ultimate Strength and Cracking of Wall Corners' Concrete, Vol. 11, pp. 31-35, (Jul. 1977) Park, R., & Paulay, T., 'Reinforced Concrete Structures' John Wiley and Sons Inc., London, ISBN 0-471-65917-7, (1975)
Phillips, D. V., & Zienkiewicz, O. C., 'Finite Element Non-linear Analysis of Concrete Structures' Proc. qj'the Inst. of'Civil Engineers, Part 2, No. 6 1, pp. 59-88, (Mar. 1976)

87.

88.

315

89.

Phillips, D. V, & Wu, Z. P., 'An Orientated Embedded Bar Formulation with Bond-slip' Numerical Methods in Engineering; Theory and Application, Ed. Pande and Middleton, J., Vol. 1, pp. 320-328, (1990) Popovics, S., 'A Review of Stress-Strain Relationships for Concrete' ACI Journal, (Mar 1970) Prakhya, G. K. V., & Morely, C. T., 'Tension-Stiffening and Moment-Curvature Relations of Reinforced Concrete Elements' A CI Structural Journal, Vol. 87, No. 5, (Sep/Oct 1990) Ramirez, J.A., & Breen, J.E., 'Evaluation of a Modified Truss-Model Approach for Beams in Shear' ACI Structural Journal, Vol. 88, No. 5, (Sep/Oct 199 1)

90.

91.

92.

93.

Ranjbaran, A., 'Embedding of Reinforced Concrete Elements Implemented in DENA' Computers & Structures, Vol. 40, No.4, pp.925-930, (199 1) Rashid, Y. R., 'Ultimate Strength Analysis of Reinforced Concrete Pressure Vessels' Nuclear Engineering and Design, Vol. 7, No.4, pp.334-344, (Apr 1968) Reinhardt, H.W., 'Crack Softening Zone in Plain Concrete Under Static Loading', Cement & Concrete Research,Vol. 15, pp.42-52, (1985) Reinke, H.G., 'Assessmentof Concrete Tensile Strength in the Design of Structural Concrete' Thesis, Institut fur Massivbau, Stuttgart, ( 1986)
Renuka Prasad, H. N, Charmakeshava, C., Raghu Prasad, B. K., & Sundara Raja Iyengar, K. T., 'Non-linear Finite Element Analysis of Reinforced Concrete Corbei' Computers & Structures, Vol. 46, No. 2, pp 343-354, (1993) Ritter, W., 'Die Bauwelse Hennebique', (Hennebique's Construction Method), Schweizerische Bauzeitung, Zurich, v. 17, pp. 41-43,49-52,59-61, (1899)

94.

95.

96.

97.

98.

99.

Rogowsky, D. M., MacGregor J.G., & See, Y. O., 'Tests of Reinforced Concrete Deep Bearns' ACI Journal, (Jul/Aug 1986)

100. Rogowsky, D. M., &MacGregor J.G., 'Design of Reinforced Concrete Deep Beams' Concrete hiternational, (Aug 1986)

316

101. Saatcloglu, M., & Razvi, S.R., 'Strength and Ductility of Confined Concrete' ASCE Journal of Structural Engineering, Vol. 118, No. 6, (Jun. 1992)

102. Saint Venant, B de., as quoted in 'Theory of Elasticity', by Timoshenko, S., & Goodier, J.N., "ed., p.33, (195 1) 103. Scanlon, A., & Murray, D. W., 'Time DependentReinforced Concrete Slab Deflections' Proceedings ofASCE, Journal of the Structural Division, Vol. 100, No.ST9 pp. 1911-1924,(1974) 104. Schlaich, J., Schaefer, K., & Jennewein,M., 'Towards a Consistent Design of Structural Concrete' Journal qj*the PrestressedConcrete Inst. 32, pp.74-150, (1987) 105. Schlaich, J., & Schafer, K., 'Design and Detailing of Structural Concrete using Strut and Tie Models' The Structural Engineer, Vol. 69, No. 16, (Mar 1991) 106. Schlaich, M., & Anagnostou, G., 'Stress Fields for Nodes ofStrut and Tie Models'

ASCEJournal of StructuralEngineering, (Jan. 1990)


107. Siao, W. B., 'Strut and Tie Model for Shear Behaviour in Deep Bearns and Pile Caps Failing in Diagonal Splitting' ACI Structural Journal, Vol. 90, No.6, (Jul/Aug 1993) 108. Somerville, G., & Taylor, H.P.J., 'The Influence of Reinforcement Detailing oil the Strength of Concrete Structures' The Structural Engineer, Vol. 50, No. 1, (Jan.1972)
109. Somerville, G., & Taylor, H. P.J., Discussion of reference above The Structural Engineer, Vol. 50, No. 8, (Aug. 1972) 110. Somerville, G., 'The Behaviour and Design of Reinforced Concrete Corbels' Shear in Reinft)rce(l Concrete, SP-42, ACI, Vol. 2, Detroit, pp. 477-502, (1974)

Stroband, J. & Kolpa, J.J., 'The Behaviour of Reinforced Concrete ColumnBeam Joints, Part 1: Corner Joints Subjected to Negative Moments' ResearchReport 5-83-9, Delft University of Technology, Dept. of Civil Engineering, (1983) 112. Suidan, M., Schnobrich, W. C., 'Finite Element Analysis of Reinforced Concrete' ASCE Journal of' Structural Division, Vol 99, No. ST 10, 2109-2122, pp . (Oct. 1973)

317

1] 3.

Sundermann, W., & Schaefer, K., 'Tragfahigkeit von Druckstreben und Knoten in D-Bereichen' Deutscher Ausschuss fur Stahlbeton, Heft 478 Berlin(1997)

114. Swann, R.A., 'Flexural Strength of Corners of Reinforced Concrete Portal Frames' Technical Report TRA/434, Cement and Concrete Assoc., London, pp. 1-14, (Nov 1969)
115. Tan, K. H., & Mansur, 'Partial Prestressing in Concrete Corbels and Deep Beams' ACI Structural Journal, Vol. 89, No. 3, (May/Jun 1992) 116. Tan, K. H., & Naaman, A. E., 'Strut and Tie Model for Externally Prestressed Concrete Beams' A CI Structural Jounial, Vol. 90, No. 6, (Nov/Dec 1993)

117. Tan, K. H., Weng, L. W., & Teng, S., 'A Strut and Tie model for Deep Beams Subjected to Combined Top and Bottom Loading' The Structural Engineer, Vol 75, No. 13, (Jul. 1997) 118. Tan, K. H., Kong, F.K., & Li, W.W., 'High Strength Reinforced Concrete Deep and Short Beams: ShearDesign Equations in North American and UK Practice' ACI Structut-al Journal, Vol. 95, No. 3, (May/Jun 1998)
119. Tasuji, M. E., Nilson, A. H., & Slate, F. O., 'Biaxial Stess-Strain Relationships for Concrete' Magazine offoncrete Research, Vol 3 1, No. 109, (Dec 1979) 120. Taylor, C. P., Cote, P. A., & Wallace, J.W., 'Design of Slender Reinforced Concrete Walls with Openings' ACI Stiwctural Journal, Vol. 95, No. 4, (Jul/Aug 1998) 12 1. Taylor, R., Maher, D. R. H, & Hayes, B., 'Effect of the Arrangement of Reinforcement on the Behaviour of Reinforced Concrete Slabs' Magazine offoncrete Research, Vol. 18, No. 55, pp. 85-94, (Jun 1966) 122. Thurliman, B., Grob, J., & Luchinger, P., 'Torsion, Biegung und Schub in Stahlbetontragern (torsion, flexure & shear in RC girders)'

Inst. QfStructural Engineering, ETH Zurich, (1975) 123. Van Mier, J.G.M., 'Strain-softening of Concrete Under Multi-axial Loading' Dissertation, Eindhoven University of Technology, Holland, (1984)
124. Van Mier, J.G. M., 'Examples of Non-linear Analysis of Reinforced Concrete Structures with DIANA' HERON, Vol. 32, No. 3, (1987)

318

125. Vecchio, F.J., & Collins, M. P., 'The Responseof Reinforced Concrete to Inplane Shearand Normal Stress' Publ. No. 82-03, Dep. of Civil Eng., Toronto University, (1982)

126. Vecchio, F. J., & Collins, M. P., 'The Modified Compression -Field Theory for Reinforced Concrete Elements Subjected to Shear' ACI Journal, (Mar/Apr 1986)

127. Vonk, R.A., 'Softening of Concrete Loaded in Compression' Dissertation, Eindhoven University of Technology, Holland, (1992)
128. Wegrnuller, A. W., 'Elasto-plastic Finite Element Analysis of Concrete Slab Structures' Pi-oceedings of ICE, Technical Note TN99, Vol. 57, pp. 535-543, (Sep 1974)

129. Wood, R.H., 'The Reinforcement of Slabs in Accordance with a Predetermined Field of Moments' Concrete, Vol. 2, pp.69-75, (Feb. 1968) 130. Xie, Y. M., & Steven, G.P., 'A Simple Evolutionary Procedure for Structural Optimization' Computers & Structures, Vol 49, No. 5, pp.885-896, (1993) 131. Xie, Y. M., & Steven, G.P., 'Optimal Design Of Multiple Load Case Structures using an Evolutionary Procedure' Engineering Computations, Vol II, pp.295-302, (1994) 132. Yuri, Y. M., Alshegeir, A., & Remirez, J.A., 'Strut-Tie Model Design of Disturbed Regions in Concrete Structures' Proceedings, ASCE Structural Congress XII, Vol. 1, pp.233-238, (1994) 133. Yuri, Y. M., & Ramirez, J.A., 'Strength of Struts and Nodes in Strut-Tie Model' ASCE Journal of'Structural Engineering, (Jan. 1996) 134. Yuri, Y. M., 'Non-linear strut-tie Model Approach and its Application Tool for Analysis and Practical Design of Structural Concrete' Kvimgpook National UniversitY, Korea (1997) , 135. Yuri, Y. M., 'A Refined Strut-tie model Approach and its Application Tool' ICE Proc., Structures & Buildings Journal, No. 140, pp. 13-24, (Feb. 2000) 136. Zienkiewicz, O.C, 'The Finite Element Method' McGraw Hill Book Company, 3rd Edition, (1977)

UN I I-h I-,, -

319