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 noncommercial 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
June 2000
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
2.4
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 StressStrainRelationship 46' .................................................................. 3.3.6 Element Stiffness Matrix and Force Vector 47 .......................................
3.4 3.3.7 Numerical Integration ........................................................................ Nonlinear 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 .
4.3
4.4
Introduction
Concrete Constituent Behaviour .................................................................... 5.2.1 Uniaxial Compression ...................................................................... 5.2.2 Uniaxial Tension .............................................................................. 5.2.3 Biaxial 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 CompressionCompression Yielding ................................................. 5.4.1.2 TensionCompression Yielding ......................................................... 5.41.3 TensionTension Yielding ................................ ... .. ..... ...... ................. Concrete Nonlinear Behaviour ...................... ............................................... 5.5.1 Compressive StressStrain Relationship ............................................
5.4
5.5
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 ...................................................................................................
6.4 6.5
Preparation 236 of Models ..................................................... .... . ... ... ................... Formwork 236 ...................................................................................................... Concrete 236 ......................................................................................................... ReinforcingSteel 238 ........................................................................... .. ... .. ...... ... StrainGauges 238 ............................................................ .. .. ................................. Experimental Procedure 239 ................................................................................. DoubleSidedCorbels 240 ................................. .. .. ............................................... SingleSidedCorbel 240 ..................................... .. . ............................................... ComerJoints 240 ..................................................................................................
236 ...................................................................................................
8.4
8.5
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 AlNuaimi, FrancescoBasile, Igor Bojanic, Alan Cuthbertson,GraemeForbes, Domenico Gallipoli, Vassiliki Kochila, BoonTiong Lim, Andy Macauley,Andrey Shvidchenko, Xiaoya 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 2D inplane, 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 struttie 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, nonlinear 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 struttie methodcan produceresultsof comparable accuracyto nonlinearfinite 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 nonlinear 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
Md
MP M11MY9 My
Mn, Mt, Mnt
Area of steel Area of steel in xdirection Area of steel in ydirection 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
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 xaxis 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
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 struttie model addressed. for a given structural system.In many structuresthere may be various load paths available and hence no unique struttie 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 redistributioncan be allowed for, while preservingthe required from the structure. Details of struttiemodeltheory 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 struttie 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 inhouse nonlinear 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,struttie 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. testseries 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
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
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 45degreetruss 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 struttie 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
elastic stress paths of the structure which have been realised through finite element
The procedure developed by Alshegir and Ramirez (1992a), was further extended by Yun (1997,2000), by carrying out nonlinear 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 struttie 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 struttie model does not exceedthe ductility capacity of the structure.
Chapter 2
is linear A 'Bregion' 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 nonlinear,the regionis described for discontinuityor disturbance in this case.Such asa 'Dregion'. The term 'D' stands loads, bends disturbance distribution of stress a can occur at concentrated comers, and openings.Some typical Dregions of a structureare shown in figure 2.3.1(b). Because the strain distribution is significantly nonlinearin theseregions,Bernoulli is no longervalid. The designof Dregionshasin 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 Bregions can be readily corresponding appliedandonly the strut andtie modelsfor the Dregionsneedto be developed
Chapter 2
Chapter 2
VT
fig. 2.2 Truss Analogy in Cracked Beam
Dregio
Bregion
Dregion
compression tension
Chapter 2
h 4
h2
&h
Ea
h
th2 hl
h,
i) Geometrical Discontinuities
h
hl
2xh
T
7" ,;:!
IIIIIII1I11
t
ii) Statical and/or Geometrical Discontinuities
Chapici 2
(distance
2 @0.3b b Dregion
along section)
/b
I'll
1 I (a) .3
CFY 0
Section 11
Section 22 u
Bregion
5b 43
l
O.8b
2
10
Chapter 2
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
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
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
Tie 
P WD
I
12
Chapter 2
2.4
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 struttie model reason may detailing.
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 struttie 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
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 aOand b/aI, (fig 2.4.2(c)).
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. CCCnode: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
iiiiiiiiiiiiiIIIIIIII
smeared CCTnode
. 7T
compression tension
a)
sinearedCCTnode
singular CCCnode
b)
IttIIIIIIIIItIIIIIII
smearedCCCnode
sin,gular CCTnode
TZ
C)
15
Chapter 2
cr , fd
fcd
cc
a) Fan shape
b) Bottle shape
c) Pristil shape
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
1)
ii)
(N
I
d
onchofoge ,. j length
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
ii)
\
fig. 2.4.3(d) TTTnodes, compression strut replaced by bonded tie
,e zone
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
2. CCTnode: 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
2.4.4.1
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 struttie 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 microcracks have to
19
Chapter 2
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
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 multiaxial. on the existing state of stresswithin thestrut,theattainable compressive strength or effectivestress f, within the concrete the will vary. For the caseof biaxial compressiontension, increase. For compressive strength of theconcrete will decrease asthetensile stresses thecase the compressive of biaxialcompressioncompression, 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
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 stressstate 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 prestressed 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 biaxial 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 uniaxial or biaxial stress. Le al: 5.0.0. From the biaxial 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
Concrete St
t Condition
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 spandepth ratio less than 2.0. Struts fomdng arch mechanism Arch members in prestressed 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
6.5% 8.5% 2%
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 twodimensionalconcreteunder 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
22
Chapter 2
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 520%. This method is the advantageous sincethe useof finite elementanalysis, makesit is easyto determine stressstate characteristics of an individual strut. Using the biaxial 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 Dregion 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 uniaxial 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
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
0.4 0.3
0.2
0.1 0iiiiii
0
...........................
lower limit
10
20
30
40 (ollo2)
so
60
70
80
24
Chapter 2
0)
but with crodaad. trunsverse reinfOltMent (daypew) in the bellyregion uncracked. ptainconaPle
ItnemEnt
b)
ar
4+P,, 0
fig. 2.4.5(c) Dimensioning of Bottleshaped 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
25
Chapter 2
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 prestressing, and hoop
reinforcement
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 Dregion 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
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 struttie 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
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/bl):! 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 reentrant comers will cause stress concentrations which are representedby a singular node.
28
Chapter 2
Node Type
Proposed by
0.85 f'
0.65 f, ' 0.5 fc' 0.8 fe, fe':5 27.6 MPa (0.90.25f, '/69) fc' 27.6: 5 f, ': 569 MPa 0.65 fr' f,,' '? 69 MPa k fc"(A/Abf'5+
Compressioncompressioncompression 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:
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
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
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(iii). 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
2.4.6.2
Smeared nodes
It is nonnal for Dregions 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
0.21D
a,
lb, net
i) multilayered tie
cd
a3O
(7, (C OS2(p)
fig. 2.4.6. I (a) Stressesin Typical CCT Nodes, Schlaich et a]. (1987)
32
('11LIptel 2
4 T2
ic
C( 


rol. 
i) TCT Node
T
T2 c
lb
C3
C4
a3 4
(Tc3
CU
a4
C3
C4
cl C5
Co a,
Gc2 C2 tttttt
ao 4
C5 tttttt CFc5
a5
C,
C,
CvC I
33
Chapter 3
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
34
Chapter 3
The finite
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)
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
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 straindisplacementrelationship: 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, .
dVj 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
(3.7)
36
Chapter 3
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.
andload vector. equationsfor the unknown nodal 9 Solution of the resulting linear simultaneous variables. 9 Evaluationof the elementstresses.
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 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
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
11 i+
TlTli
42)
38
Chapter 3
2
401
4yIII' 4T i
7
corner
39
Chapter 3
These shape functions are part of the serendipity family (Zienkiewicz 1977) and are
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
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, EIHafez (1986) and W, elementswith five degreesof freedom (u,V,
40
Chapter 3
Bensalem (1993) used an eight node isotropic element with five degreesof freedom
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:
* *
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

(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
IDW(X, Y) ox ax
+,, (X, Y)
Yl
DW(X, Y) +Y(X, Y) Dy
(3.16)
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
(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 inplane 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
where [BI] is a 5x5 matrix which contains Cartesian derivatives of the shapefunctions
DNj D4 DNj
j 0
DNj Dx DNj Dy ax a Dy a
DNj ax ax Z, + aNi Dy ay O, q
(3.18)
DY
DNi' ax
DNi' ax
(3.19) DNj
l .a
ax
DTI
ay
OTl
DNj
ay J
DNj
Dy
DNi' ax DNj Dy
ax
Dy
ax
Dy
LOaq Ohl i
43
Chapter 3
I dz T
I dz T
(TY
44
Chapter 3
x
y
midplane
actual deformation
deformatioi assumed normalto midsurfac deformation I Fig 3.3(c) Cross Section Deformation of Mindlin Plate
MXY
MY
44
Qyz
MXY
x
Fig 3.3(d) A Typical Mindlin Plate (positive as shown)
45
Chapter 3
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
(3.22)
Dy T4 =8
Dy
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)
a hx
a Dy ax h Dyj
46
Chapter 3
0
(3.25)
E
_V2
2(1.2) 0
0 IV 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.
(3.26) 0 0.
0
In cases wherethe steelis positionedat an angle counterclockwisefrom the xaxis, to the global Cartesian the local modulusmatrix is transformed axis.
[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
(3.27)
andnumericallywritten as:
nmm
[K] =
Yal:
YaWjWk[B]T
(3.28)
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
(3.30)
48
Chapter 3
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
+1 m
lwif(4i,
+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=14 are given in table 3.3.7
49
Chapter 3
3.4
In a nonlinearproblem 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 nonlinear analysis cannot be found right away. In nonlinear 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 nonlinearanalysisare reduced. basedupon the incrementaliterative method.The generalprocedurein this method involvesadapting 8u until increment Au by iterativeincrements the total displacement The incrementaldisplacements tolerance. equilibrium is reached within a predefined 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 NewtonRaphson method.Within the method
50
Chapter 3
itself there are two variations, the second of which is known is the Modified Newton
Raphson method.
DR Ki = DAU
(3.34)
iteration, the stiffness relation shown above is In the normal NewtonRaphson (eq. 3.33) the predictionof the iterativeincrements evaluated everyiteration.Therefore is basedon the lastknownor predictedsituation,regardless of whethera stateof tonRaphson 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.
51
Chapter 3
d. displacements
As can be seenfrom figure 3.4.3 the unloading branch of the loaddisplacement 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 Arclength 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 outof balance residual forces until a desired level of hasbeenachieved. Convergence accuracy criterion canbe basedon the outof balance force norms,the displacements or the internalstrain energy. The methodadoptedfor the presentwork is basedupon the outof 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
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 nonlinear 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
force (f)
f"t
fim
fe.
force (f)
t fext
displacement (u)
fig. 3.4.2 Modified NewtonRaphson iteration
54
Chapter 3
41 0
1 11 +1/43
a,
2 +1
1
2
1 H 1
Id
+ _F 7
2 36
r31 , 0 + 2 36 430 1 + 2 36
IV
f
fext3 fext2 f,,, ti
cl u
c2 u u0
a) loadcontrol
b) displacement control
55
Chapter 4
4.1
Introduction
This chapterdetailsthe visualisationprocess usedin the formation of struttie 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
defining As a means themajorstress of automatically struttie 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 struttie 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
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
by Oncethe main stresspathshavebeenisolated,the struttie 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 2025%. The second and third case illustrate the effect of spandepth ratio on the load path behaviour. Struttie models from the visualisation processare illustrated in figure 4.2.3(d). In the case of resulting the deep beam, the resulting struttie 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
#+
\'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,
119
ft .o
adr IF
compression tension

t
1 .o
.:
e
.,
t.
a 2
lb
4a
#4
..
0: 0a
40
.,
.*6
vi e0
40 0
tf,
iii) rr=25%
58
Chapter 4
i) rr=O
.I*
ii) ir=10%
i1i) rr=20%
59
Chapter 4
Model Visualisation
qjjj
i) rr=O
"AA VIVA
Z4 61,WA
""A
t44
A V Ar 1 0 Ir 0 011;
AA V. W,. wxO W00
:'
,II
, W'X OWAO
.,
El compression m tension
ii) rr=15%
4
ni) rr=20%
Chapter 4
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
61
Chapter 4
In the case of the deep cantilevers, the struttie members are not always aligned
aroundthe exactlyalongthe main stress paths.This is dueto the fact that the stresses nodalzonesaresmeared.
4.3
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.
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
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 stressstaincurve. 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 reexamined (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 uniaxial stresses in their original directions, i. e. dowel action and bending of the bars is ignored
e The bars are elasticperfectlyplastic 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
Compressi
64
Chapter 4
(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 xaxis, 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)
(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
(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 xaxis, 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
(Mx*
MX)COS2()
+ (M
y*
My)sin
2E)
+ 2Mxysin0cosO : 0.0
(4.7)
(4.11)
Dividing by
COS20
=0 dtanO
(4.13)
d2F(O) dtanO
>0
(4.14)
(4.15)
d'F(O) 2o d tan
A+B(C)
2+ 2C(BB
)=O
(4.17)
66
Chapter 4
or AB
C2 =0 _
(4.18)
(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
(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 WoodArmer 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 nonlinear stressfield. It was found that in many cases, design from the nonlinear field helped to reduce steel congestion by 'smoothing' out the peak moments occurring at concentratedloads or supports.
67
Chapter 4
ff

MV
mly
M,
y my
mx
mxy
mxy
mx
MY mly
68
Chapter 4
(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)
hence My* = My + M,
or ,
MY*= MY +1ml
(4.23)
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
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)
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)
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
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
*
*
equation 4.30.
A schernatic representation of the design equations for bottom steel is given in figure 4.3.4(a).
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
Wma,
and
My*max
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 xdirection, M, *,,,,, is provided for, but in
the ydirection 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 .
72
Chapter 4
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
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 nonelastic layoutsfor deepbeams. to derivereinforcement patterns stress
73
Chapter 4
my/im,yi
M, +IM,,
/M, l
lm, ),

M, *= 0 MY*= 0
M'.
= M,
+I
M2, YIMY
MY *=0
Mmy=
My 2
'y
(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
setting:
Ayfy
nx xyY, ttttt
nn
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*
ay = a, sin
'rxy = ((Y2
(Y2 COS2 0+ (y
 al)cos0sinO
If tensile steel is to be provided, then (y, = 0.0, and equations 4.344.36 become:
(yx (yy
(Y2sin
2 0+(y
(4.37) (4.38)
75
Chapter 4
IrXY= (02)cos0sinO
(4.39)
hence:
2o
(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 midsurface 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
Ax, fx
(T)
ll'y
ny
11,
fig. 4.4. I (b) Sign Convention for Inplane Normal & Shear Forces per unit length
77
Chapter 4
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
Method of Solution Minimisation of (p,, + py) Direct solution Direct solution Minimisation of (p,, + py)
fcu
Direct solution Direct solution Direct solution Direct solution Direct solution
09G2 fcu
Px= P. =O
(ax *+GY
0,
Thus:
(yx * +(Yy +
78
Chapter 4
((Yy 
ax
4.4.2.2
For this situation,the minor principal stressin the concrete reaches ultimate strength, 4.344.36 be i.e. cr 0.0. Equations 2= a, < and can written as fe ,,,
20f, ax = Gl COS
,,,
sin
20+
Cyx
ay = a, sin
20_f
0+ (T COS2 y cu
thus:
ax+fc
,U
=Glcos
20+f
0+(; COS2 cu
(yy
ox
+fcu
(TX*
=(a,
+fcu)
Cos
2o
+ fc
a
*) + TXY2 = 0.0
(4.47)
79
Chapter 4
0,
Thus: I
T=O,
px =I/
fs'(axf
Irxy
fs'(cryf lTxyl)
px = 0, and py =I/
fs'((;
Ir Oyf
4.4.2.3
Mixed Cases:
.u
(y
X*=[fcu/2.0](I`
 cos20)+aX*
(4.48)
80
Chapter 4
y=
y*=[fcu
/2.01(1+cos20)+
aY
(4.49)
f, Txy , =
(4.50)
2, r
2 1/2
cos20 = [I C.
p,, =I/f,
Similar to case7;
81
Chapter 4
boundary curves were derived by equating the design equations for two cases.The
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
82
Chapter 4
Curve
Equation
Curve
r
Equation
I Irxyl
Cry I =2 F'rY1 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
10
x XY I 1TxYFT
 1/2) CY
CY
f +1
y y12
fcU lrxyl 4 11
ITXYI ITXYI
1/2)
00 12
ax IT 12 xy
fcu + IT I xy
fcU oixy, 4
83
Chapter 5
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 nonlinear important be The have to most of theseeffectsare: considered. effects
tensile cracking yielding of the steel nonlinear 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 stressstrain 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
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
Concretebeing a brittle material,there exists within the results of concretetests, a marked statistical scatter.Figures 5.2(ab) shows examplesof this for Young's in in behaviour Variation the tension. of one of stressstrain 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
The study of concrete under uniaxial compressionprovides a good premise from detail its behaviour under more complex stressstates. to which
85
Chapter 5
1 66
276 7 10
4681i0
;, 0
VflcPC IM
40 f, ,. MPG
100
50
ACI
Range formula for which ACI was derived code
A
30 (40.000 (w, /145), 0 . 1e) psi
145 E C( W
M PC
x 103
20
2##,
0,6
It,
& 10 &
QD psi
1000 0 ?('0
2000
BODO 9,100
130
0
IleC
fig. 5.2(a) Young's modulus E, vs. cylinder compressive strength f, ', (w= concreteunit weight, ACICommittee363 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
illustrated in figure 5.2.1(a). From the experimentalresults, the following basic observations canbe made: ', f, its 30% concrete stressstrain of maximum compressivestrength 9 up until
linear is relationship roughly ' in f, increase deformation is 30% to observed up stresses above gradual at ,a 0.750.9f, "t bending more sharply on approachto peak strength f, ' after reaching peak strength, the stressstraincurve 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 stressstrainbehaviour 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 stressstrainplot as the compressive strength is increased.
Numerous formulae derived from standard mathematical functions or from curve fitting techniques have been proposed to approximate the uniaxial compressive (1995). A (1970).
stressstrain response; Saenz (1964), Mansur et al. (1995), Almusallam review of the various proposed formulae
Examples of various numerical uniaxial compression models are presented in figs 5.2.1 (ce). 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 uniaxial compressive stressstrainresponsethat
87
Chapter 5
1.0
Axial strain Lateral 't strain
f',  67 N/MM2
10
12
fig. 5.2.1(b) Compressive stressstraincurves for concrete with different f, ' (Chen 1982)
88
Chapter 5
C max
f 2 C2 n%mx c2
'2)
T_ _
EC2 1
EO E
&1 0 I
c cu
2
CL
ul Ir
STRAIN
(z
10'3 )
89
Chapter 5
2A5 Curve
No, 1 2 3 4 5 Tyoe Granite Gravel Gravel Gravel Gravel Aggreqate Size 3/8 3/16" 3,, 83/16" 3/16B S. 7 3,83/16" 383/116
Age (monthi) 2 2 3 2 1
1,75
1 os 
Tensile sirain M
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
5.2.2 Uniaxial 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 manymicrocracks at the mortaraggregate 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 stressstrain testedin relationshipfor a numberof different concretespecimens
uniaxial tension is shown in figure 5.2.1(a). In general, the responseis almost linear
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 biaxial compression. strength increase of roughly' 16% f, " is observed under conditions of equal biaxial compression, and an increaseof around25% f, ' is achievedwith a stressratio of initial stiffnessin biaxial 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
In the case of biaxial tension to uniaxial tension is tensile strength similar ,a achieved. The stressstraincurves for biaxial and uniaxial compression are similar. For the tensioncompression case (fig. 5.2.3(c)), the compressive strength decreases almost linearly as the tensile stressincreases.
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 nonlinear 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
61
C, pp
E3
I
C2, C3 Q(i IL
F2
6; C) AG
2() 5cm(2in) (79 in) Ej, E. E3 +3 tensile strain +2 +1 01 2 3 mrrttm(QOOlirViN compressive strain
fig 5.2.3(b) Concrete biaxial behaviour: compression compression (Kupfer et al. 1969)
93
Chapter 5
"I 12
E 2'E3 E3
10 09Q8 
EJAE3
00 
/ El  

OL 4  Ei3 03 f N
;
t
05
05
10
15
5pp pp
295 kpicmz
4 200 psi)
010 C3
008
E,
6', 162
1/0
'1
CN Ow
1/1 1/Q55
QO4 comprvssive
002 strain
+002
CID4
006
008
alo
94
Chapter 5
95
Chapter 5
The subsequent development of the rotating crack model allowed the corotationof the crackwith the principal strainaxes.
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 loaddisplacement 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
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 nonlinear 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.
96
Chapter 5
Primary
Crack
Secondary Crock
CRACK
LINKS
Simile crack
Double crack
97
Chapter 5
or
98
Chapter 5
The biaxial yield criterion used in this work is basedupon the experimental work of Kupfer et al (1969). The yield surfacesfor concrete under biaxial 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,,,,
(5.2)
(5.3)
The following describes the derivation of constants a and b where. f. " is the uniaxial cornpression strength, previously stated as 1.16 strength (ft1f. "). is the equivalent strength under biaxial compression
5.4.1.1
a) For uniaxial
33
=a
(5.4)
T,y
0.0, =
99
Chapter 5
2f"
1.16
33
a LI 6(bf,
(5.5)
Solving for a and b, the biaxial compression yield equation is given by:
Toct
cyoct
0.1714fe, J'C'
(5.6)
5.4.1.2
Tension Compression
(Ty 111f. following the same procedure leads to:
r22
m =0.0 3 (1 + m)
(5.7)
5.4.1.3
TensionTension
For biaxial tension the following simple circular yield criterion is implemented:
2 CY2
0.0 1.0 = 
(5.8)
100
Chapter 5
07
Compression
01
Initial
yield
surface C 0 E C1.
fig 5.4.1 Yield Surfaces of Concrete under Biaxial Stress (Chen 1982)
101
Chapter 5
5.5
(5.9)
I.
da dF2.
(I
Ec
 V(X)
On reaching the peak stress ap at strain Fp,the slope of the stress strain curve
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
E, is the initial modulus of elasticity of concrete for uniaxial 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 stressstrain 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 stressstraindiagram. Subsequentloading surfaces are assumedto have the shape of the limiting yield surface. The intermediate surfaces are
repIacing
/, ". Tile following equation was proposed by Johnarry (1979): the ultimate strength.
tensile strength, E, =
concrete elastic modulus, Ej = instantaneous elastic modulus. The instantaneous elastic modulus is calculated up to peak strain FP 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
5.5.2.1
Single Cracking
or tensioncornpression, are
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
104
Chapter 5
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 .
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
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
(5.18)
c2s2
cs
and
[T]
S2 2CS
C2 2CS
CS 2c2 S
(5.19)
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 loaddeforination 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 stressstrain curve for concrete. The second method involves modifying the stressstraincurve 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 stressstraincurve 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 oneway spanning slabs with different have shown that the effect of tension stiffening decreaseswith increasing ratios steel
106
Chapter 5
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 < Fcr
then
if
Ccr!Ci:! C2F,,
then
a=
Cl ft
if
C2Fj> C2F,,
then
G=O
where cyand Fiare the local stressesand strains orthogonal to the crack, the cracking is The f, /Ec the the tensile strength of concrete. Fcr 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
Cracking
concrete stress
steel stress
CY
ft
cl
Fcr
C2 c,
108
Chapter 5
sheardisplaccnient
1.0
Fcr
109
Chapter 5
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 (Ff/ Fcr) (5.20)
defined below: fictitious is the the to as Ff strain normal crack where
cf=
Cos
Ocr +yy
sin OcrCosocr
0, ft/Ec, is the Fy the Fx, yy is and and are strain inplane strains F, critical cracking =
5.6
Modelling of Reinforcement
The modelling C, of steel bar behaviour is less complicated than that of concrete since its behaviour is largely uniaxial due to the onedimensional nature of reinforcing The in bars typical of characteristics stressstrain 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 bifinear representation is sufficient to model the behaviour of the steel. and this can be modified to take account of strain elastoplastic hardening (fig 5.6(b)).
A(T= ESAc
110
Chapter 5
On reaching the yield stress fy, the incremental stress relationship becomes:
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(ce))
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 stressstrain 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
Chapter 5
In this method, it is possible to model the steelconcrete 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, ElMezaini 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 nonlinear behaviour such as cracking and dowel action.
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
900
800
700
600
CY
500 0 400
300
200
100 01i 0
50
100
150
200
strain
x1OA3
fI,
fs
A
cy C,
A
Ey F,
i) Bilinear (hardening)
n) ElasticPerfectly Plastic
113
Chapter 5
Y
CYY CY, y
01
////
//////////// 01
o' 001
114
Chapter 5
Modelling
independently. The bar be evaluated each element can stiffness contribution of formulation for inclined reinforcement proposed by Ranjbaran (1991) was
5.6.3.1
At a typical point P in the reinforcement (fig 5.6.(e)), the strain in the concrete with is follows: XY 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 RrR
(5.24)
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
)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
xx'
+m,
Fyy
(5.25)
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
fB
u. iT J, BVDBj AQ u 21
iT DBj AQ
fB
il , iT ful B DBj AQ vv
iT DBj AQ
(5.28)
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: .
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
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
117
Chapter 5
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
and hence:
2 ir 2 (1
)(L +j2 12 11
(
+2(JIIJ,,,
22
) ar 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:
(5.37)
118
Chapter 5
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,, ) riI ... ... mi  (ri ) (ri )(ri ) (ri r,, ri+l riI  r, ... ... (rrj) (r
(5.38)
Hence for a straight reinforcing element, equation (5.36) may be written as:
22 I
r(h
+I
(b + a
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
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
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
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 inplane 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 nonlinear phases while larger increments may require more iterations. It was shown by AbdelHafez ( 1986) that the effect of increment size on the resulting solution is not significant.
120
Chapter 5
Concrete Properties
Steel Properties
E,=21500
N/mm 2
, T
E,=214000
2 N/mm
2
fy=460 N/mm
For numerical analysis, a symmetrical quarter of the slab was analysed. Firstly, the
and W
elements. and
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
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 loaddisplacernent 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 stressstrain curve. From figure which 5.7.1(e), it can be seen that the presenceof tension stiffening in the model improves the loaddeflection responseat the service load level.
121
Chapter 5
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 loaddisplacement 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 loaddisplacement characteristics is
observed. The steelstrains 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 (hi).
122
Chapter 5
615
970
615
545
""
107( 2160
a) plan view
""
545 2100
ci.
145 145 160
145
1
c.1.
322 11
77 Id 1
1401,
300
LI 1
155
1
c.1.
figs. 5.7.1 (ac) Details of Hago's Model number 3 (all sizes in rnm)
123
Chapter 5
1.4
1.2
m1 0.8
2 0.6 M 2
CL CL
*0106
'o 0.4
0.2
0
0
0 10 20 30 40 50 (mm) 60 70 80 90 central deflection
1.2
Co
0.8
0.6
CL CL
0.4
0
experimental ts *without (c1=0.5, c2=20.0) 0ts 6 ts(c1=0.5, c2=15.0) Ots (cl=0.5, c2=10.0)
0.2
fig. 5.7. I (e) Slab n.3 testedby Hago; effect of tension stiffening
124
Chapter 5
1.2
0.8
c
0.6 m a,
m 0.4 rL m
0 006 0
0.2
OL 0
10
20
30
40
50 (mm)
60
70
80
90
central displacement
1.2
.T0.8
0.6
CL cc
0.4
* 0.2 0
fig. 5.7. I (g) Slab n.3 tested by Hago; bottom steel strains at centre
125
Chapter 5
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
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 loaddisplacement curves obtained in detailed figures 5.7.2(de). A are numerically and satisfactory experimentally
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
914.4
a) Plan view
a.
c.1.
25.4crs
b) reinforcement
25.4 crs
c.1.
c.1.
V
128
5 Chapter
rL 6000 4000 2000 0 02468 d1splacement(mm) 10 experimental C] without ts br present analysis *
H
12
'o
129
Chapter 5
layouts for Details S8 the of reinforcement eachof theseslabsare chosen. and were
5.7.3(ad) 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
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
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
loaddisplacement The factor B=0.4. relationshipfor eachslab is displayed retention betweenexperimentaland in figures 5.7.3(eh).It can be seenthat a good agreement numericalmodelswasachieved.
130
5 Chapter
C.'.
76 crs
33 crs 76 crs I
5OCTS
305 crs
30 crs I. C.
b) Model S2
c.1.
105crs
105crs
90 crs
40 crs
.
ci.
1 C.
d) Model S6
c) Model S8
figs. 5.7.3(ad) Slabs tested by Taylor et al; reinforcement details (all sizes in mm)
131
5 Chapter
140
120
100
80
60
40 * 20 I present analysis
D experiment
00 0
10
20
30
40
50
60
70
80
90
140
120
100 z
0 80
cc 7a 0
60
00
20 40 60 central deflection (mm) 80 100 120
132
5 Chapter
120
100
80
60
40
20
10
20
30
40
50
60
70
80
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
133
5 Chapter
The first threebeamsin the serieshad the samespandepth 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 (ad). 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)
f,,, (N/mm2) ft (N/mM2) Ratio Span/Depth Ratio ShearSpan/Depth DesignLoad Pd(kN) SteelProperties
134
5 Chapter
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 midspan 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 midspan 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(ef). From the loaddisplacementcurve, 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 5.7.4gh) (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 loaddisplacement 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 loaddisplacement beams TGRAS2 and TGRAS3 is less ductile (figs 5.7.4(1m)).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
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 spandepth was much ratio of beams,less brittle loaddisplacement 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(pq)). 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 loaddisplacement and loadstrain plots (figs 5.7.4(no)). Significant in (up to the the the was observed midspan 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
425 900
150 +
425
425
bl
150
lid
425
26
600
26
300
26 26 28
15Q 35 0.
YOU
01d
YOU
N1
425
150 A J.
425
425
150 +
425 .1
900 26
600
26
300 IS 35 0
7UJ
26 26 28
you
137
5 Chapter
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
NI1
N1
325 17 35. 01
138
Chapter 5
II
II
II
Ii I.
3.5
4.5
1.6
1.4
1.2
1,:
0.8 cc 0 0.6 CL cc 0.4 0.2
?"+
0.5
1.5
2 strain/yield strain
2.5
3.5
139
5 Chapter
_____
_____
_______
_______
______
______
II
I
___
i
___ ___
I /I
I I
_ _
II / // 1I II
\
___ 
I I
I I
,, II
II II
I\ 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
\
\
05Pd)
140
5 Chapter
.
/ 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 \
141
5 Chapter
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
142
5 Chapter
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
OL. J 0
0.5
1.5
2.5
central d1splacement(mm)
1.4
1.2
M1 w 0 c
0.8
cc
II II Ii I.
0.2
OC
0.5 1 1.5 central displacement (mm) 2 2.5 3
143
5 Chapter
1.4
1.2
a m
"I
C)
0.8
II
Ii II
0.6
0.
M 0.4
0.2
1.2
o 0.8
0.6
ta
rL
00 0
144
5 Chapter
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
_______ _______ 
1
1 %
wo
11
1%
11
11
11
11
,%
fig. 5.7.4(q)BeamTGRAS4:NumericalCrackPattem(P=0.725Pd)
145
5 Chapter
corbel resultingfrom the natureof its eccentric with testingthe onesided associated
loading. The example chosen for this analysis was part of a series of tests carried out
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
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 photoelasticanalysiswas 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
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 columncorbel 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 elastoplastic the model used van In MohrCoulomb 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(de) 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
1000
50
800
50
1000
600
Chapter 5
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
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
DV 0.4
1.OOE+02
2. OOE+02
3. OOE+02
4. OOE+02
5. OOE+02
appliedload(kN)
Io0.5 1 1.5
2.5
stralWeld strain
150
Chapter 5
Jr )c
g)
h)
figs. 5.7.5(gh) Niedenhoff Corbel M2/B2: Comparison between experimental and numerical crack pattern
151
Chapter 5
in frames. Likewise, beamcolumn behaviour a the connections portal of particular in this carried out orderto testthe numericalmodel's was structure numericalstudyof
behaviour. Previous beamcolumn 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
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 triaxial stress may occur in very wide frames such as
152
5 Chapter
hence frame 70 thick the generation of the tanks was only mm. and etc, retaining walls, triaxial stresses was believed to be negligible.
In the present analysis the frame was analysedusing displacement control (chapter 4) loaddisplacement be to the of relationship obtained. part which allowed unloading The experimental loaddisplacement 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 1618 kN. This cracking in The the the comer. stresses principal stress plots (fig of caused redistribution 5.7.6(fg)) 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 biaxial 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 midspan at a load level of 24kN (fig 5.7.6(d)).
153
5 Chapter
2@ 2(6)
A
C.'.
120 5@ 104 2(6) 70 sectionAA c C J 170 2@ 20 95 120 Luo160 sectionCC covcr on stirrups5mm stirrups2.8 120
870
154
5 Chapter
30
25
20
13 cc 2 15
C)
'1,1
10
experimental
10
12
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
. I 12
fig. 5.7.6(c) Frame A7 testedby Stroband & Kolpa; Concrete stressesat inner comer
155
Chapter 5
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 midspan
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
156
Chapter 5
Modelling
It o
0 00 VI * 0 '410 o.
Compression IC11"Imi
'4
..
%:
J
.r.
9iI
*'
157
5 Chapter
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, underprediction <0.1, of will result ultimate subsequent of values for B. 0.4 adopted was a value of analysis,
158
6 Chapter
SlabDesign
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 spandepth ratios stipulatedin BS8110Part 1. chosen
The direct design approach, described in chapter 3, was used to derive the layouts. For SMI5, design derived from the required numerical slabs a reinforcement steel areasat (rrO) 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 redistribution 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 2140x 100 + centralcolumn Square simply supported 2l4Ox2l4OxlOO on adjacent sides+ columnsupportat oppositecomer Square supported on four comers 910 x 910 x 45
6.2
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
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 6.2.26.2.4.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 rrO. 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
N1
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 rrO. 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)
163
Chapter 6
Slah Design
i) (ii= 0)
164
Chapter 6
Slab Design
165
Chapter 6
Slab Design
41%,
0. +
.,
+ 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
166
Chapter 6
Slab Design
167
Chapter 6
Slab Design
Ii
I1
I I 1 1
f 1
f f
f7 f
tt
tt
i f 4 + + + + + ++ k*
kxj,
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
+ +
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,
je
A
dp
(ii) rr = 12%
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
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 xdirection, the slab in figure 6.3.1(a),whereAs,, and Asxtop,refers to top steel in the xdirection.. 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 rrO 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)
(rrO)
(rr=40%)
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.
1749.2 1241.4
In the numerical analysis of each model, a total of 30 increments was used, with an OlPd initial increment Of OlPd in the elastic stage,then 0.05Pdduring cracking and O, towards ultimate load. The loaddisplacement 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 midspan 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 midspanat 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 momentcurvaturerelationship is The sectional behaviour as represented
detailed at the centre and near the load point in figures 6.3.1(fg). The moment
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 curvaturerelationships are similar.
171
Chapter 6
Slab Design
I
, Ll
172
Chapter 6
SlabDesign
I. 
150crs
75crs
150crs
i) A, at bottom(rrO)
150crs
75 75 iq 0!4 150crs
173
6 Chapter
SlabDesign
150crs
75crs
150crs
H4
>l(
150crs
75 75 150crs
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)
0.00
(0.00)
5.10 (0.00)
14.38
(0.00)
0.00 (0.00)
2.96
(0.00)
0.00 (000)
0.00
(0.00)
0.00 (0.00)
0.00
(0.00)
0.00 (0.00)
0.00
(0.00)
0.00 (000)
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
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)
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)
6.23 (0.00)
21.60
(0.00) '(0.00)
0.00 (0.00)
3.12
(000)
0.00 (0.00)
0.00
(0.00)
0.00 (0.00)
0.00
(0.00)
0.00 (0.00)
0.00
0.00 (000)
0.00
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
1.2
m1
0.8
.00.6
0.4
0.2
0 c
0.8
0.6
0.5
1 strain/yield
1.5 strain
2.5
6 Chapter
SlabDesign
1.2
0.8
0.6
0.4
0
0 0.5 1
1.6
2.5 Ry/Ft
3.5
4.5
1.2
0.8
0.6
0.4
0.2
*U
(rr0) (rr40%)
fig. 6.3.1(g) Slab SMI Principal Moment near load point vs. Load
178
6 Chapter
SlabDesign
0.5
1.5
2 RY/R
2.5
3.5
0.2
0.4
0.6
0.8
1.2
1.4
179
6 Chapter
SlabDesign
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 rr30% 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 rrO rr30%. observed was respectively areas
SteelVolumes
(Cm 3)
(rr=O)
(rr30%)
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 rrO. steel around provided comers can steel
At the centre of the slab, M,, and I/Ry were calculated as 58kNm/m and 0.0001mnf I respectively, for each design. The momentcurvature 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)ir=O
182
Chapter 6
Slab Design
(N
cc,
ul
l
r
r_
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 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
40crs
75crs
75crs
14 4 4
I 50ci s
40crs
1
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 (503)
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)
9.33 ).()( 29.62 (50.3) 45.93 (50.3) 57757 (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)
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
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)
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
0 c
0.8
0.6
rL 0.4 CL m
0.2
0 0
(rr=O) (rr=30%
0
0234567 strain/yield strain
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
Iss
Chlplcl 0
1.2
0.8
0.6
0.4
0.2
(rr0)
J)
(rr=30%,
0uI
023456
7 Ry/R
1.2
0.8
0.6
0.4
0.2
(rr=O) (rr=30%)
199
Chapter 6
Slab Design
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(iii). 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
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 loaddisplacement 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 midspan 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 110 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%,
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 mornentapplied load
relationship at the load point shows that eventual softening occurred in the rr=30% design at an ultimate load of 1.2Pd, (fig. 6.3.3fg). 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
fig. 6.3.3(a) Slab SM3, Symmetrical Quarter, Numerical Steel Areas (MM2)
192
Chapter 6
Slab Design
75crs
150crs
75crs
150ers '
Nid
i) A, at bottom(rrO)
NI
I
.
50crs 75
300crs 1 75
75  300crs k 014
75 50crs k 014 0
193
Chapter 6
Slab Design
I
300crs
37.5crs
300crs
75crs
37.5crs 014
300crs
iii) A, at bottom(rr30%)
iv) A, at top (rr30%) 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)
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 (000) 0.00 (0.00) 0.00 (0.00) 0.00 (000) 4.73 (0.00) 13.24 (14.2) 18.73 (28.3)
5.32 (0.00) 0.00 (0.00) 0.00 (000). 0.00 (0.00) 0.00 (000) 6.28 (000) 13.24 (28.3)
14.24 (0.00) 2.72 (0.00) 0.00 (000) 0.00 (0.00) 0.00 (0.00) 0.00 (000) 5.70 (28.3)
26.28 (28.3) 6.50 (0.00) 0.00 (000) 0.00 (000) 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 (iL0) 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)
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 (rr30%) 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(rr0) (rr30%)
10
12
0.8
1.2
C
ID
0.8
0.6
0.4
0.2
I *D
(rr0) (rr30%)
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) (rr=30%)
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
fig. 6.3.3(i) Slab SM3: Principal Moments near column support vs. Load
199
Chapter 6
Slab Design
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. top
3356.2 470.4
1.23 1.43
From figure 6.3.4(b), it can be seenthat some of the peak moments around the comer 'smoothed' by load the time rr12% is reached. A 23% increase were out points and
from rrOto rr12%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 loaddisplacement response(fig. 6.3.4d) shows that an ultimate load of 1.2Pdand be It the that 1.31? can also seen and rr12% respectively. at rr_O achieved was d behaviour of the model designedat rrOwas 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 rrO 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 u0. For the second design , at rr12%, 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(gh), 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 rrO,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 rr12% somezones from the momentcurvature 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
Chapter 6
Slab Design
) c::
42
C N
Cj ,0
4
203
Chapter 6
Slab Design
150crs
75crs
150crs
75
75 150crs i4 014 75crs 014 4 I 50crs
50crs 75
150crs
75 75
0i4i4
i4
150crs
75 50crs
014 1.
204
Chapter 6
slall Design
300
37.5 crs
50 crs
75crs
75crs Oio
50 crs
37,5 crs 0! 4
300
L1
450
75crs
75crs
450
75 30crs
(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)
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) 9.91
(0.00) O. (JO
(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
).00) (0.00) ).()( 2.85 0.00 0.00 ).0( ).00) (0.00) 3.21 17.39 0.00 ).()0) (50.3) (000) 33.40 1 19.22 1 5.65 (503 (50.3) (50.3)
fig. 6.3.4(c) Slab SM4: Numerical & (Provided) Steel Areas in mrn
206
Chapter 6
Slab Design
0.00 1 0.00
(25.2) 91.38 (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)
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
13.28
0.00
0.00
0.00
0.00
26.25
0.00
0.00
(0.00)
oo) (O.
11.09 (000) 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
(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 ). ()(
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
1.2
C,
o. 8
0.6
CL OA to
0.2 0iiii
0 0.2 0.4 0.6 strain/yield 0.8 strain 1
(rr= 12%
1.2
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
2.00
2.50
3.00
fig. 6.3.4(f) Slab SM4: Top Steel Strains near colunin SLIPPOIt 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
2 2
0.5
0.4 0.3 0.2 0.1
0 EI
0.2
0.4
0.6 applied
1.2
1.4
209
Chapter 6
Slab Design
455 I17.
455
This slab was centrally loadedand pinsupported 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)
(rrO)
(rr12%)
Total
A. bottom
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 rr12%, while an increase of 330% is observed for top steel at rr12%. 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 rrO,(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 loaddisplacement 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 rrO and rr12% 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 midspan 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). O8Pd
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 rrO. 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 0.0002mm1 respectively, at rr=O. In the second design at this area, M,, and I/Ry were 8.5kNm/m and 0.0002mm1 respectively.
211
Chapter 6
Stab Design
00
C'4
212
Chapter 6
Slab Design

65crs
120crs
32.5
120crs
65crs
i) A,,at bottom(ffO)
1
152.5
it NI4
bl
(ii) Ar,at top (rrO)
fig. 6.3.5(b) Slab SM5: Steel Layout, all sizes in mm, all bars 6mm.diameter
213
6 Chapter
SlabDesign
32.5crs
195
130
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)
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
(0.00)
0.000
0.000
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)
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
15.030 (14.2)
86.860 (84.8)
11.760 (14.2)
17.026 (84.8)
5.165 (0.000)
0.000 (0000)
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)
fig. 6.3.5(b) Slab SM5: Numerical & (Provided) Steel Areas in mm2
216
Chapter 6
Slab Design
LM .
1.2
0 c Im a 0.8
0.6
40
0.4
0.2
Drr1
2%J
0 0.00
0.50
1.00
1.50 $train/ylold
2.00 Wain
2.50
3.00
3.50
1.2
0.8
0.4
0.2
00 0.00
ii
1.00 2.00 3.00
4.00 strain/ylold
5.00 $train
8.00
7.00
8.00
9.00
217
Chapter6
SlabDesign
1.2
0.8
0.6
0.4
0.2
0 001234567
1.2
0.8
0.6
0.4
0.2
Li 0
0.2
0.4
0.6
0.8
1.2
1.4
1.6
218
Chapter 6
Slab Design
1.2
0.8
0.6
0.4
0.2
0 02468 Ry/R 10 12 14 16
0.8
0.6
0.4
0.2
n
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 applied load/design load
219
Chapter 6
Slab Design
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 xaxis roughly from 45' horizontal. the transformed providing orthogonal steel
Steel Volumes
(CM)
(rrO)
(rr=12%)
Total(12%) Total(O)
Total
A, bottom
Total
53.6 77.5
41.4 64.1
53.6 179.7
1.0 1.03
In this case,the numericalsteelareas at rrOand 12%are equal,which is in contrasts increase The SM5 was observed a'20% where at rr12%. resulting slab with in figure is 6.3.6(a). layout The loaddisplacement shown numerical reinforcement 6.3.6(b). in figure An load lIPd 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. 6.3.6cd). the the top comer support at the at same steel of Inspection of the moment curvature relationship at the comer support shows that 6.3.6ef). 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.6.3.6gh).
220
Chapter 6
Slab Design
32.5crs
32.5crs
195crs
195crs
65crs
22crs
I+
22crs
65crs
(ii) A, at top (rr12%) 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)
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 (rr12%) 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
(113.08)
(28.27)
(0.00)
(000)
(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
(000)
0.00
(000)
0.00
(0.00)
0.00
(0.00)
0.00
(000)
0.00
223
Chapter 6
Slab Design
1.2
0.8
0.6
0.4
0.2
20/6 1
2) .! 0.6
CL 40
0.4
0.2
*rr12%
04 5 10 straln/yield 15 strain 20 25
0.8 c 2 0
0 13 lp 0.6
0.4
0.2
I"
04iii0 0.2 0.4 0.6 0.8 strain 1 1.2 strainlyield
rr=12%
1.4
1.6
224
Chapter 6
Slab Design
0.00
1.00
2.00
3.00 Ry/R
4.00
5.00
6.00
I0
0.20
0.40
0.60
0.80
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 0rr=12% I
0.2
0.4
0.6
0.8
1.2
226
Chapter 6
Slab Design
6.4
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 rr30%,wherethe natureof the evolvedmeshled to a large increase
load Given designer from the a predetermined 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)
(rrO)
(rr15%)
Total(15%)
Total(O)
Total
A. bottom
Total
53.6 77.5
46.3 66.7
62.2 839
1.1 1.1
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 midspan 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. lfg).
227
Chapter 6
Slab Design
(il) rr=15%
fig. 6.4. I (a) Slab SM7, principal moments
228
Chapter 6
Slab Design
30
ci
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
65
65
22crs
bb 22crs
65
65
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)
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 (000)
0.00 (0.00)
0.00 (0.00)
0.00 (0.00)
0.00 (0.00)
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)
1.2
0.8 .0 V
0 "a
CL m m
0.6
0.4
0.2
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
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
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 rrO 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 rrO. 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 predetermined 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 redistribution accommodate Property
(rr) 0
SMI
40% 0
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
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
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 11Pd column support 0.7 Pj 1.2P, j centre 1.0P', 53.5 77.4 1.1 0
4587.4
0.81"d Centre
235
Chapter 7
Experimental Program
Five structuresdesignedusing the struttie 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 7day 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
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
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), Lcylinder P is the and where from The load. each control specimenwere averagedto give the results ultimate in valuespresented the next chapter. experimental
Universal Olsen Testing Tinius Machine fitted the using with an Stype 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, stressstrain
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
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 microcrack 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
7.3.2 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 setupis 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.
testing.
240
Chapter 7
Experimental Program
2P
1'Top loadliq, Platcli ()I Losenhausen Machinc
N
I1
Pl
241
Chapter 7
Expenniental Program
It'
242
Chapter 7
Experimental Program
i 9 E
bolted to tab floor 625 IOOXIOOXIO hollow sections hydraulic. jack & load cell .,k 50
1080
41 <1
III]
11511
Expeimiental Setup
243
Chapter 7
Expeninclital Program
i) Plan view
i) Section AA
't,
/I
im
44
Chaptcr 8
This chapter details the design by strut and tie models of a number oftypical Dregion type structures. The main objective of this work is to assess(lie SUItabIlItyOfthe Struttie method in achieving the required performance from a designed StiLICtUrC.
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 struttie frame corner joints. All of these designs were assessed through nonlinear analysis. in joints laboratory. tested the to this, tile were physically and corner corbels in addition Details of the experimental setup were given in chapter 7.
As in the case of slabs, the service and ultimate load characteristics of the StiLICIIIICS, is important It from investigated. that the structures resulting the struttie 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 spandepth 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
by BS81 10. In all designs, adequacy of anchorage and 1)()11(1 (11C ()1,
245
Chapter 8
8.2
Deep Beam BI
This model was a typical simply supported deep beam with an effective spandepth 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, uniaxial cube crushing strength of
concrete, uniaxial 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 struttie 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 jc.,,,, jltjjjg by introducing (fig8.2a(iii)) ties the along strut at the third points. model struttie 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
Member 1
Force (kN)
A. required
(MM2)
A, provided
(MM2)
No. of Bars
A, provided A, required
4 7
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 biaxial 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(dh). From the the are shown The results of nonlinear analysis
figure 8.2(d), it be loaddisplacement 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 midspan at
1.25Pd,which is close to the load predictedby the struttie 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 biaxial I. this initiated tensionconcrete zone a state of at was
is 0.8f,, hence The idealisation stress of an ultimate reached. struttie 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
in from figures be 8.2(ij), However, the the stresses stress plots seen can as plates. by diagonal to those the struttie 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(ij). the principal stress plot that the largest can seen given are concentratedalong the outer edge of the bearing plate. compressive stresses
248
Chapter 8
1) (110)
compre "
ii)
(rl)
El
000,
F0 ol 40
n5
...
n,
tic
fig. 8.2(a) Bearn B 1: Elastic Stress Patterns and Strut Tie modcl
249
Chapter 8
, q)K
Igo
120
1010
60 '30
IIII0d IT 11II, bid 4+04
2(1()
40
80
90
235
80
SO
235
90
10 .
250
Chapter8
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)
.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:
1onumed7cal I
0.8 2 0.6
cL 0.4 a
CL
251
Chapter 8
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
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)
Stresses
252
Chapter 8
0.8
0.7
Al
:CB
0.6
0.5
U
0.4 U) 0.3
"
("
"
0.2
Li
0.2
0.4
0.6
0.8 load
1.2
1.4
applied load/design
1
OE 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
253
Chapter 8
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
StressStatc
tens i oncom pressi on approaching peak tensioncompresmon approaching crushing
254
Chapter 9
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 :Sc Schematic h: at ic
Pd
Pd
Material Properties (Design) P,1=350kN concrete: 2 f,,,=50N/min 2 f,=3. ON/in111 E, =21.5kN/m 1112 steel: fy=50ON/nim 2 E, =200kN/mn 12
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
thickness=250mm
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 fiom 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 biaxial 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
255
Chapter 8
compression tension
i) rr=0
75 1_1
n2 ja amo
II
i\)
fig. 8.3(a) Corbel C2A: Elastic Stress Paths and Struttie illo(tel
256
Chapter8
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
* 02"; Fc2
jo
257
Chapter 8
2 0.5fcu, 24.9N/mm
258
Chapter 8
180 50 3012
L 300
0
strain gatioc
125 i) elevation
200
"Is'
",
Wwwwow
259
cluipicr 8
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
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
1.2
0.8
02
IIj
V 0
e za 0 '0
CL m
0.1
09
261
Chapter 8
7 /_.
fj,
T
I
Ig8.3(k) ,iij1.1
6 22
Chapter 9
I N/nun
I (H(nun
t 1) N/mm
65m
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
263
Chapter 8
1.2
1 08
0.6
I
OA
02
0,1
F)
0.35 03 025 02 0.15 0.1 005 0 0 0 04 06 2PPII*d load/design load 08 1
Chapter 8
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
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 corbelcolumn 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 struttie 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(tv). 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
columncorbel 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. struttie of
265
Chapter 8
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
(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 (00) 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
300
35C
F014 11 10
125 i) elevation
200
1.2
0.8
2 0.6 m T 'a
CL
m 0.4
0.7
0.8
0.9
displacement (mm)
fig. 8.3(t) Corbel C2A: Comparison of Direct Design and StrutTie Design
268
Chaplei 8
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
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
269
Chapter 8
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 corbelcolumn stress concentrations reduce order in Geometric C2A. the given model and material properties are previous occurred in the table below. Schematic
P', Pd
50 150 150
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, =200kNhnin 2
200 k 1 4 r19 I250 1,200,1 I
thickiiess=250iiim
previously,
calculated
as 3010mm
bars main
reinforcement
reinforcement.
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(ab).
270
Chapter 8
(1.42Pd) occurred
was 650kN,
junction. loading,
0.7Pd. The largest of these cracks was 0.1 illin in the diagonal strut region was observed
on successive
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
main steel at the upper columncorbel strains show a good comparison significant widening
and experimental
of the cracks both at this point and in the diagonal fig. 8.4(h). The corbel at ultimate
strut, as call be in
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(mn) 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 struttie 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 colurnncorbel 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, struttie more in a predicted elernent analysis of this nodal region was carried out using a displacement control displayed The in figures 8.4(op). The funnelling of the this results of are analysis. be from the principal stressplot. can clearly stresses seen compressive
271
X,
180 50
150
150
350
bi 4
0
125
i) elevation
200
250
ii ) Side
c1cvatiOll
V
A
Formwork
272
Chapter 8
273
chaptel
8SI
ILI I&IICI
)C I '[ I
a Co 0
c1 AM 0
0.8 m
' CL 0.6
'
0.4 0.2
numerical
0.2
0.4
0.8
1,2
1.4
1.2
2 C 0,
08
0.6 CL
a
0.4
tnu
0.2
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
0.8
1.2
1.4
Ij
274
Chapter 8
1.2
Co
C
a) 7A
0.8 (12
0.6 1
CL
CL cc
0.4
0 0d2
dl
0.0005
0.0005
0.001
0.0015 strain
0.002
0.0025
0.003
0.0035
275
Chapter 8
7 =
(j=5.8
100111111
7.8 N/nim 2
65ninli
500
350
*I
B (129.5,391.1)
A (120.3,342.1 )
125
325
276
14
12
la
ob
08
06
04
0,2
B
1
08 LL 0.6
0.4 A 0.2
Bj L77077
04 0.6 0.8 load 1 1.2
Z;
277
( '11kj)kl
Yt III &II"I),.
I,,II
11 1
120nim
70nim
compression tension
93.3mm
80111111
'sSLItc
El
m
278
Chapter 8
8.5
Corbel C4A
C4A was a onesided 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/nun 2 f, =3. ON/min 2 E, =21.5kNhnni') steel: fy=50ON/mm 2 2 E, =200kN/niin thickness= I 50nini Material Properties (E xperimental)
1,250j00j F,
concrete: f,,,=49.6N/nun2 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 struttie 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(ab)
279
Chapter 8
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(gi). 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 columncorbel 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(nin). 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
lln
150
150
0
crs
350
00 clS
1.4
250
1014 200
ig
lb
i) elevation
.. ___________________ ____
281
Chapter 8
282
Chapter 8
283
Chapter 8
1.4
1.2
1
C
0.8
O6
0.2
0.4
0.6 strain/yield
0.8 strain
1.2
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
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
284
Chapter 8
77
(T=7.7 N/nun2
100111111
3.7 N/mm 2
65ni
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
285
Clllplcl
SI1,1),
100mm Demec
X]
dI
0 0
dl d2
. .............
0.0005
0.0005
0.001 strain
0.0015
0.002
0.0025
0.003
0.70
0.60
0.50
A
0.40
0.30
0.20
0A
0.10
0B
0.00 0.00 0.20 0.40 0.60 applied 0.80 load/design load 1.00 1.20 1.40
0C
0DI
286
Chapter 8
8.6
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
concrete: I'
=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 StrutTie
The results of the elastic analysis and corresponding struttie 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)
287
Chapter 8
At the inner corner of the structure, i. e. node 2, the concrete is in a state of biaxial 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 struttie model, the force in this strut is equal to 107kN, 2<0.6f,,,. A in 14.3N/inm 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
i) (rr=O)
n) (rr=30%)
I( 1k N
()
121
ir
I WkN
I OOkN
I NAN
iv) Dimensioning
289
Chapter 8
100
17
A
I Z150
750
100
d
20)
100
1,4 ,
:
fig. 8.6b(ii) Corner FJ IA Reinforcement & Formwork
290
Chapter 8
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 12.96kNin, 0 08Md). The numerical ultimate moment was recorded its 14.5kNin, (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(ef). The numerical and experimental strains presented in figs. 8.6(gh) show a reasonable correlation. steel
The numerical compressive stresses show that stress concentrations occurred in tile inner corner, which as previously stated is under biaxial 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
(0.3Md)
292
Chapter 8
1.4
1.2
E 0 E 0.8
11 ex
JE
1.2
0
0 0.2 0.4 0.6 strairNeld strain 0.8 1
1.4
12
m1 tu
'70 (D
0.8
2 0.6 'a
CL
m 0.4
0.2
0.2
0.4
0.8
1.2
294
Chapter 8
50nim
14.3Nj caF=
/MM2\\\
I OOkN
750
0D
(56.25,693.7)
0C 600
A( 131.2,568.7)
450
150
300
295
Cilaplel 8
0.9
B
0.8
0.7
0.6
0.5 (0)
A 
10 0
A B
0.2
0.4
0.6
0.8
1.2
1.4
C
0.5
D
0.4
0.3 Z;
v'
0.2

0.1
6c
10D
ou 0
0.2
0.4
0.6 applied
1.2
1.4
296
Chapter 8
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
concrete:
150T 750
L
Ik
concrete:
1 ' =41.3N ,,,
150
steel: f =504N/nini
2 (012)
III 2
A, provided
(MM2)
297
Chapter 8
I OMN
()
imp I WkN
I MkN
I OOkN
i) Struttie Model
ii) Dimensioning
298
Chaptei 8
StILIt&'.TIC DCSll)
2(012)
150 spiral section AA Cover oil stirrups StI HLIPS0 section CC 15111111
06 50mm
150
12
(a' I 100
Detills
299
Chapter 8
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(1m). 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
HI
72
//'
I
I.
301
,, I Ic I ), "Wil
Ei"
c'
Ii
/
Corner FJ2B at ultimatc ( hg,. t7l 8.7(J')
, i "I
,I
"
302
Chaptel
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
E 0.8 0 E c P 0.6 E 0 E
1 =
0.4
IF,
0.2 0 0Aenta numeric al expe rim
on 0
0.1
0.2
0.3
0.4
0.6
M080.9
0 0.8
r 0.6
0.4
0.2
numerical
0expenmenlal
F' ..........
0,9
0.1
0.2
0.3
0.4 strain/yield
0.5 strain
0.6
0.7
0.8
303
Chapter S
'Strul
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
00
0.2
0.4
0.6
0.8
1.2
0.2
0.15
0.1
0.05
D
01 0
............
1.2
0.2
0.4
0.6
0.8
304
Chapter 8
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
II 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
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 struttie models, assuming that steel yielding governs failure, finite elements and experimental ultimate loads is presented in table 8.8(b). In each case, the struttie model provided a good prediction of the ultimate load behaviour of the structure, comparable with that of the finite element model. Model (struttie) _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
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
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 struttie 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 nonlinear 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 2way 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
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
*
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 nonlinear 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
corbels.
308
References
AbdelHafez, L. M., 'Direct Design of Reinforced Concrete Skew Slabs', Ph.D. Thesis, University of Glasgow, Civil Engineering Dept., (1986) 2. AbdelKader, M., 'Prediction of ShearStrength of Reinforced and Prestressed Concrete Beams by Finite Element Analysis', Ph.D. Thesis, University of Glasgow, Civil Engineering Dept., (1993)
AbdulWahab, 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. 11512 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.4956, (Mar 1994)
AlMahaidi, R. S.H., 'Nonlinear 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., 'StressStrain Relationship of Normal, High Strength and Lightweight Concrete' Magazine offoncrete Research, Vol 47, No. 170, pp. 3944, (May 1995)
10.
Alshegeir, A., & Remirez, J.A., 'Struttie Approach in PretensionedDeep Beams' ACI Structural Journal, Vol. 89, pp.296304 (1992) Alshegeir, A., & Remirez, J.A., 'Computer graphics in Detailing Struttie models' ASCE Journal offomputing in Civil Engineering, No.6, pp.220232 (1992)
12.
Armer, G. S.T., Discussion of Ref. 129. Concrete, Vol. 2, pp. 319320, (Aug. 1968)
309
13.
Bazant, Z. P., & Cedolin L., 'Fracture Mechanics of Reinforced Concrete', , ASCE Journal of'Meclianical Division, Vol. 106, pp. 12871306, (1980)
14.
Bell, J.C., & Elms, D. G., 'Finite Element Approach to Postelastic Slab Behaviour' ACI Special Publications, SP3015,pp.325344, (Mar 1971)
Bensalem, A., 'Direct Design of RC Structures using Nonelastic 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 1556, (199 1) BS81 10 Part I&2, 'Structural Use of Concrete' British Standards Inst., ( 1985) Buyukozturk, 0., Nilson, A. H., Slate, F.O., 'StressStrain Responseand Fracture of a Concrete Model in Biaxial Loading' A CI Journal, Vol. 68, pp. 590599, (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.3594 10, (Jun 1972) Cervenka, V., Gerstle, K., 'Inelastic Analysis of Reinforced Concrete Panels Part 1: Theory' IABSE Publications, Vol. 3111,(197 1) Cervenka, V., Gerstle, K., 'Inelastic Analysis ofReinforced Concrete Panels Part 2: Experimental Verification and Application' IABSE Publications, Vol. 3111,(197 1) Chen, W. F., 'Plasticity in Reinforced Concrete', McGraw Hill, New York, ISBN 0070106878, (1982)
Clark, L. A., 'The Provision of Tension and Compression Reinforcement to Resist Inplane 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. 624644, (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., 'Nonlinear Finite Element Analysis of Concrete Slab Structures' Proceedings oj'ICE, Vol 63, Part 2, pp. 159179, (Mar 1977)
28.
29.
30.
Crisfield, M. A., 'NonLinear Finite Element Analysis of Solids and Structures, vol. 1: Essentials' John Wiley & Sons, (199 1) Desayl, P., & Krishnan, S., 'Equation for the StressStrainCurve of Concrete' Journal of the American Concrete Institute, (Mar 1964) Eibl, J., Akkermann, J., Idda, K., LuceroCirnas 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) EIMezaini, 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. 740754 (Apr 1989) Foster, ST, & Gilbert, R. I., 'Experimental Studies on HighStrength 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. 310322, (1985)
311
38.
Gupta, A. K., Akbar, H., 'Cracking in Reinforced Concrete Analysis' ASCE Journal qf Structural Eng., Vol. I 10, No. 8, pp. 17351747, (Aug. 1984)
39.
Hagberg, T., 'Design of Concrete Brackets: On the Application of the Truss Analogy' ACI Journal, Vol. 80, pp. 312, (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 Shearfriction Theory for Bracket Design' ACI Journal, Vol. 7 1, No.2, pp.312, (Feb 1974) Hillerborg, A., 'Reinforcement of Slabs and Shells Designed According to the Theory of Elasticity' Betong, 38(2), pp. 101109, (1953) Hong, S.G., 'Truss Model for Tension Bars in Reinforced Concrete Bearns: TensionTensionCornpressionRegions' 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, 4347, (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. 177189, (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., 'Elastoplastic 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. 93105, (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. 172176, (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.819 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 0216926955, (1990) Kotsovos, M. D., & Pavlovic, M. N., 'Structural Concrete Finite Element Analysis for Limit State Design' Thomas Telford Publications, London, ISBN 0727720279, (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. 1661, (Feb 1965) Kupfer, H., Hilsdorf, H. K. & Rusch, H., 'Behaviour of Concrete Under BiaxIal Stresses' ACI Journal, Vol. 66, pp. 656667, (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., 'StressStrain Responseand Fracture Compression' in Uniaxial Biaxial Concrete and of ACI Journal, Vol. 69, pp. 291295, (May. 1972)
MacGregor, J.G., 'Reinforced Concrete Mechanics and Design' PrenticeHall, 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.205230, (1998)
Mansur, A., Wee, T. H., Chin, M. S., 'Derivation of the Complete StressStrain Curves for Concrete in Compression' Magazine of Concrete Research, Vol 47, No. 173, pp. 285290, (May 1994)
67.
68.
Marti, P., 'Basic Tools of RC Beam Design' Journal oj'the American Concrete Inst. 82, pp.4656, (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.5277, (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. 366372, (May 1971) Mayfield, B., Kong, F, & Bennison, A., 'Strength and Stiffness of Lightweight Concrete Corners' A CI Journal, Vol. 69, pp. 420427, (Jul. 1972)
Meyer, C. & Bathe, K. J., 'NonI i near Analysis of R. C Structures in Practice', Proceedings ol'ASCE, Journal ofthe Structural Divisimi, Vol. 108, No. 7, pp. 16051622, (Jul 1982)
69.
70.
71.
72.
73.
74.
75.
314
76.
McNeice, A. M., 'ElasticPlastic 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.315, (1985) Morsch, E., 'Concretesteel Construction',
English Translation by E. P. Goodrich, MacGrawHill, New York, (1909)
78.
79.
Muller, P., 'Failure Mechanisms for RC Beams in Torsion & Bending' International Association fiv Britqe & Structural Engineering Publications 3611,147163 (1976) Ngo, D. & Scordelis, A. C., 'Finite Element Analysis of R. C. Beams', ACI Journal, Vol. 64, No. 3, pp. 152163, (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. 10301038,(1964) Nielsen, M. P., Braestrup, M. W., Jensen,B.C., Bach. F., 'Concrete Plasticity, Bearn ShearShearin JointsPunching 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. 393396, (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. 3135, (Jul. 1977) Park, R., & Paulay, T., 'Reinforced Concrete Structures' John Wiley and Sons Inc., London, ISBN 0471659177, (1975)
Phillips, D. V., & Zienkiewicz, O. C., 'Finite Element Nonlinear Analysis of Concrete Structures' Proc. qj'the Inst. of'Civil Engineers, Part 2, No. 6 1, pp. 5988, (Mar. 1976)
87.
88.
315
89.
Phillips, D. V, & Wu, Z. P., 'An Orientated Embedded Bar Formulation with Bondslip' Numerical Methods in Engineering; Theory and Application, Ed. Pande and Middleton, J., Vol. 1, pp. 320328, (1990) Popovics, S., 'A Review of StressStrain Relationships for Concrete' ACI Journal, (Mar 1970) Prakhya, G. K. V., & Morely, C. T., 'TensionStiffening and MomentCurvature 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 TrussModel 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.925930, (199 1) Rashid, Y. R., 'Ultimate Strength Analysis of Reinforced Concrete Pressure Vessels' Nuclear Engineering and Design, Vol. 7, No.4, pp.334344, (Apr 1968) Reinhardt, H.W., 'Crack Softening Zone in Plain Concrete Under Static Loading', Cement & Concrete Research,Vol. 15, pp.4252, (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., 'Nonlinear Finite Element Analysis of Reinforced Concrete Corbei' Computers & Structures, Vol. 46, No. 2, pp 343354, (1993) Ritter, W., 'Die Bauwelse Hennebique', (Hennebique's Construction Method), Schweizerische Bauzeitung, Zurich, v. 17, pp. 4143,4952,5961, (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. 19111924,(1974) 104. Schlaich, J., Schaefer, K., & Jennewein,M., 'Towards a Consistent Design of Structural Concrete' Journal qj*the PrestressedConcrete Inst. 32, pp.74150, (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'
Stroband, J. & Kolpa, J.J., 'The Behaviour of Reinforced Concrete ColumnBeam Joints, Part 1: Corner Joints Subjected to Negative Moments' ResearchReport 5839, 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, 21092122, pp . (Oct. 1973)
317
1] 3.
Sundermann, W., & Schaefer, K., 'Tragfahigkeit von Druckstreben und Knoten in DBereichen' 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. 114, (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 Structutal Journal, Vol. 95, No. 3, (May/Jun 1998)
119. Tasuji, M. E., Nilson, A. H., & Slate, F. O., 'Biaxial StessStrain 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. 8594, (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., 'Strainsoftening of Concrete Under Multiaxial Loading' Dissertation, Eindhoven University of Technology, Holland, (1984)
124. Van Mier, J.G. M., 'Examples of Nonlinear 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. 8203, 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., 'Elastoplastic Finite Element Analysis of Concrete Slab Structures' Pioceedings of ICE, Technical Note TN99, Vol. 57, pp. 535543, (Sep 1974)
129. Wood, R.H., 'The Reinforcement of Slabs in Accordance with a Predetermined Field of Moments' Concrete, Vol. 2, pp.6975, (Feb. 1968) 130. Xie, Y. M., & Steven, G.P., 'A Simple Evolutionary Procedure for Structural Optimization' Computers & Structures, Vol 49, No. 5, pp.885896, (1993) 131. Xie, Y. M., & Steven, G.P., 'Optimal Design Of Multiple Load Case Structures using an Evolutionary Procedure' Engineering Computations, Vol II, pp.295302, (1994) 132. Yuri, Y. M., Alshegeir, A., & Remirez, J.A., 'StrutTie Model Design of Disturbed Regions in Concrete Structures' Proceedings, ASCE Structural Congress XII, Vol. 1, pp.233238, (1994) 133. Yuri, Y. M., & Ramirez, J.A., 'Strength of Struts and Nodes in StrutTie Model' ASCE Journal of'Structural Engineering, (Jan. 1996) 134. Yuri, Y. M., 'Nonlinear struttie 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 Struttie model Approach and its Application Tool' ICE Proc., Structures & Buildings Journal, No. 140, pp. 1324, (Feb. 2000) 136. Zienkiewicz, O.C, 'The Finite Element Method' McGraw Hill Book Company, 3rd Edition, (1977)
UN I Ih I,, 
319
Mult mai mult decât documente.
Descoperiți tot ce are Scribd de oferit, inclusiv cărți și cărți audio de la editori majori.
Anulați oricând.