Documente Academic
Documente Profesional
Documente Cultură
SUBJECTED TOFLEXURE
by
BSc, MSc
Engineering
of Surrey
A Thesis Submitted
for
the Degree of
SUMMARY
of creep in reinforced
stresses. Consequently, there is a need to consider the stress redistribution and relieving effect in design in order to make a proper capacity of the structural member in
and analytical
and extended to undertake the task of evaluating in a full-scale structure. The structure under
abutments.
Field measurements are obtained from the instrumented abutments which are subjected to sustained lateral embankment backfill. Laboratory
columns of the
pressures frorn'the
tests on two beams and small companion of the basic time some recommendations, in the
specimens provide the data necessary for the, computation dependent parameters of the concrete used on site. Finally guidelines and design charts
design of flexural
ACKNOWLEDGEMENTS
I wish to express
my sincere
gratitude
to my supervisor,
Dr. Peter
offered
My thanks are also due to Mr. S. Robinson, Mr. M. Jackman and Mr. P. Haynes, who gave technical I am also grateful advice and assistance in the experimental of Commonwealth work.
to the Association
Universities,
CONTENTS
Page No. Summary Acknowledgements List List of of figures tables 11 M viii xii
List
of plates
x1i xiii
2 3 8 8 9 9 10 11 12 13 14 15 18 19 20 21 22 22 24 24 26 28 30 31 31
2.1 Introduction 2.2 Microstructure of concrete 2.2.1 Hydrated cement 2.2.2 Pores 2.2.3 Water 2.3 Time dependent phenomena 2.3.1 Shrinkage 2.3.2 Creep 2.3.3 Influencing factors 2.3-. 4 Effects on structural
behaviour
2.4 Practical prediction methods 2.4.1 CEB-FIP 1970 2.4.2 CEB-FIP 1978 2.4.3 ACI Committee 209 2.4.4 Bazant-Panula 2.4.5 Comparison of various methods 2.4.5.1 Time function 2.4.5.2 Concrete constituents 2.4.5.3 Size and shape effects 2.4.5.4 Environmental conditions 2.4.5.5 Age at loading 2.5 Summary
iv
laws for
creep
33 33
of
law
34 34
3.3 Age-adjusted modulus 3.3.1 Fundamentals of law 3.3.1.1 Basic theory 3.3.1.2 Verification of theory 3.3.2 Computation of aging coefficient 3.4 Viscoelastic aging 3.4.1 Rheological models 3.4.1.1 Basic models 3.4.1.2 Models for concrete 3.4.2 Formulation of law
3.5 Summary
35 36 36 37 39 40 41 41 45 46
49
4 Analysis
of creep effects
51 51 52 52 53 54 56 58 61 61 62 63 64 65 65 70 71
74
4.1 Introduction 4.2 Simplified method 4.2.1 Basic assumptions 4.2.2 Stress-strain relationship 4.2.3 Derivation of time dependent equations 4.2.3.1 Basic equations 4.2.3.2 Quadratic form 4.2.4 Extension for aging 4.3 Classical method 4.3.1 Basic assumptions 4.3.2 Stress-strain relationship 4.3.3 Derivation of governing equations 4.3.3.1 Equilibrium equations 4.3.3.2 Volterra-type integral equations 4.3.3.3 Single non-linear integral equation 4.3.4 Numerical analysis
4.4 Summary
5 Theoretical
studies
75 75 76 77 80 87 91 92
5.1 Introduction 5.2 Comparative studies 5.2.1 Neutral axis position 5.2.2 Stress redistribution 5.2.3 Strain profile 5.2.4 Stress profile 5.2.5 Variable moment
vi
5.3
Parameter studies 5.3.1 Creep configuration 5.3.2 Creep path 5.3.3 Age at loading 5.3.4 Compressive reinforcement 5.3.5 Reinforcement ratio
5.4 Summary 6 Case study 1 6.1 Introduction 6.2 Experimental work 6.2.1 Companion specimens 6.2.2 Strain measuring devices 6.2.2.1 Acoustic vibrating 6.2.2.2 Demec gauge 6.2.3 Laboratory tests 6.2.3.1 Static modulus 6.2.3.2 Strength 6.2.3.3 Thermal expansion 6.2.3.4 Shrinkage 6.2.3.5 Creep
117 119 119 120 120 121 122 123 123 125 125 126 127 127 129 129 130 133 136 136 139 140 143 145 145 148 148 151 153 153
154 154 155 158
wire
gauge
6.3 Results and discussions 6.3.1 Modulus of elasticity 6.3.2 Strength 6.3.3 Coefficient of thermal expansion 6.3.4 Free shrinkage 6.3.4.1 Normal drying 6.3.4.2 Varying temperature 6.3.4.3 Varying humidity 6.3.4.4 Retarded drying 6.3.5 Creep function 6.3.5.1 Normal drying 6.3.5.2 Retarded drying 6.3.5.3 Varying humidity 6.4 Summary 7 Case study 11 7.1 Introduction
7.2 Full-scale work 7.2.1 Site details 7.2.2 Details of abutments Instrumentation 7.2.2
7.3 Experimental work 7.3.1 Details of beams 7.3.2 Instrumentation 7.3.3 Shrinkage measurement 7.3.4 Loading arrangement 7.3.5 Creep test
vii
7.4
and discussion Free shrinkage Creep function Restrained shrinkage Creep effects
7.5
Summary
207 215
and worked
LIST OF FIGURES
Page No.
Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure
Figure
2.1 2.2 2.3 3.1 3.2 3.3 3.4 4.1 4.2 4.3 4.4 4.5 5.1 5.2 5.3 5.4 5.5 5.6
5.7
11 12 16 43 models 44 45
components of time dependent deformations models (a) Kelvin model (b) Maxwell
The deformation response of basic rheological (a) Kelvin model (b) Maxwell model Burgers model The deformational response of a Burgers model concrete system fibre section
46 in bending 55 62
Creep of cracked reinforced Beam geometry and coordinate Mechanical material Strain model of a typical through
distribution region
beam depth
Integration Effects
of creep on reinforced
Neutral axis histories for doubly reinforced different analytical methods Tensile steel stress histories for singly using different analytical methods
Tensile steel stress histories for doubly reinforced using different analytical methods Compressive steel stress histories section using different analytical
Maximum compressive reinforced section Maximum compressive section reinforced concrete stress using different concrete stress using different viii
Figure
5.8
86
ix
Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure
5.9 5.10 5.11 5.12 5.13 5.14 5.15 5.16 5.17 5.18 5.19 5.20 5.21 5.22 5.23 5.24 5.25
reinforced
section section
using using
88 89 90 93 94 97 98
Combined effect of shrinkage and creep on (a) strain distribution and (b) concrete stress distribution for Stress profiles methods analytical for Stress profiles methods analytical Variable Neutral singly reinforced section section using various using various
doubly reinforced
stress
histories
direct
and bending
Influence of age at loading for loading at 7 days Relation between creep after stress-strength ratio Effect of compressive position Effect stress of compressive
5.26
110 on maximum compressive ill on neutral axis in singly 113 113 in 114
Effect of reinforcement ratio on maximum compressive concrete stress in singly reinforced section Effect singly ratio of reinforcement section reinforced on tensile steel stress
Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure
5.31 5.32 5.33 5.34 6.1 6.2 6.3 6.4 6.5 6.6
ratio
on neutral on tensile
axis steel
in doubly 115 stress steel in 115 in 116 116 131 131 of 135 137
Effect of reinforcement ratio doubly reinforced section Effect of reinforcement ratio doubly reinforced section
on compressive
stress
Effect of reinforcement ratio on maximum compressive section concrete stress in doubly reinforced Development of modulus of elasticity Development of modulus of elasticity Determination concrete of coefficient with with time time
expansion
Comparison between A-side methods Comparison between E-side methods Comparison between A-side method Comparison between E-side method Comparison between A-side methods Comparison between E-side methods Comparison between A-side method Comparison between E-side method Elevation Elevation of Bridge of A-side
data and prediction data and prediction 138 data and Bazant-Panula 141 data and Bazant-Panula 142
Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure
6.7 6.8 6.9 6.10 6.11 7.1 7.2 7.3 7.4 7.5 7.6 7.7
shrinkage
creep data and prediction 146 creep data and prediction shrinkage shrinkage data and Bazant-Panula data and Bazant-Panula 150 Interchange 156 156 157 160 strains for load cases considered shrinkage 165 168 147 149
creep functions
strains 170
xi
Figure
Figure Figure Figure
7.8
7.9 7.10 7.11
shrinkage
strains
171 172 173 caused 175 beams 178 179 180 beams182 Column 185 Column 186 beam 187 188 189 218
Concrete tensile stresses by restrained shrinkage Applied Total Total Total moment histories time time time dependent dependent dependent of
adjacent during in in
to reinforcement test test test on test beam beam from plot test for
deflections on strain-time
Figure
7.17
of bending
on strain-time
plot
for
Figure
7.18
Bending moment histories backfilling Total Total Neutral time time axis dependent dependent depth
for
site
columns
after
capping
7.19* 7.20 B. 1 B. 2 B. 3 B. 4
stress
ratios
Compressive
stress
LIST OF TABLES
Page No. Table 3.1 Comparison . shrinkage Neutral different of influence of aggregate type on creep and 28 stress after 240 days under 10C on creep ill of predicted long-term deflections of beam 118 mix proportions cube strengths strengths time dependent 181 121 132 133
Table
5.1
Table
5.2
reinforcement
Table
5.3
tensile
LIST OF PLATES Page No. Plate Plate Plate Plate Plate 1 2 3 4 5 Bridge Acoustic 15 on the M25 interchange vibrating wire strain at Wisley gauges 4 124 124 128 162
Demountable demec gauge Creep rigs Loading rig for for small test specimens 1 beams
xii
NOTATION
area steel
of
section
steel
tension steel
steel
of
concrete
Es Ell e F f, g, h ic K, K2 K3 K4 k kT L
of
steel
function
appearing
in
viscoelastic
analysis
temperature
xiii
xiv
m n P p pf R 1
rb
weighted
integral
for
beams
S* T To t to t, t*
Uh
creep function
temperature
load application
time at load application time at which beam fibre compression activation quantities humidity at depth y first goes into
V, W w x9y)z A, B, C, D, E b, a, c a, 6
6b 6P 6S CC
coordinates quantities appearing defined respectively quantities basic appearing in methods of creep analysis in effective and
deformation
parameters
of beam
xv
cc
CS Esh El S C0EI E0 it E C V
ac
US
of S T 0
concrete stress
tension steel stress
compression integration steel stress representing age at loading
variable
creep coefficient ratio of concrete stress at time t stress at initial loading to concrete
X w
CHAPTER I
INTRODUCTION
1.1 Background
materials
under sustained loading. This type of behaviour is termed creep and affects the strength and long-term performance of structures. In addition concrete
also exhibits changes in strain with time when no external stress is acting, due to movement of water to or from the ambient medium. This change Is referred Hatt 1 to simply as shrinkage or swelling, respectively. was probably the first to acknowledge the significance of creep concrete
an structural
beams to sustained loading and measured the deflections. However in his findings, no reference was made to shrinkage of concrete, so that the increase in deflection effects of the reinforced beams presumably includes the from the
Introduction
Although shrinkage of concrete was observed much earlier than creep, its structural significance was recognised by Whlte2 only in 1911 when he
mentioned the stresses developing due to shrinkage in a paper to the American Society for Testing Materials. Since then, the problem of shrinkage and creep of concrete has remained of interest theoretical point of view and for structural both from the
design purposes.
Ever since time dependent phenomena in concrete became recognised, there has been an overwhelming number of publications on the subject. Various aspects of creep and shrinkage, such as hypotheses on the mechanism of the phenomena involved along with observed properties of the material and problems related to structural structural behaviour, have been studied. In of these physical
properties of concrete is recognised as being a possible cause of cracking. They also influence the deformations and are responsible for the transference the structure of load from concrete to steel in compression members, while is in service. of observed data on actual structures has conditions of one or
more of the variables, under which the actual structures exist. However, real structures are very different differences from small scale specimens. Apart from age at loading and rate of
strength development, they also exist under varying environmental conditions of temperature, vary simultaneously. laboratory humidity and stress, some or all of which may the relevance of
Introduction
performance of the
(a) the relevance and importance of small scale tests in the prediction of the shrinkage and creep properties of the concrete
in the structure, (b) the effect of painting the concrete with a waterproofing agent on
the shrinkage and creep behaviour, (c) the validity of using the mean axial creep properties to represent members subjected to
(d) the magnitude and effects of the phenomena of shrinkage and creep an the long-term behaviour of the structure.
data obtained while the bridge is under service. The results presented here form part of a continuing study into the time dependent deformations occurring in the columns of the abutments.
In reinforced
material,
paste which ages, shrinks and creeps. The steel reinforcement virtually
stable with time and restrains the concrete from attaining Thus, prior to conducting a structural
full deformation.
analysis which
-1
I
I,
introduction
observed data on an actual structure, of concrete such as creep coefficient, strain have to be determined. Several international
the basic time dependent parameters creep function and free shrinkage
Codes of Practice
basic parameters from variables which influence time dependent behaviour of the concrete. Factors such as the concrete constituents, environmental
conditions, size and shape of the members and age at loading are considered. These procedures along with the method of Bazant and Panula 6278 are assessed in terms of their prediction of the time dependent
parameters of the concrete used in this study. The present state of understanding of the time dependent behaviour of concrete together with a review of current prediction methods is discussed in the next chapter. An additional problem was encountered due to the fact that the columns were waterproofed with bituminous paint and that the prediction methods do However, the method of Bazant and Panula (BP) of painting
could be conveniently modified to take into account the effect and results from controlled
tests on painted companion specimens were used The evaluation of the physical
properties of the concrete together with the shrinkage and creep characteristics, Chapter 6. Once the basic parameters are evaluated, the structural analysis can under the additional influence of painting is given in
proceed and although creep and shrinkage are interdependent, the prevailing assumption is that they can be treated separately in order to simplify the
analysis. A greater part of the research programme has been devoted to developing the analysis for time dependent effects due to creep in members
under flexure. In the analysis, creep behaviour is modelled by a stress-strain relationship and this relationship is used to represent the
Introductlon
response of the concrete under load. The simple and most widespread stress-strain relationship which incorporates a fictitious effective for creep
modulus, is applied to the analysis of a cracked section under aging of concrete can also be taken into into the relationship and
bending. Furthermore,
of the earlier law. Creep of concrete can behaviour and using elementary beam
also be modelled by viscoelastic theory, Sackman and Nickell cracked singly reinforced 9
extended for the general case of a doubly reinforced section subjected to a flexural moment. The theories behind the stress-strain relationships for
creep are given in Chapter 3 and the following chapter gives a detailed account of the formulation stress-strain relationships. A comparative study of the various analytical methods for creep of f inding of the methods of analysis which utillses those
constitutes part of the research programme, with the intention out which method best suited the problem of determining strains in the instrumented reinforcement
parameter study and the results are reproduced in the form of charts which are suitable for design purposes. Results from the comparative parameter studies are presented and discussed in Chapter 5. The experimental instrumentation work carried out in conjunction with the and
creep tests on companion specimens and two test beams, which are scaled down versions of the site columns. Results from these tests were used primarily prediction to verify the theoretical techniques adopted. Finally, a of the
Introduction
moment acting on the columns as a result of sustained lateral pressures from the backfill prediction are made using the proposed analytical method. The
of the creep and shrinkage properties in the larger test beams time dependent
and site columns along with the comparison of theoretical strains with measured values are presented in Chapter 7.
CHAPTER 2
MATERIAL
BEHAVIOUR
2.1 Introdiction
Concrete has at least two fundamental differences in comparison with many other common structural materials. Firstly, unlike the metals, water, is
products of hydration,
unhydrated cement and aggregate is heterogeneous and when reinforcement present, anisotropic as well. The second difference arises from the fact
that the properties of concrete change with time and are greatly affected by temperature and relative humidity, and therefore by the environment.
The shrinkage and creep in concrete are commonly defined by the free shrinkage strain and creep coefficient and they are very much dependent on
the above mentioned variables. In design, the approach with regards to the evaluation of these time dependent parameters is based on simplified procedures with correction variables. This chapter provides the background into time dependent phenomena in concrete and their subsequent ef f ects on structural behaviour. An factors to account for the effect of the
understanding of the mechanisms involved and the factors which influence them is essential so as to enable the formulation procedures in the form of practical prediction of the simplified
methods.
Material
behaviour
2.2 Microstructure
of concrete
in ine f a matrix embedded aggregates and of coarse hydration, cement consists of crystalline
The gel plays the dominant role in time dependent deformation behaviour the the quantitatively. modifies aggregate while materials in cement are believed to be unaffected
The crystalline
by ordinary drying. As
is deemed of concrete
necessary to give a proper pespective of the behaviour of concrete and assist in the derivation of simplified but reliable material laws.
The chemical reaction between water and cement leads to the formation of a hydration product of colloidal size and properties. At any stage of
hydration the hardened paste consists of hydrates of the various compounds, referred to collectively as gel, of crystals of calcium hydroxide, some
minor components, unhydrated cement, and the residue of the water-filled spaces in the fresh paste. These voids are called capillary within the gel itself there exist interstitial pores, but
The resulting mass of cement gel is very porous and due to its high specific surface, a large amount of free water becomes adsorbed on the is to from the If prevented, the of water or paste movement cement surface. reactions of hydration would ultimately use up the water until too little humidity within
is left to saturate the solid surfaces, and the relative the paste decreases. This is known as self-desiccation.
Material behaviour
10
The gross space available f or the products of hydration consists of the absolute volume of the dry cement together with the volume of water added to the mix. The loss of water due to bleeding and the contraction the paste while still plastic accounts for a small reduction in the gross than the twice the The more occupy gel particles space as paste sets. volume of the dry cement but due to their size, the shape of the individual however, by be determined There are, strong observation. cannot particles indications that they are in the form of thin, rolled or crumpled sheets and between each sheet there exists a molecular layer of water. of
2.2.2 Pores
Within the hardened cement paste there exists two distinct pores. Firstly,
types of by
the part of the gross volume which has not been filled
throughout
of the hardened
paste.
of hydration,
and the
of the capillary
capillary
interconnected
interstitial
particles and are much smaller than the capillary about one-third
of the volume of the gel solids in normal drying for a given cement
conditions. However, the actual value is characteristic but is largely independent of the water/cement progress of hydration.
is formed at all stages and that continued hydration does not affect
Material
behaviour
11
products already in existence. Thus as the total volume of gel increases with the progress of hydration the total volume of gel pores also in Figure 2.1.
interstitial Solid dots represent cel particles; spaces are gel cores; spaces such as those marked C Size of gel pores is exaggerated. are capillary cavities.
11
2.2.3 Water
The cement paste is hygroscopic owing to the hydrophilic character cement coupled with the presence of sub-microscopic content of the paste depends on the ambient humidity. capillary pores, owing to their comparatively
of
ambient relative
the gel pores even at very low ambient humidities. Water in hydrated cement is held with varying degrees of firmness. At one extreme there is free water while at the other, chemically combined
water forming a definite part of the hydration products. In between these two categories there is gel water held in a variety of other ways.
Material
behaviour
12
is called
adsorbed water, and that part of it which is held between the surfaces of certain planes in a crystal is called interlayer or zeolitic water. Lattice
water or sometime referred to as water of hydration is that part of the water of crystallization which is not chemically associated with the and is
beyond the range of the surface forces of the solid phase. A diagrammatic representation of the probable structure Figure 2.2. of hydrated cement is shown in
00x
0 00
00
00
00
00 0 0
00 0
00 0 0
00
000000
Tobernorite z sheets
Interlayer water
interparticle J* bonds
000
000 0
70
of hydrated cement
12
of concrete is partly
a result of the
response of the various components of the gel to stimuli pressure in, and the hygral equilibrium creep deformation
of, the cement paste. The bulk of of water and the pressures
to capillary
Material
behaviour
13
and changes in the surf ace energy of the solid gel particles.
2.3.1 Shrinkage
deformation
that concrete
undergoes when not subjected to load or restraint. characteristic of any structural material
The shrinkage
shrinkage found by measuring linear strains on a control specimen under constant ambient humidity and temperature. Shrinkage can be divided into
its various components namely capillary drying shrinkage. Water in the capillary surface tension. The capillary
As the concrete begins to dry out, the homogeneous in is the narrow spaces trapped the and water Subsequently, menisci are formed between pressure, thus causing
volume changes caused by hydration and other chemical reactions are termed as chemical shrinkage. Volume change due to hydration of cement is proportional of hydration and is accelerated by a raise in temperature. both colloidal particles and crystallised to the degree
During hydration
these hydration products are not stable with respect to a decrease of the relative humidity. They tend to lose some of their hydrate water, resulting the calcium hydroxide formed during
in further
Material
behaviour
14
hydration can react with the carbon dioxide in the surrounding air. This reaction known as carbonation is limited accompanied by a decrease in volume. Drying shrinkage is defined to be the volume change of a colloidal inert system as its moisture content is changed. The physical presence of free water within the gel subjects the solid particles to internal pressure is determined internal tension. The by pressure resulting surface created by the thickness of the adsorbed water film on the gel particles, particular relative humidity at a to surface zones only and is
the pores can either cause the adsorbed water to be removed or it may increase internal the or either Correspondingly, will pressure re-enter. decrease thus affecting the magnitude of drying shrinkage.
2.3.2 Creep
If concrete is subjected to sustained loads, it continues to deform further with time and this phenomenon is now commonly referred to as creep. The creep strain is a measure of the difference loaded specimen and an equally old identical in deformation between a
specimen that has undergone the same history of environmental The instantaneous elastic deformation
which is the ratio of the creep strain to elastic strain. Creep of concrete occurring under conditions of no moisture movement to or from the ambient medium is termed as basic creep. This component of creep is unaffected by the size of the concrete member. However, creep with drying known as drying creep, will have the
occurring simultaneously
Material
behaviour
15
of wqter and cement hydrate. The walls of the the motion of adjacent water
molecules which are held together by van der Waals forces. As molecules of held in the layer in permanently, both the not are pores and adsorbed water a disturbance of the energy balance within the microstructure from sustained mechanical loads, causes redistribution resulting
of the water
molecules from areas of higher pressure to areas of low pressure and relative humidity, or to the surrounding atmosphere. These molecules of
the the cement the diffuse of of crystals surfaces solid along water hydrate resulting in a change of mass and thickness of the adsorbed layers, to believed is together. This bringing the widely and gel particles closer be the dominant mechanism of creep of concrete under moderate stress levels. A diagrammatic dependent time the of components various summary of
Creep and shrinkage of concrete are influenced by a large number of factors, generally discussed under the category of composition, environment and maturity. Concrete is usually characterised by the aggregate type and
humidity, factors in the the the size and principal are environment while shape of the concrete element, temperature and stress.
The influence of water on the magnitude of shrinkage is twofold. Concrete with a high water content possesses larger amounts of free water for evaporation and also results in the formation gel, and both contribute of a higher volume of ratio
Material
behaviour
16
(DSHRINKAGE to -EFROM
Wli'-G-LUIJED
G) CREEP cr U) to c) ELASTIC NOMINAL STRAIN TIME IN HYGRAL CREEPOF A LOADEDSPECIMEN WITHTHE AMBIENT MEDIUM EQUILIBRIUM
Map
IEEP
Material
behaviour
17
increases the size of the pores in the paste structure, sustained load, the redistribution
so that under a
readily and causes a higher rate of creep. Aggregates can be assumed to be relatively stable and have a restraining the fraction effect inert and volumetrically
of aggregate in the concrete mix or using aggregates with a causes a decrease in creep as well as shrinkage. of fresh concrete or the
f in influence hardening so ar as creep and shrinkage rates of setting and they change the volume concentrations capillary water. The greater the degree of hydration of the cement at the time of load application, the lower the rate and total amount of creep. One explanation for this is that the expulsion of water from the gel becomes more difficult as the porosity is decreased through hydration. The effects of ambient temperature and humidity on shrinkage and creep of hydrate adsorbed water and
are important as these factors influence the volume proportions of hydrate, adsorbed and capillary diffuses the moisture rate at which water, and rise is
within the concrete and evaporates to the atmosphere. A temperature accompanied by an increase in thermal activity correspondingly increases diffusion controlled of the water and
increase also accelerates hydration which in turn reduces affects the pore water content in the hardened and
film the thickness the water adsorbed governs of paste which cement movement of water molecules. In conditions of high ambient relative
humidity the rate of loss of moisture or water vapour to the surrounding atmosphere is reduced, and slows down the movement of water molecules to the outer surface, thus reducing creep and shrinkage.
Material
behaviour
18
It is a well established fact that the larger the mass of concrete, the lesser the amount of shrinkage and creep. In a large concrete structure, the path taken by the water molecules to reach the outer
surface is greater and with a resulting increase in the frictional resistance to the movement of water from the interior, in relation to a
small concrete specimen. This probably accounts for the small magnitude of shrinkage and creep observed in large masses of concrete.
2.3.4 Effects
on structural
behaviour
Structural
free to the frequent if were of cracking. concrete cause probably most deform, any shrinkage would be of little consequence, but usually it is
restrained, tensile stresses will developed. These stresses are relieved by creep, but cracking will occur once the tensile strength of the concrete is exceeded. Cracking may not only impair the ability carry its designed loads, but it may also affect the appearance of the structure. The durability of any structure to
and damage
by the entry of
water through cracks, which corrodes the steel, leaches out soluble components, and the concrete deteriorates. Creep, in general, tends to relieve the stress in concrete, especially when reinforced. Thus, when a sustained load is applied to a reinforced
concrete member such as a column, creep of concrete causes a gradual reduction in the load on the concrete and a corresponding increase in the load on the steel. In various structural elements such as continuous beams
in the the most highly stressed of stress some relieves and slabs, creep in the the increases adjacent stress portions of concrete, so and portions
Material
behaviour
19
cracking. However, creep may cause excessive deflections and curvatures, and from an aesthetic point of view would be undesirable.
2.4 Practical
prediction
methods
Prediction characteristics
methods provide designers with the means of estimating known using concrete of shrinkage and creep of a given
the
basic based situations, the on simple Basically are methods parameters. such as unlaxial creep or relaxation or mean deformation of a whole
cross-section that may be in a non-uniform into various Codes of Practice their simplicity, practice. There exists fundamental differences some of them, time dependent deformations several coefficients
In the dependent parameters. principal on which are are presented as a sum of the various the level of simplicity varies vastly,
components. Furthermore,
to procedures which
can estimate the magnitude and the development of the deformations with time using high order mathematical behaviour. All these prediction methods are formulated from numerous shrinkage functions to model the material
the inconsistency Thus, the data results. of scatter and amidst creep and the probabilistic in be borne the mind. evaluated should values nature of
Material
behaviour
20
concrete
structures contain equations and charts that may be used for assessing creep and shrinkage. The procedure includes the most important such as mix proportions, humidity, variables
and creep of concrete under normal conditions. The base value of shrinkage is modified by correction into account the influence of environmental constituents to give the ultimate factors to take
development of shrinkage strains with time are evaluated by multiplying ultimate value with a function that varies exponentially with time. An
additional correction
introduced in the evaluation of the creep coefficient. The approach to the formulations was to single out each influencing
factor independently of the rest, and to take the product. as estimates of the creep coefficients further and free shrinkage strains. However, as shown by some of the factors were
smaller magnitude and slower rate of shrinkage in larger sections are attributed to the size effect which is dependent on humidity.
There is no technique for consideration of changes in environmental conditions, such as variations in ambient humidity minor influences like compaction, workability, and temperature. Other
mentioned. Hence, it is obvious that the values derived from this method cannot be more than approximate. in a situation shrinkage and creep are very important, where the effects of
lower bounds should be considered, using magnitudes 15% above and below the estimated figures.
Material
behaviour
21
14
proposed certain revisions to the CEB-FIP 1970 into the CEB-FIP Model Code 1978 and
the German Prestressed Concrete Code DIN 4227 1979. The purpose of its introduction into the Model Code was to provide designers with a more
refined approach to the problem of creep and shrinkage over its predecessor, The CEB-FIP 1970 Code. The method discards the use of the charts which gave the relationship between the base values of shrinkage and creep, and the relative Instead, a table giving the appropriate relative humidity.
normal conditions of exposure for the concrete. There is allowance for the effect of concrete consistency defined by the slump on the base values of
creep and shrinkage. Major modifications are introduced into the evaluation of creep. Creep part termed as creep or
flow and a recoverable part known as delayed elastic. The creep coefficient is evaluated by taking the sum of both terms. The basis of the formulation lies in the rate of flow method15 which has been developed specifically cater for creep under variable stress and also for cases when there is complete unloading of the concrete. The development of the creep or flow of curves for different member to
thicknesses, while the delayed elastic component is defined by a single curve, for both cases of loading and unloading.
Material
behaviour
22
209
by is base multiplying modified creep or shrinkage of value whereby a with correction factors to allow for the influence of different
variables.
rest.
Factors like effects fine consistency, of curing period, concrete Code 1970 the ignored by CEB-FIP are aggregate and air content, which were included. Instead of utilising coefficients, charts to evaluate the correction factors or
the method gives a set of equations. Most of these equations of the correction factors and the variables
hyperbolic functions.
2.4.4 Bazant-Panula
shrinkage
of concrete is through the use of equations which define the development of shrinkage with time, and the dependence of shrinkage on humidity, and temperature. size, age
development of shrinkage and the size dependence of humidity effects treated according to diffusion theory.
Creep is broken down into its various components, namely basic and drying creep. Basic creep is formulated using a double power law with four 7 development the defines the This of relationship parameters material . basic creep component with time and also describes its dependence and the
Material
behaviour
23
dependence of the elastic modulus with time. The models for predicting basic creep and shrinkage are combined and in drying to environment, a cover creep extended The principle behind the formulation at constant temperature8.
the drying creep curve and the basic creep curve are different.
the is drying that the superposition of a roughly curve creep considers basic creep curve and a curve which exhibits the same dependence on size, drying the dependence The creep humidity of as shrinkage. shape and by Neville been had earlier much established also component on shrinkage on the basis of LIHermite's relationship work'8. Thus, on this assumption, the 17
for creep under simultaneous drying is established through the drying term. more complex with a
summation of the basic creep and an additional The actual formulation for drying is slightly
breakdown of the additional drying creep term into two components. These components represent the two fold effect of drying upon creep, namely that of
moisture caused by drying reduces creep19. The dependence of shrinkage on size and humidity effects is assumed to hold true for the drying creep components. A notable feature of the proposed method is the use of a creep function instead of a creep coefficient to evaluate creep. Within the
be to be proportional stress assumed range, creep may working of concrete to the applied stress and obey the principle the Thus, of superposition.
use of the creep function, defined as the sum of the elastic and creep The is in by time valid. equally a constant caused unit stress, strains advantage of using the creep function over the creep coefficient occurs
have influence the an almost equal effect variables, certain which of when be the to such the are considered as component, creep and elastic on influence of a mineralogical type of aggregate. Since the creep coefficient
Material
behaviour
24
is the ratio of creep strain to elastic strain, the effect would tend to cancel out. By using the creep function could be accounted f or properly.
A more recent improvement to this procedure includes the use of a double power logarithmic law provides a relatively 20 law to describe basic creep The double power . good description of the existing test data on
basic creep. However, there exists some deviations in this law, particularly in representing the final slope of the curves of strain versus of load duration where it appears to be too steep when long
the logarithm
term tests are considered. This is remedied by a creep law which exhibits a transition from the double power law to a straight line in the logarithmic line has
the same slope for all ages of loading and the transition older concrete.
The farAors which influence the properties of shrinkage and creep are expressed in the formulation of the prediction methods in the form of
mathematic, ii relationships or functions. A comparison of the various d on the representation methods ba,.,.. of the influencing factors is important
An ,-,pect of particular
of short time
developm, "'t Of the shrinkage and creep strains with time in the form of logarithmic, exponentL11, hyperbolic and power expressions.
Material
behaviour
25
22
have
been widely used for modelling shrinkage and creep versus time. This form of equation represented concrete shrinkage quite well but its tendency to flatten out underestimates long term creep. Exponential equations for creep
24 23 25 and later by Ulitskii were introduced by Thomas , McHenry who also proposed the same equation f or shrinkage. Shrinkage has also been 26 represented exponentially by investigators such as Lyse , Wallo and 28 27 Kesler . The development of a simple power equation by Shank and a 29 logarithmic expression by the U. S. Bureau of Reclamation provided relationships which do not tend toward an ultimate been used to define creep. The CEB-FIP 1970 method as proposed in its original form, defined the development of shrinkage and creep by a single relationship expressed mathematically which could be was of value and these have
derived for a range of thicknesses from 200mm to 50mm. Later, curves for different
a family
size also entered into the picture. In an attempt method, a correction to rectify factor some of the shortcomings in the CEB-FIP 1970 was introduced into the Model Code to account on member size in the calculation for the
component of creep are defined by two separate charts. Each curve from the family of curves on the charts describes the development for a particular member size which has already been corrected for the influence of humidity. The recoverable part of creep is defined by a single curve, applicable for all member sizes.
Material
behaviour
26
function to describe creep and shrinkage. Equations rather than charts, which relate to the development with time are given. The one significant difference between the ACI 209 and the CEB-FIP 1970 lies in the fact that and shrinkage-time curve for all concretes member thicknesses
and the same curve applies to both shrinkage and creep. The simple models presented in the Codes of Practice considered that
the shape of both the basic and drying creep curves with time to be the same. However the approach taken by Bazant, although less simple, puts the problem in its right perspective and makes it physically more correct. mentioned earlier, the drying creep curve consists of the basic creep component and an additional drying creep term and the shapes of the two component curves are not identical. Test data for basic creep have As
indicated that the curves normally rise at an increasing slope in log time as far as the measurements go. Thus based on experimental evidence, Bazant
proposed a double power law to describe basic creep. The shape of the drying component should resemble that of shrinkage. A hyperbolic function with an exponent, as opposed to the simple form adopted by the ACI 209 method, is used to describe the development of shrinkage with time. Bazant found that this relationship experimental data examined. gave the best fit curves for all the
The composition of concrete may be described in terms of the aggregate, cement and water content. In concrete it is the cement paste which creeps and shrinks, while the aggregate is usually inert and modifies
the creep and shrinkage. The overall effect is that the rate and ultimate
Material
behaviour
27
mix proportions.
The largest single factor which influences the magnitude of shrinkage and creep is probably the water content. Together with hydration, it determines the porosity of the concrete. The greater the porosity, the greater the shrinkage and creep of the concrete. A higher cement content would mean more cement paste and results in higher shrinkage and creep. The CEB-FIP 1970 and ACI 209 methods account for the influence mix proportions by multiplying creep with correction the base value of either the shrinkage or of the
the CEB-FIP 1978 method make no reference to the mix proportions and are rather ambiguous. The proposed procedure of Bazant, evaluates the base value from the mix proportions of the concrete through a series of empirical equations. This presents a more versatile approach in comparison
to the other methods which merely corrects a reference value, based on some mean value of past experimental investigations. of mineralogical
A factor which tends to be overlooked is the effect type of aggregate which has a significant
elastic properties of the aggregate. A comparison of the influence of the aggregate type is illustrated in Table 3.1 whereby the shrinkage and creep
of limestone concrete is taken as unity, and all the other values are expressed as a ratio of the shrinkage and creep of this concrete. The modification dilution effect caused by the aggregate can be accounted for, by the
in which the non-creeping and non-shrinking aggregate there is restraint from the aggregates
accordingly. Methods such as the CEB-FIP 1970 and ACI 209 account for the influence of aggregate through the dilution effect correction by including a of the
Material
behaviour
28
Aggregate type
Creep
Shrinkage
30
effect
ignored. As an exception to this, the method of Bazant evaluates the base value not only in terms of the mix proportion but also includes the dependence on strength. The strength of the concrete is dependent on the elastic properties of the aggregate and indirectly aggregate type into the computation. enters the influence of
The size and shape of a concrete member influence the rate of loss of moisture through drying, or rate of gain of moisture as in swelling, under given storage conditions. These affect creep but also their magnitude. The CEB-FIP 1970 Code caters for these effects correction factor by applying a of curves values of not only the rate of shrinkage and
Material
behaviour
29
effective correction
thickness. In comparison, the ACI 209 method only allows a on the base value, while the relationships defining the rate of
shrinkage and creep are assumed to be applicable for all shapes and sizes. The correction factors from both these methods are based on experimental member sizes and shapes.
the drying process to diffusion and since caused by drying, the size
dependence of these two phenomena could be explained by the diffusion theory. The usually assumed mathematical law for water diffusion in drying equation,
aw at -
Cd
V2
(2.1)
where Cd is the diffusion coefficient w is the humidity of the concrete t is the time
The proposed size dependence for both shrinkage and drying creep is 31 based on the above relationship and is expressed by a parameter called the shrinkage square half-time. The overall effect of this parameter is to in
shift the plot of shrinkage and drying creep components horizontally log-time, with a change in size. This seems to be In agreement with data conducted for size and shape effects on shrinkage
experimental
32
In contrast, prediction methods such as those of the ACI and CEB-FIP Codes, which account for size and shape effects by the multiplication correction factors simply shift the plot vertically. This would tend of
values for large concrete members, in comparison to the values from the BP method.
Material
behaviour
30
2.4.5.4 Envirorynental
conditions
Both shrinkage and creep are influenced by the drying conditions of the environment. The principal and temperature. AS humidity factors in the environment are the humidity decreases, shrinkage and drying creep values methods.
of temperature
The ACI 209 gives no reference to the influence of temperature, the CEB-FIP 1970 method only accounts for the effect of temperature
drying creep during curing. There is no factor to account for the variation of temperature during storage or loading, or for the increase in creep rate above that of normal room
dependence of creep and shrinkage can be explained by the rate of hydration of the concrete and it the drying period into an drying time. The use of the approach for considering
can be taken into account by converting equivalent hydration period or effective concept of effective
variation in temperatures
The CEB-FIP 1978 method includes a correction on the drying period. Temperature differing
33 the dependence of hydration slightly refined relationship which modelled , 34,35 on temperature through the rate process theory
Material
behaviour
31
maturity
of the concrete,
A possible explanation
for this is that the movement of moisture from the gel becomes more difficult as the porosity is decreased through aging or hydration. is accounted for by all the prediction methods discussed, of correction factors which are based on experimental
This effect
by the introduction
evidence. These are evaluated either from charts as in the case of both the CEB-FIP methods or by an equation as given by the ACI 209 Code. A notable feature of the double power law for basic creep in the BP method is the age-shift property inherent in the relationship. same shape of creep curve in log-time but the curve is shifted horizontally. reflected loading The law simply defines the ages of loading, is again ages at
2.5 Summary
state of knowledge in the area of time dependent deformation concrete. it should be realised that neither the microstructural nor macroscopic observation are sufficient
processes involved in creep and shrinkage. Some physically meaningful mechanisms have been described but the actual behaviour depends as well on phenomena such as crack formation and internally created states of stress,
which modify time dependent deformation. The various prediction methods so relevant to national Codes of
Practice, have shown and emphasized that research on creep and shrinkage
Material
behaviour
32
CHAPTER 3
3.1 introductlon
The creep behaviour of concrete is represented by a time dependent stress-strain relationship and this relationship forms the basis of the in reinforced concrete structures. Consequently,
law which have been developed. The stress-strain relationships examined in this chapter includes the modulus, the
simple and most widely adopted law in the form of an effective age-adjusted effective relationship. modulus and a general viscoelastic
stress-strain
analysis of a cracked reinforced concrete section, as shown in the next chapter, in order to investigate the effects of creep. All these laws are linear, whereupon creep is assumed to be proportional to the stress applied, and as a result the principle of
superposition holds. This is known to be acceptable for stresses that do not exceed 0.4 of the strength of the concrete 38,39,40, provided that large in
reversals of strain and repeated loads are excluded, and the variation temperature and water content Is not too severe.
33
34
3.2 Effective
modulus
This law also known as the reduced or sustained modulus, is the simplest form of stress-strain relationship for creep. It has been widely For a large number of
cases, such as creep under constant stress and when the effect
of aging is
negligible, as in old concrete, the law is usually of acceptable accuracy, 9 41 and Sackman according to Pauw
Under constant sustained stress, it is assumed that at any time t, the concrete may be considered as an elastic material elasticity, given by
a Eeff a
(1
having a modulus of
(to)
V)E: (%)
+ 0(t,
(3.1) +o Eef f=
(to ), (to ), )= E(to 0 E
where
effective
stress,
modulus
and modulus at time t
strain
)= qCC(t'tO
(tItO )= 36
creep coefficient,
is simply a
35
method of analysis which in general, is considerably easier to perf orm than one based on stress- strain relationship. 42 However, past investigators such as Ross pointed out, that the any other unsound f or varying stress because it disregards linear
approach is theoretically
the stress history. The strain at any time is assumed to depend on the stress at that time only. Furthermore, the effective modulus of elasticity
applied at age t=to, and with the assumption that creep at any time is proportional to the stress at that time. This would tend to overestimate
creep deformation under increasing stress, and conversely underestimates the deformation under decreasing stress. The effect of age at loading and the results. The
the elastic modulus at the time of loading also affect method also has the disadvantage of predicting upon unloading.
modulus, is its inconsistency in modelling creep when aging of In an attempt 43 to overcome this problem, but
proposed a new law. This new X, In order to obtain an X, is called the aging for the
modulus primarily
of aging. 13 In earlier studies by Trost and Neville the term , X, was referred
This technique was later extended and the 44 developed by Bazant who abandoned the term, relaxation
36
relaxation
problems.
Within the working stress range and provided that strain reversals are excluded, creep of concrete may be assumed to be governed by the linear 44 principle of superposition in time. Bazant produced a general formulation based on the following theory.
When the strains vary linearly with the creep coefficient, can be expressed in the f orm,
its history
cW-
Elt)
C. +
E:10 (t, t
for 0)
t>t
(3.2)
and also
cy
(t)
=0
for
0<t<t
(3.3)
The solution for the stress history can be obtained simply from the solution for q=0, c, = 1, that is, from the relaxation function which is
stress caused by a unit strain. if the conditions described by Equations 3.2 and 3.3 are satisfied, then the following holds true. The stress, a (t) varies linearly with
37
), the relaxation function, and the stress-strain relationship 0 be written in the form of an incremental elastic law.
(t't ER
may
A a(t)
(3.4)
in which
A E: (t)
=C
(t)
-C
(t
),
(t) ACY
=a
(t)
-G
(t
(3.5)
a(t 00
E(t ) 0 2, (t, %
(t, t
0M-e
(t
(3.6) 0 (3.7)
o (t, to)
where
modulus
3.3.1.2 Verification
of theory
The unlaxial creep law may be expressed in either of the following equivalent forms,
two
(3.9)
(t)
(3.10)
38
unit stress applied at time t', and the rest of the terms are as def ined earlier. The relation between functions ic and ER may be obtained, by considering the strain history to be a unit step function, that is, c=1 fortato and c=Ofort<tc). the stress response is
Thus by definition
ER(t,to)
and substituting
c" = 0, yields
1=fJ
0) aER(tl t td
(3.12)
0) +
JC(t, t')
dt'
JC(t, t')
(3.13)
[E(to)
)/0 t) ER(t, -
(tto)
(3.14)
Substituting
this relationship
Equation 3.4, and noting that cr (to)/ E(to) = co then Equetion 3.4 becones ,
W= (to) Mto) cy cy +
td3 ER(t, -
ICI -Eo3
for tZ
(3.15)
39
Substitution of this expression and Equation 3.2 into Equation 3.9 yields,
co + l [E(t:0) JC(t, to) - 11 = Jc(t, to) a(to) aE( V, to) dt' at! (3.16)
ft JC(tltl) t
and rearranging,
) Jc(t, t,, ) ( co -cI) E(t,, cl = V, to) 3EW JC(tltl) t+ 0 dt' at, and if (3-17)
cl)
if
satisfied,
co E: l,
then dividing by (c0-E: 1), Equation 3.13 is obtained. Thus Equations 3.4 through 3.8 are correct and exact for any values of co and c, .
The determination
of the aging coefficlent, X , requires the stress from the chosen creep function. 45 integral in Equation 3.13
For this purpose, time t may be subdivided by discrete times (r = 2,3 tr -t n); one r-1 ...... load
t, = 0. If the integral equation in Equation sum using the trapelzoidal rule, then
+ JCr,
s-1)
(3.18)
40
After
subtracting
the f orms of Equation 3.18 f or t=tr equation for the increments, AER=
r IIAE S=l
Rs
(j
Cr, s +i
Cr, s-1
Rs
(3
-1
XII S=l
(3.19)
where ic,,
the is E(to) ic(trts) ERI and starting value = = s Computation of the relaxation function and the aging coefficient
from
Equations 3.19 and 3.8 is a simple task which can be undertaken by a computer. A listing of a program based on these equations, written Fortran IV, is given in Appendix A. The use of the trapelzoidal in
rule which
is a second -order method have resulted in a degree of accuracy which is 43, tr chosen are small satisf actory provided that the initial time steps 2 and that they increase in a constant ratio, such as Atr/At r-l 10
IA6
Within the range of working stresses, the response of concrete under sustained compressive load, can be approximated quite well as a combination of elastic and viscous behaviour. This type of material behaviour is hence
termed viscoelasticity, and is more commonly associated with the mechanical behaviour of plastics. As far as concrete is concerned, attempts to simulate t !me-de formation by a model consisting of elements, each
41
characteristic,
is largely empirical
that is, elastic and viscous, which are behaviour of concrete are A perfectly if the
represented by a spring and dashpot element, respectively. elastic body is one that exhibits completely
reversible deformation.
load-de formation relation is linear, the body is represented by a linear spring with an equation,
6s
=wP
(3.20)
P is the applied load An ideal viscous body undergoes a shearing deformation at a rate which is a function of the applied shearing stress. The mechanical device representing an ideal viscous body is a dashpot, with a piston moving through a fluid of viscosity, v. moves with a velocity dydt Under a sustained load P, the piston
such that
d6p
pv
dt of the piston
(3.21)
where
6p is the displacement
42
as a Kelvin or Voight model, and a Maxwell model, as shown in Figures 3.1. In the Kelvin model, the spring and dashpot are in parallel so that they undergo the same displacement, the total force on the model being the sum of the forces on the individual elements. In a Maxwell model, the spring and dashpot are in series so that they take the same total load, the total displacement being the sum of displacements of the two elements. It should be noted that, since no force, however large can cause a finite instantaneous deformation of a dashpot, a Kelvin model alone cannot or relaxation.
Considering the Kelvin model in more detail, when a load is applied suddenly, the element exhibits no instantaneous deformation, load is carried initially transferred since all the
of a Kelvin model of
the spring alone, that is, Pw. If the elements are linear, then the model can be represented by the rheological equation,
1d p= -6s+ w vdt 6p (3.22)
and since
6s =6p=6,
then
pv
d6 dt
(3.23)
43
The solution of the above equation, with boundary conditions 6=0, when t=0 is
Pfe
-tWV
(3.24)
Ii)
P (a) (b)
Figure 3.1 Basic rheological models (a) Kelvin model (b) Maxwell model
Thus the deformation of the Kelvin model increases exponentially time, and there is complete recovery upon removal of load. This deformational illustrated behaviour under sustained load and after its removal is
with
in Figure 3.2 a. If the elements of the model are non-linear curve would change, but the general
Considering a Maxwell model, when a load is applied the extension of the spring is 6S =wP, d6P P rate dt =vand the deformation of the dashpot proceeds at the
6=6
S+
6p
(3.25)
44
and since the load carried by the two elements in series is the same, then
ss =d6p =v
(3.26) dt
differential p d6 dt
W+-= dt v
(3.27)
Load removed
-0 9
-4 4,
c C,
-4
I
0 Time, t
0
(b)
Tine,
(a)
Thus, it can shown that when the viscosity of the dashpot is high, a Maxwell model degenerates into a Hookean spring. on the other hand, for a low value of w, that is, when the spring is very stiff, the model
degenerates to a dashpot. Essentially the Maxwell model represents a liquid, since its deformation is unlimited. Figure 3.2 b shows the
behaviour under a sustained load and after its removal. Unlike the Kelvin
45
is non-linear, the displacements OA and BC are not equal to one another, and if the dashpot is non-linear, the line AB becomes a curve.
The Kelvin and Maxwell models are often used in combination to simulate the visco-elastic response of concrete creep. One such built-up
model is the E3urgers model, which is a series combination of a Kelvin and Maxwell model as shown in Figure 3.3.
0
lk
to
Under a sustained load, the model deforms instantaneously, followed by a time dependent deformation increasing at a decreasing rate and tending asymptotically to an inclined straight line, as illustrated in Figure 3.4.
On removal of the load, an instantaneous recovery takes place, which exponentially deformation approaches a horizontal line with time. A part of the under load is not recovered. Thus, eventhough the behaviour of
Stress-strain
laws f or creep
46
Burgers model and the creep behaviour of concrete is far from satisf actory 13
Load sustained
Load removed
go c c -1 4J
I
Figure 3.4 The deformational
i Time,
response of a Burgers model
A modified Burgers model, which makes use of varying retardation and a Maxwell dashpot having limiting and Kesler 46 deformation,
time
for which the comparison was done. Over the years, a number of rheological models of varying complexity, mainly based on the visco-elastic concept of
47 48 concrete have been developed by various workers such as Ross , Flugge 38 49 Cowan Gopalakrishnan50. Freudenthal and ,
47
is deformed, experiment has shown that the stress in to the strain at the same location and time,
is proportional
and does not depend on the past history of strains to which the material has been subjected. This enables a linear theory of elasticity formulated upon the assumption that stress a (x, t) is a function Similarly, experimental to be of c (x, t).
viscous f luid is dependent upon the current rate at which strain is changing with time, that is a (x, t) can be expressed as function of (x, t)/Gt. Thus a viscous fluid "remembers" the past history of 3E: deformation only to an extent which can be expressed in terms of the the basis
current strain rate. These two simple physical observations f orm of classical linear theory of viscoelasticity. Linear viscoelastic materials exhibit a relationship
subjected at all times T up to and including the time t. Mathematically this can be expressed in the form
t (X, t) =FE:
(XE)
(3.28)
where
F is a functional
of -c(x, -r)
if only the linear mechanical behaviour is considered, F becomes a linear continuous functional and Nagy 51 of its argument function. the form, Accordingly, Rlesz
(t) cy
t fC
(t, T) (T) dT
(3.29)
3C
and represents
the material
response
3.29 allows
of material to produce
aging, a given
in concrete,
48
history is not necessarily the same as that required to produce the same strain history at a different time. The material response function in the function of in an
above relationship can be replaced by ER, the stress relaxation the material, alternative and the linear integral relation can also be written
form by transformation
a(t)
f Tr -CO
C(T)
ER(t? T)
(3.30a)
or
(T) 2a E (t, T) dT (3.30b)
a(t)
=-E -CO
element at a
given time is expressed in terms of the strain history in that element. Conversely, one can also consider the dependence of the current strain on the past history of stress and that by analogy write,
ta
(t) E: f TT (-r) a J T) c(t,
dT
(3.31a)
-00
or
t J -Co where ER IS the relaxation function dT T) c(t, (3.31b)
JC is the creep function, and is sometimes referred to as the age of the material of loading. at the time
49
The functions ER and ic can either be determined from experiment from the physics of the material structure, and it is possible to derive
or
forms of these
functions for various spring and dashpot models can be evaluated readily from the material constants52, although it tends to get rather complex as these
functions can be determined quite easily by performing creep or relaxation tests on the material itself.
3.5 Summary
The creep laws which have been examined in detail here, form the basis of the analysis of creep effects mathematical limitations. in the subsequent chapter. These laws are
However the magnitude of the error depends very much on the of the concrete.
The stress-strain relationships are given in a general form and the basic material parameters either in the form of the creep coefficient into these relationships. or creep
Although discussions have been mainly centred on three different stress-strain relationships, other forms of linear creep laws have also been proposed by various investigators. Creep laws in the form of a
53 54 first-order differential equation were proposed by Glanvil-le Levi and , 56 Arutyunian55. All these laws correspond to some viscoelastic model . Overall, the effective modulus along with the age-adjusted effective
modulus, are the simplest f orm of stress-strain relationship for creep. In effect by using these laws the problem is reduced to a series of elastic
analyses. Unlike some other methods it does not require solving of differential or integral equations. The age-adjusted effective modulus
50
relationship
is slightly more involved, since it requires computation of However this is only necessary if the effect of
aging becomes significant. The concept of viscoelasticity theoretical makes it possible to formulate
models of creeping concrete. However, an analysis which law in the form of differential or integral
incorporates a stress-strain
resort to numerical methods to obtain the solution. A more generallsed form of viscoelastic stress-strain law has been presented, and this could be based on different
developed further
rheological models.
CHAPTER 4
4.1 Introduction
The occurrence of stress redistribution major and significant effect on the structural
hypothesis. Various analytical forms of the problem soon developed, such as 58 59 9, 13 14 those proposed by Glanville Torroja Sackman Neville Rusch and , , each one varying in approach and complexity. In this chapter, the case of a cracked beam is studied and the stress-strain relationships derived previously are used in conjunction conventional reinforced modification effect beam theory. A simplified with
of the effective
are developed and presented. The objective of the analyses is to arrive at the relationships which would allow for the computation of the position of the neutral axis and of
changes in stresses in the concrete and steel with time under the effect creep. This would assist in the calculation of time dependent deflections within the section.
51
52
Although the analyses are developed for a beam, they are also applicable for any reinforced concrete section subjected to a flexural
moment, such as the case of a column being subjected to lateral loads. Furthermore in structural design, it is always assumed that under working in resisting tension and thus an
The hypothesis postulated by Faber In his analysis forms the basis of the effective effects modulus method. A similar treatment to the problem of creep
is presented here, using the analysis of a cracked section at the limit state. A triangular stress distribution is assumed for
serviceability
the concrete in the compression zone, which generally occurs at loading levels encountered under working conditions.
Under the assumption of a cracked section, concrete in the beam is capable of sustaining compression only, while tension is being totally resisted by the reinforcement. Furthermore by confining the analysis to the proportional to elastic proportional to
strain, and within the compression zone, both are directly the compressive stress in the concrete. The usual basic assumption of elementary
plane sections remain plane during bending, so that across the depth of the section there must be a linear distribution does not slip relative to the immediately of strains. The reinforcement surrounding concrete of the beam, and the
53
adjacent concrete.
of load,
Fc (to)
(to)
(4.1)
cc (to) Ec(to ) ac(to), cc(to) is the modulus of concrete at loading, and are the instantaneous concrete stress and strain
where
With the assumption that both elastic and creep strains are proportional to stress in the concrete, the required time dependent modulus,
Ef jt) =-=
ccW
+c cc
(4.2)
where cc (t), ec(t) are the concrete stress and strain at time t E cc is the creep strain in the compression zone.
proportional
in the concrete, than it can be def ined in terms of the creep coefficient, 0 (t, to). Thus, Equation 4.1 can be rewritten
CC Efic(t) =
in the form,
(4.3)
54
(4.4)
It can be seen that the stress-strain relationship Equation 4.4 is slightly different
expressed by modulus
of an extra term,
concrete stress.
Equation 4.4 can be expressed in a more convenient f orm, as
(1 4.5
is the modular ratio at Instant of loading, is the modular ratio at time t, is the creep coefficient, and
4.2.3 Derivation
With reference to Figure 4.1 and considering horizontal internal f orces at the time loading,
equilibrium
of
As Es c s(to) =J
(to) bd cL,
(4.6a)
loading,
As Es gs(t) =
ac(t) b
(4.6b)
55
Ed
cross-secion
Strain
diagram
Notation
b h d d s AI s k cc Es s t0 t A
width depth depth effective area area neutral strain strain strain time time
of of to
of of
compression loading
at after
Initial loading
56
Similarly
writing
the equilibrium
M(t
cr c(to) c)+aI
b k(to )d
)d k(to (d -3s0
(t ) (d - d')
(4.7a)
M(t) =ia
(4.7b)
Equations 4.6 and 4.7 represent the time dependent equilibrium equations upon which the subsequent derivation will be based.
By substituting
p=
p E c =J s(t)
(4.8)
Now
(4.9)
k(t)d k(t)d - dl
k(t)d
and
W=E E'
(t)
sc
(4.10)
n(t) (I - k(t)) =I
57
k(t)2
4')
=0
(4.11)
k(t)
(p +
pl)2n(t)2
(p n(t) +2 + pl-d-) d
n(t) (p + pl)
(4.12)
The solution for the instantaneous case Is obtained when t=t.. (t, to) = 0, n(t) = n(t ), and 0 I
Thus,
k(t
000d0
n(t ) (p + plA
n(t ) (p + pl)
(4.13)
Equation 4.13 is the standard and recognised expression in reinforced beam design reinforced for calculating , the neutral axis position in a doubly
section.
Referring
k(t)
(t) E: cs
E(Q
+ E:
since E
f4t)= .L
Es n(t) =E fic(t)
ac(t) (t)
(4.15)
(4.16)
58
and similarly
a, Q= S(
ac(t) NO
dl ) (i -1) t k-
(4.17)
Under constant applied moment, the instantaneous moment of resistance and the moment of resistance after time t, must be the same. Thus from Equations 4.7, M(to) = M(t) and the identity yields the following,
and rearranging,
k(to) d' k(to) (1- 3 )+ P, GS(to) (1-8C(t 0) k(t + '3 (1k(t) P' cy's(t) c(Q
d')
From Equation 4.17, cysl(to)and asl(t) can be expressed in terms of ac(to) and ajt) expression, respectively, and the following is obtained from the above
-a-
Therefore,
(4.18)
59
n(t)
(4.19)
%(t 0)
Substituting
(4.20)
the relationship in Equation 4.20 into Equation 4.18, dl dl 1 Ut )+ pl n(t) (1 -kTQ-d ) (1 -8- ) k(t) (1 3 dl dl li(-t-o) '+ ) k(to) (1 p, n(t. ) (i --1 (1 ) -) - -j kTt -8
0)
by k(t) throughout,
k(t)2 (1
k(t) _ k(tO
k(t ) (1 -
I-k(t)(P 2[p + plg pl)](l + d3dd k(to k(t )(I p, n(tXl+ 3k this expression in a simplified 2A-
m )+ k(t) B1 (1 3
(4.21)
d' -d
60
with time is given by Equation 4.21 and this can be expressed in quadratic f orm. Thus,
2B k(t? 33E
PIC
2) =0
(4.22)
k(t) =
b ,/
-b2
4ac
(4.23)
2a
where
a2 b and c 3 (2 3+-! 13 A - P' + P' C pf C 2)
(2 A _pf C+
p, C1 and D
0, then
A= 8p
k(tc)) k(t ) (1 3
2p 3
(ip
and 2
61
Once the position of the neutral axis Is located with respect to time, the stresses in the concrete and steel under the effect of creep can be
computed and studied. The maximum compressive concrete stress after time t can be found from Equation 4.18 while the stresses in the reinforcement given by Equations 4.16 and 4.17. are
reinforced concrete beam can be modif led to take aging into account. The aging coef f 1clent, x, fictitious is introduced into Equation 4.4, which defines the relationship is therefore
(t, to)
of which is described in
the computation
n(to) (1 +X(t, tO)O(t, tO)). This is the only modification include the effect of concrete aging In the analysis.
range, time dependent response of concrete In compression could be approximated quite well by the assumption of linear viscoelastic Thus, Faber57 who initiated the effective behaviour.
62
of creep
in a singly to cover
is extended which
the effects
of compressive case.
reinforcement
as well,
represents
The geometry and coordinate system of a beam of rectangular cross-section is shown in Figure 4.2. Up to some initial to, the beam is supposed to be in a completely time, denoted by
beam is subjected to a transverse load which gives rise to a bending moment distribution M(x, t).
-ve
p-I
Neutral ails d
Linear
strain
distribution -I .
+Ve
The basic hypothesis of beam theory that is, plane sections bef ore bending remain plane after bending applies. This implies that the longitudinal strain c(x, t), is then linear in y, thus
63
(4.25)
In the foregoing analysis, the concrete is assumed to behave as a linear viscoelastic materIal any magnitude. Furthermore, in compression and to crack under tension of the reinforcement is linearly elastic and there
stress-strain relationship
derivedin
Section 3.4.2 describes the behaviour of the concrete. Thus rewriting Equations 3.30 and 3.31 with the subscript c denoting concrete, then for t> tithe material behaviour law for a typical concrete fibre can be equations,
(X, Ec t)
(x, T) JC(tT)
dT
(4.26a)
Ec(x, t)
t
=-fac
(x, T) aT
ic (tj)
dT
(4.26b)
ti
a (x, t) =fE cc ti t
(X, t) 0c fEc (XPT) aT
(X, T) ER (tj)
dT
(4.27a)
ER(tj)
dT
(4.27b)
in the fibre. The other terms are as defined earlier in Section 3.4.2. The stress-strain relationship appropriate to each fibre making up the
concrete of the beam may be thought as being the law governing the
64
mechanical model shown In Figure 4.3. The knife edges, A, come apart in tension, and the general linear viscoelastic model, B, which is effective
Cc
Rigid rods
<Ik
Mife edge
AA
Rigid rbd
material
4.3.3 Derivation
of governing equations
The major task in the analysis Is the determination deformation substitution equilibrium integral
of the basic
functions, a and B, of Equation 4.25. It begins with the of the viscoelastic stress-strain relationship into the
equations of the beam. As a result, a single non-linear k(t) is derived. once k(t)
65
4.3.3.1 Equilibrium
equations
ff and
ff A
y, t) dydt=
(4.28)
(x, dy t) dt = Y, Y a
M(X, t)
(4.29)
area
of the beam. At a typical cross-section there is a moment, M(x, t), but there is no resultant axial load.
4.3.3.2 Volterra-type
integral equations
over the cross-sectional area of the beam, since the relations hold in different regions of the beam
stress-strain
t. when the load is first applied, there is an immediate strain response, as indicated by the line to. Above yo the concrete Is in compression while below it the concrete has cracked and carries no stress. As time progresses, the compressive stresses in the concrete relax and the neutral axis penetrates into the initially crack partially initially cracked zone, so that the
cracked zone which had been previously ineffective. cracked zone contributes
in the initially
from the time they first become compressed until the current time. The time t, at which a fibre located at depth y, first goes into compression is
66
Reinforcemen
t
-4
41
4J
d 0
(D
Reinforcemen
For the cases under consideration, the neutral axis of strain coincides with that of the stress. This fact is used along with the notions mentioned above, in the equilibrium equations and results in
-b
k(t)d 1 t.
Ec(x, -r)
-la aT T
4(t,
T) dT
1 dy
+ As
Es
r:.s(x, t)
t) =0 + As Es cs'(X,
(4.30a)
and
-b
k(t)d j
c) t*
2(x, c C T) aT ER(tj)
dT 1y
dy
+ As
Es
(x, t) d c.
(4.30b)
where t* = k-l(y/d)
67
Substituting
f or the strain,
-b
fo
k(t)d
if, [a(, [) + t*
tI
aT
ER(tpT)
dT
dy (4.31a)
+ As Es [a (t) +B (t)
and
k(t)d ty
(T) t *[Ct +-66 (T)
-b
fo
If
ITT
ER(t'T (t)
dT
+AEa ssd
A's Esa
In the above equations, for the purpose of convenience, explicit mention of the span coordinate x is supressed. The space integration extends over the currently y= compressed region of the concrete from y=0 includes the contribution of each started to
fibre in this region only from the time, t*, that compression first
in the fibre, up to the current time, t. The terms outside of the integral signs represent the contributions reinforcement. Thus the governing equations of the problem, Equations 4.31, are two simultaneous, non-linear, By introducing Volterra-type integral equations In cc and . from both tensile and compressive
concrete, and the modular ratio, Equations 4.31 can be expressed suitably non-dimensionallsed,
k(t)d t
f0T
fft.
(T Ct
(T)
S*(t, T)
dT- I
dC
+np[
a(t)
+B (t)
+np,
[ a(t)
d' +j
B(t)
(4.32a)
68
and
k(t)d -fo
Iftj
dT I
+np[a
(t) + (t)
+nP,
W+u0
d'
(t) I -a-
d'
p(t)
(4-32b)
where
y/d
p= n and I(Q
As/bd Es/Ec
N/bd pl = ,
2 M(t)/bd Fc (28)
Figure 4.5 and then adding up all such strips to cover the region of integration shown. It is convenient to change the order 'of the iterated first along a vertical strip, as
shown in Figure 4.5, and then to add up all such strips to cover the integration region. The equations resulting from this process are,
tI to
fok(T)
;C TT
)3 S*(t, T[a
(T)
(T
dE
dT
[a (t) n +pIn -p
(4.33a)
ana
t k(T) a
10[ II to
TT
s*(t, T)
(T) +
(T) d'
d d'
dT
-P n[a
(t) +
pl n
R (T) (t) + EF
a-
-M(t)
(4.33b)
69
y
k(t)t
k(tOld
to
Time
f to
S*(t,
(T )k2 (T)
dT
-p na+p,
(t) +0 d
(4.34a)
and
ft
3[
[a n -P
to
5T
S*(t, T) +a (t)
a(T) k p, n[
2(r)
+1
B(.
r)
k3(T)
dT (4-34b)
dl 3J
are
70
t
0
dl
to
r) +0 (T) 0
e(tT )d -T (4.35a)
e(tT)dT=
and
a (t)
k+
2 (t)
3- 0 (t) t
CL (T)
1ta
k3 (t) dl
(T)
pnf a
[a to
(T) +B (T)
dT
] yT
e(tT )dT
-pF ng[ d
fto
+ ja
e(t, -r)
= -M(t)
(4.35b)
e(t,T)
mM
EC(28)
tI ft, p(t) 0
J tj) e(tT) dT
3T
The two governing simultaneous, non-linear equations, Equations 4.35, can be reduced to a single governing non-linear integral equation. Subtracting Equation 4.35b from 4.35a, gives
+ pl nf
6 (t) k2
(t)
TT e(t, T) dT=
-M(t)
(4.36)
then
a(t)
Oft)
A]b d
a(t)
t ft'O
(t) k(t)
= -m(t)
[i p, n +
e(tT ) dr
and simplifying
further,
71
(t) k2(t) a -L + pl n
d
Ia
(t) k(t)
t0T
pn
t1a
IT knT
e(tT)
e(t,-u) dT
dT =0 (4-38)
-pl na 0d
Equation 4.37 yields an expression for a(t) in terms of k(t), and this is then substituted into Equation 4.38 in order to arrive at the solution for k(t). Oncea directly ,a and k are evaluated the stress distribution follows is
as = Es cs =Es(
a+ 0)
The stresses
from
The governing non-linear Integral equation is solved numerically, discretising the equation and approximating some time _,, the integral.
by
and its
value is sought at t,
1 ai k W
+ p, n
ai k, + pl n d] Pd
dd
(T)
d'
dT=
1a e(tiT) -R k-ktT
-Ml
dT
(4.39)
[1 a( T)
11 e(tiT RTT aT
72
Evaluating the second integral above using the trapelzoldal Equation 4.39 becomes,
rule,
ki 6 ai
12_I
iki+p,
n [I d
+ pl n
dd
fa t0
f[h
e(ti, ti)
ai (i
1
ki
hi_l gi_l
m1
(4.40)
dl
Tnl T 'r
3T
dr e(ti, T)
Rearranging Equation 4.40 and expressing it in terms of ()t,, the following is obtained, 61mi+ A) pl n (I [W hi-, gi-11 + i_, All[l dd A'l I+3 ki ki - k2i
(4.41)
6 PI n [h id e(ti, ti)][l -1The identity substituting p, =0 for a singly reinforced and d' =d 6mi
ai
3ki-
ki2
k1 (3 - ki)
9 in the analysis carried out by Sackman by initially discretIsIng and rule.
subsequently evaluating the second Integrals using the trapelzoldal This gives,
where
Vi-i
(T) ot
kT
aT
e(t,,
T) dT
73
and
Into _,
Aik-'i + Bi k-2i + Ci ki + Di =0
Ai = Vi_l + hi_l fi_l + R! Wj_j + Ehi_l gi_l pp 1 (1 dl hi-I W gi-11 -3V + +3-! -m' _ j_j pnpV P' hi-, Wi-1 gi-, -3 IT -3V p d'] d
Bi
i-I -3
fl-i
fi-I
[hi_, )] hi_lgi_, ]l t, e(ti, + -61m, d'] [h [l hi_, f ti)1 pl n e(ti, +6 1-1-d i_, d] [hi_, [1ti)] Wi-, pl n e(4 +62 pd (1 hi-, ]1[hi-, ti)] e(ti, n gi-, mi pl + + -691 pd _dl)[Wi_l t gL 1 [hi-, hi_, ti n e(ti, gi-, pl +69 pd -
Di
The proper root, k 1, of the cubic equation can be obtained by using a suitable method of approximation. A good starting value of ki is necessary process. This is achieved
simply by using the value of k at the previous time stepi that is kj-, as the initial approximation. The numerical procedure is initiated by setting
t j= to, and this reduces the cubic equation to the simple quadratic equation associated with reinforced concrete beam design.
74
4.4 Sumary
both methods are based on conventional beams. However, the use forms of
governing equations for the position of the neutral axis. In the simplified method, the problem is reduced to an equation of the second order, while in the classical approach, using the concept of viscoelasticity, is essentially third order. The computational work in the analysis presented were quite involved in Fortran IV, were the problem
and programs for both methodsof analysis, written prepared. A typical problem involving viscoelastit
computer time. Listings of these programs are given in Appendix A. Both methods of analysis are based upon the assumption that the neutral axis shifts towards the tensile reinforcement with time and that of the
neutral axes which occurs if shrinkage and thermal strains are included, are considered later in Chapter 7. Furthermore, the methods
stress, although in real structures this Is unlikely to be true. However, for practical sufficient limitations purposes, an analysis based on cracked section is usually members subjected to working loads. All these
for structural
subsequently would not result in governing equations which are highly complex.
CHAPTER 5
THEORETICAL
STUDIES
5.1 introducUon
of neutral
compressive
concrete
stresses
of creep
in a reinforced
have to be assessed in terms by comparing of the problem. theory provides beam. methods with
of accuracy
applicability, theoretical
either solution
experimental
According a correct
to Bazant theoretical
in a cracked
In this chapter,
and
age-adjusted moduli are compared with the viscoelastic Parameter studies based on different reinforcement ratios and effect
aging method.
of compressive reinforcement
75
Theoretical
studies
76
of
the concrete in the compression zone cause the position of the neutral axis to shift towards the tensile reinforcement. stresses are redistributed, Correspondingly the flexural
tensile steel stresses as the compressive concrete stresses decrease. All these effects occur gradually and are very much dependent on the creep properties of the concrete. By using the methods of analysis elaborated in the previous chapter, curves depicting changes in the neutral axis position, stresses and strains within the reinforced obtained. Comparative studies on the analytical methods employed were carried section can be
out on both singly and doubly reinforced for the doubly reinforced
d/dI = 0.185, which are the values for one of the test beams describedin Chapter 7. For the singly reinforced section, an arbitrary value of 1.5%
for the tensile steel ratio was chosen. The expression for the creep function given by the ACI 209 method was utillsed throughout mainly because the methods of analysis. The
Young's modulus for steel and concrete at 28 days was 200 kN/mM2 and 25 kN/mM2 respectively. , beam
The constant moment applied to the singly and doubly reinforced was 77.1 kNm and 97.7 kNm, respectively. moment of the sections under consideration.
This corresponded to the cracking The moment was applied when the or 240.
Theoretical
studies
77
in a reinforced concrete cross-section, the neutral axis position determines the local curvature and deflection. Thus any movements in the
position of the neutral axis, such as those caused by creep of the concrete, strongly influence these two major variables. When subjected to a sustained flexural in fibre a reinforced compression extreme moment, the strain in the concrete section, increases due
to creep. Consequently, the neutral axis moves downwards, as shown in Figure 5.1, and causes the curvature and deflection to increase.
CS(t)
I
b
-____
0cj L0cto,
-l
-d
concrete
For different
moment, the
Figures 5.2 and 5.3. As exhibited by the plots for the singly reinforced section, the position of the neutral axis, as given by k(t), changes rapidly during the first three months after loading. The shift then becomes more gradual and carries on at an almost constant rate. The shape of the
Theoretical
studies
78
0.65
10
10
60
-, 240 240 60
0.55
0.50 , 4-' 0 a)
>
240
0.45
0.40
Numbers on curves indicate at loading in days. Effective Viscoelastic modulus eff. aging
age
Age-adjusted
mod.
Days after
loading
section
Figure 5.2 Neutral axis histories for singly reinforced using different analytical methods
Theoretical
studies
79
0.65
10
10
10
t. 55
0.50
co
0.45
0.40
age
Days after
Figure 5.3 Neutral axis histories for doubly reinforced section using different analytical methods
Theoretical
studies
80
curves resemble the creep curve f or the concrete and as shown later, is very much dependent on the creep function or creep path used to represent creep of the concrete. By using the effective simplified modulus (EM) in the analysis, which Is a
approach, the values of k(t) are on average 10% higher than those analysis. This corresponds to a lower
position of the neutral axis and as such the effects of creep are overestimated. Thus, if the viscoelastic values are representative of the of
true and actual values, a designer would make a conservative the creep deflections By incorporating effective
estimate
and curvatures in his design, by using the EM method. an aging coefficient through the age-adjusted
modulus (AEM), the computed values of k(t) are closer to the solution. This is particularly significant at early ages of
viscoelastic
loading when the concrete is affected by aging. However for higher ages at loading, no apparent gain is obtained through the aging coefficient, as its
value approaches unity and the solution from the AEM tends towards the EM. Although these observations and figures apply to the plots for the singly reinforced section, a similar form is also obtained for the doubly
a sustained flexural
moment are
time progresses have to be evaluated or predicted. This forms an important part of the design procedure when the effects of creep are to be included, so as to ensure that the safe working stresses in both the steel and concrete are not exceeded.
Theoretical
studies
81
As the neutral axis moves downward due to creep, the compression region increases. This causes the extreme compressive concrete f lbre stress to decrease and subsequently slows down the strain increase in the fibre. However, the lever arm of the compressive concrete force is reduced slightly during this process, thus causing the steel stresses on the
tension side to increase slightly. In a doubly reinforced section, a decrease in both the concrete stresses and lever arm must be accompanied by an increase in the steel stresses on the compression side to provide the same moment of resistance. Hence, to balance up the internal f orces, the tensile steel stresses also increase. All these effects are illustrated There is a significant in Figure 5.1.
increase in steel stresses during the initial by the plots of steel stress histories in
stress at loading are obtained after 18 months of loading, whilst the compressive steel stress of the doubly reinforced section doubles during
the same period of time. The general pattern of these curves is similar to those obtained f or the plots of neutral axis histories, with the stresses increasing rapidly during the first few months after loading and then
showing a much reduced rate of increase with time. On the other hand, as the steel stresses increase, there is an appreciable decrease in the maximum compressive concrete stress as '8. The decrease in concrete stress ranges indicated in Figures 5.7 and 5. from between 20-30% of the initial reinforced value at loading for the singly
section. The plots exhibit a sharp decrease in concrete stress during the first few months which is later followed by a more gradual decrease. As mentioned earlier, the EM method has a tendency of overestimating the effects of creep if compared with the viscoelastic solution. This fact
Theoretical
studies
82
145
10
10
140
10
##,
0100 ..
60
60 40 ... ...........
135
240
fn c Q)
130
age
Effective
Days after
Figure 5.4 Tensile steel stress histories for singly reinforced using different analytical methods
section
Theoretical
studies
83
115
10 10
60
240
110
(0
4J (1)
105
age
Age-adjusted Viscoelastic
100
mod.
60
120
180
240
300
360
420
480
540
Days after
loading
Figure 5.5 Tensile steel stress histories for doubly reinforced using different analytical methods
section
Theoretical
studies
84
10
120-
10
0-% 44
60 ** /
100-
60
10
cu
24D
4 (1) W
80 cu
E 0
5--o-O 00*
AU -
60
I,
I/
60
.
-. . ',
'7/,
V
Numberson curves indicate at loading in days.
40 Effective modulus , Age-adjusted eff. mod. Viscoelastic aging 0 60 120 180 240 300 loading section 360 420 480 540
age
Days after
Figure 5.6 Compressive steel stress histories for doubly reinforced using different analytical methods
Theoretical
studies
85
10
E
cn
4)
c 0 L) W
cl E 0 u >Z ca x
60
120
180
240
300 loading
360
420
480
540
Days after
Figure 5.7 Maximum compressive concrete stress histories for singly reinforced section using different analytical methods
Theoretical
studies
86
age
Ef fective
Age-adjusted viscoelastic
E N,
4 0)
cu k
B
240 60
(U 41 a.) k C. ) c 0 L) Q) W
cl E 0 L)
C13 x
240 240
6
60
10
60
10
Figure 5.8 Maximum compressive concrete stress histories for doubly reinforced section using different analytical methods
Theoretical
studies
87
is affirmed
compressive concrete stress than the other methods. Thus, the values obtained from the EM method represent an upper bound for deflections, curvatures and steel stresses associated with creep in reinforced concrete
members under flexure. The AEM, as before, yields values which are closer to the viscoelastic method, for early ages at loading. Under working loads, steel stresses increase due to creep. The magnitude of the increase might be of little side, but in the case of a doubly reinforced significance on the tension
stress of the compression steel is appreciable. This would require a designer to carry out an estimate and check that the safe working stress is never exceeded.
beam theory applies, namely plane sections remain plane af ter bending. Therefore a linear distribution of strain is assumed to occur across the changes
with time as the neutral axis shifts and the strains in both concrete and steel increase. Figure 5.9 shows the strain profiles for the singly reinforced for different ages at loading obtained using the viscoelastic section
method.
while those obtained using the other methods of analysis are of similar form. By assuming that concrete below the neutral axis is cracked, creep is confined to the compression zone of the section. This accounts for the
Theoretical
studies
88
OD x No c
U) 4)
0 04
CY) c
0
L. C)
00 x
CO 0
co 4 c:x c .i -e
E ol c 0) ca C)
ca (1) 0
%G .1 cl 9' L) Co 1 0 'A
0) c 10 c D c 0
c2
(n
(0 C:)
'0 cu
0)
0
c2 CD CN x m1
. \o
>1
Co (U cro c:c
0) c
0 C. V) (L) 40 w cl c
Co a
Go
Ul%
0 x
-,a
co w
ca
C4
0 --q
LL
c cc 0 , --I 4J co (D ED
%0
1 0 ,
%D
8v
Theoretical
studies
-0
CX)
89
10 %0
4j LO C!
(0
0 c C
CD C)
x
(U
E cn c 0) m
L)
cu
C1
cr) c c:
4-b U)
%o_
-0
-0
AIr
(1) 0 L)
OD c 'l C:)
64 41 IA x 4
c (n D c 0
(0 CD \o
4-) L) Q) U)
(D L)
S 0
C
jo
(D -
11
T) c
4J (1) cu cm 9c
43 1 0
W 0
(0
41
(n
'D Co 1 0E 1 8
-0
0
CO
.4 0
-4 C -4 10 4 4-I
6. cl
ct) V) M 0 P-1
I10
LA W
0 LL CM c "A 13 (a 0
co 0) CM cx
r; %0 -
10
; 74
;a WN
Theoretical
studies
90
free
A,' ForCeS
Z'
on
steel
AS
orces on oncrete
measured
shrinkage
strained
shrinkage
Creep strains
Shrinkage I
strains
Combined
strains
Gov.
Ton.
C. P.
Tan.
Conp.
Ton.
Strain
distributinn.
across
section stresses
Shrinkage
stresses
Combined
COM.
Ten.
COMP. (b)
Ten. section
Figure 5.11 Combined effect of shrinkage and creep on (a) strain distribution and (b) concrete stress distribution
Theoretical
studies
91
relatively
in the neutral axis position and compressive concrete strains. It is also apparent from the profile plots that the older the concrete at loading, the
smaller the changes in neutral axis position and compressive concrete strains. With aging, concrete becomes stiffer with time, and its tendency discussed in Section
The values of strain plotted are those arising solely f rom the ef f ects of creep. However, in real life, creep does not occur on its own, but is always accompanied by shrinkage. In a reinforced member, restrained
shrinkage causes additional compressive strains across the depth of the section. Thus, by assuming uniform free shrinkage across the depth of the section, the effects of shrinkage can be included in the strain profiles by from the two time dependant
is lowered and the compression zone of the section increases as a result of shrinkage strains. Moreover, the magnitude of the maximum compressive strain increases whilst that of the tensile strain decreases. Although these effects are illustrated and surnmarlsed for a doubly reinforced section. The
section, the same argument holds for a singly reinforced corresponding effect following section.
Under working loads, the stress distribution compression zone is assumed to be triangular.
compressive concrete stress decreases and the neutral axis moves downward, but the triangular distribution still remains. The stress profiles from the
Theoretical
studies
92
effective
and 5.13, ref lect this behaviour. Also shown in Figures 5.12 and 5.13 are the stress profiles obtained from results of the viscoelastic simplified aging analysis. Unlike the other two
boundary of the compressive zone and of the various times at which the concrete fibres start to experience compression. As a consequence, the relationship initially which gives the solution of the concrete stresses in the However, the stress distribution
in the initially
with respect to the depth. The additional compressive strain caused by restrained shrinkage in the reinforced section, as shown in Figure 5.11a, produces compressive
stresses in the steel. This yields a reduced tensile steel stress while the stress in the compression steel increases. However, tensile stresses are produced instead, in the concrete adjacent to the steel. These arise as a result of the concrete being restrained the steel reinforcement. from shrinking to its full value by in the
concrete stress is raised as opposed to being lowered in the case of the neutral axis of strain. Thus as a result of shrinkage, the plane of zero strain is not the plane of zero stress and there exists a region of tensile stress in the concrete within the region of compressive strain 61
In many practical
may vary
Theoretical
studies
93
concrete
stress
(N/mm'
20 in 8 10 8 6.4 20
k(t)
> V
\u
.4
a
IA
.5
Age at loading
Effective
Theoretical
studies
94
Comoressive concrete
10 8642 10 0 8642 10 0 8
stress
(N/mm')
642 10
0 86420
k(t) 0.1
0.2
cc
w4
d 0.3
IA t
4J
0.4
0.5
0.6
0.7
Age at loading
Effective
Age-adjusted Viscoelastic
Theoretical
studies
95
construction
sequence or a gradual build up of the live loads and it causes to the loading conditions. Most are based on a constant
stress history which according to McHenry of varying stress by using superposition. The underlying principle
time t, by a stress increment applied at any time t', are independent of the effects of any stress applied either earlier or later than t'. The method of superposition requires the creep-time curves under constant
stress at the ages when either the stress Increments or decrements were applied f or the calculation simplification of strain under a varying stress. This creep recovery since it assumes that
tends to overestimate
the recovery is equivalent to the corresponding creep Itself. Hence, under decreasing stress the method tends to underestimate strains 62,42,63 by Backstrom Tests conducted . 64 the creep
The principle of superposition can be applied to the EM and AEM methods for the computation of creep effects under varying moment. However,
the relationship which defines the neutral axis location with respect to time is independent of the applied moment, as given in the previous chapter (Section 4.2.3.2). Thus the analysis Itself is deprived of the dependence
on the moment and use of it In the case of variable moment history would result in an incorrect representation of the position of the neutral axis
with time together with the histories of steel and concrete stresses. On the other hand, the viscoelastic aging method defines a relationship for
the neutral axis position which is dependent on the moment history (Section 4.3.3.2) and thus is theoretically methods. correct compared to the two simplified
Theoretical
studies
96
In order to investigate
subjected to two moment histories, as shown in Figure 5.14. The histories of neutral axis position, maximum compressive concrete stress and tensile steel stress from the two cases are depicted in Figures 5.15 to 5.17. It is obvious from the shape of these curves that flexural determined by the moment regime. Under increasing or decreasing moment, there is an appreciable shift in the neutral axis position and in both cases the movement is towards the tensile reinforcement. As a result of creep, the rate at which the strain and causes the neutral creep effects are
in the steel and concrete changes are different, axis to shift. The differencein
the strains and stresses across the section can also be observed by comparing the plots of the tensile steel stress and maximum compressive concrete stress. These results are further compared with the values obtained by
subjecting the section to constant moment values. The values of M/bd 2 for
the two cases of constant and correspond concrete after to the final stresses moment are 0.84 N/mM2 the variable the location in Table and 1.68 N/mm' moment respectively
and steel
of the neutral
5.1.
Comparing the results from Case A and the respective constant moment, it can be seen that the effects former. This is reflected of creep are higher in the case of the
value of maximum compressive concrete stress. At early ages after loading, the creep rate is high and causes greater shift of the neutral axis and redistribution of the stresses. Thus a section which is subjected to a high will utillse the available creep more than a
Theoretical
studies
97
0.8
0.6
Co
"'\
2_ //-M/bd
0.84 N/M,
0.4
0.2
0. 0
20 40 60 Time in days 80
100
120
(a)
0.8
\\-
0.6
M/bd 2.1.68 N/mM2
Co
0.4
0.2
0111111i
20
80
100
120
Figure
5.14
Variable
moment
histories
(a) Case
Theoretical
studies
98
C'ass
30
60
90
150
180
210
240
12
Case
10
U) 'A
8 cu 6 ID
Case A
x :2
30
60
90
150
180
210
240
Figure 5.16 Maximum compressive concrete stress histories due to variable moment
Theoretical
iso
studies
99
160
140
case a
'120
ul
A A 4 loo
Case A
80
60
ratio
in the case of direct stress and will be elaborated later in this chapter. Thus creep effects under gradually decreasing moment are much higher
than under constant moment. Conversely, under a gradually increasing load, the effects indicated by load than less be as also constant under much will
the values in Table 5.1. Although the discussion is mainly centred on a singly reinforced reinforced correctly section, the same applies in the case of a doubly on the compressive steel stress will be
Results from the comparative studies have revealed to a certain degree, the occurrence of differences different in the magnitude of creep due to to further
Theoretical
studies
100
ii
Constant 0.84
moment
0.550
3.52
1.68
0.550
7.68
Table 5.1 Neutral axis and concrete stress after 240 days under different loading regimes
the investigation
factors which might exercise their influence over the curvature, deflections and internal stresses and strains of the member, with time. creep of plain concrete, the role in determining the effects
ages at loading, factors such as different ratios and compressive reinforcement detail.
These parameter studies are based on results obtained using the viscoelastic aging analysis. Values from the viscoelastic analysis are
Theoretical
studies
101
Section 5.2, are used in the following characteristics ACI 209 Code.
as a
in the compression zone. This redistribution the creep curve, but is slightly
modified by a relaxation of the concrete stresses in the extreme 41 compressive f lbres . To demonstrate this, the ratio of the creep strain to the instantaneous strain, that is, the creep coefficient, the singly and doubly reinforced from direct compression. The values of creep coefficient for direct compression are computed by for bending of
using the ACI 209 method which is then used to model the creep behaviour of the concrete in the viscoelastic analysis of the beams under constant
applied moment. From the analysis, the ratio of creep to Instantaneous strain of the extreme fibre in the compression zone is taken as the flexural creep coefficient. As shown in Figure 5.18, the flexural creep
coefficients
are smaller in magnitude and they tend to stabilise somewhat This implies that
more rapidly than do the compressive creep coefficients. under the influence of relaxation flexural
stabillse more rapidly than the compressive c reep strain under constant stress. The effect of relaxation of the concrete stress is greater when the
is present and this accounts for the lower values in the case of the doubly reinforced and reinforcement
creep coefficient,
of compressive reinforcement
Theoretical
studies
102
Compressive creep
1
4J E ci
31
30
60
90 Days after
120 loading
150
IBU
ziu
Z4
ratios are discussed in more detail in later sections. Thus, there seems to be a form of relationship and flexural relaxation creep coefficients, between the compressive
and to take into consideration the factor should be introduced. creep strains are not
Moreover, it should be emphasized that the flexural simply obtained by multiplying coefficient
adopted, the outcome would be a high and erroneous estimation flexural strains with time.
of stresses
due to creep is a process which closely resembles the creep curve. To highlight this point, three different creep paths were chosen arbitrarily.
Theoretical
studies
103
Creep Is instead represented by the specific creep curve which specifies the development of creep strain per unit stress, with time. Within the working stress range, creep is proportional to the stress and specific of the
One route was the calculated specific creep curve from the ACI 209 method and this was taken to be the reference creep path. The next was a constant rate of creep from zero to 60 days, followed by a constant rate from 60 days to 90 days, to give the same creep strains at both 60 and 90 days as the reference creep curve. The third route, referred to as the initial until creep path, assumed that the initial 14 days. Thereafter reference creep rate continued at the same
value of creep strains at 60 and 90 days, as the other two. These different creep paths are illustrated in Figure 5.19. section are illustrated in
Figures 5.20 to 5.22. Similar forms are obtained in the case of the doubly reinforced section. The graphs depicting changes in position of neutral similar in
axis, defined by k(t) and tensile steel stress are distinctively configuration
compressive concrete stress is an Inverted image of that configuration. Thus, the development of time dependent curvatures, deflections, stresses
and strains caused by creep, are very much dependent on the shape and form of the creep curve. Factors which affect as concrete constituents affect and environmental the creep of plain concrete such conditions will ultimately in the
reinf orced section. It should be noted that in this study, the creep curve is derived from the ACI 209 method. Various other Codes of Practice in Chapter 3, give their own representation such as those reviewed
Theoretical
60
studies
104
50
c 4 U) a cu
40
30
F-Age at loading = 10 days Initial creep path Reference creep path Reduced creep path
20
10
10
20
30
40 r)ays after
50 loadinq
60
70
80
90
creep configurations
0.50
0.45
0.40
Age at loading = 10 days Initial creep path Reference creep path Reduced creep path 0 10 20 30 40 Days after 50 loading 60 70 80 90
0.35
of different
Theoretical
140
studies
105
135
// II II
1/
44
130
Age at loading - 10 cays creep path Reference creep pat'i Rediced creep path Initial
125
0 10 20 30 40 Days after 50 loading 60 70 80 90
of different
10
:1 E
U)
cl
cc x
10
20
30
60
70
80
90
of different
Theoretical
studies
106
different
mathematical
functions to determine
representation
configuration
and properties
as hydration progresses with time and this process is termed aging. Past investigators30,66,67 strictly maturity have shown that with an increase in age, or more
through than Influence does Age total creep other of creep. not amount and 13. The same concrete loaded at different ages the stress-strength ratio undergoes a different growth in strength and therefore, for a constant
The ef fect of age at loading on creep can be observed in the plots of the neutral axis position and stresses as shown in Figures 5.2 to 5.8. An apparent feature exhibited in all the plots is that under constant applied moment, there is a decrease in creep potential in stress redistribution. and subsequently a decrease
for higher higher lower stress compressive maximum stresses and steel axis, ages* at loading. Along with these, the rate of increase of k(t) and the steel stresses during the first few months under load, is much greater for
concrete loaded at an early age than for older concrete. This also applies to the rate of decrease of the maximum compressive concrete stress. However, after this initial period, the curves seem to be almost parallel
to each other implying that the rate is no longer dependent on the age at loacling. This is in agreement with experimental observations made by
Theoretical
studies
107
Davis
68
and Glanville
69
loading. Based on the results of five different investigators who used a wide 70 influence the L'Hermlte pattern of a general obtained range of concrete, of age at loading on creep. The comparison was based on the same stress being the days loaded 7 studied. The age and at at acting on concrete relationship as shown in Figure 5.23, is linear and reflects a decrease in
loading the loading. With higher at almost a age at ages at creep with half the is than less the value of creep occurring creep of magnitude year, of concrete loaded at 7 days. A somewhat similar relationship Ruetz's 71 tests the experimental data obtained by Reichard 72 showed the was found in
Furthermore, effect
of stress-strength
between creep
through the loading influence the creep on of age at mentioned earlier on stress-strength ratio. Under the same applied stress, concrete loaded at ratio compared to concrete
loaded at a much earlier age, due to the growth in strength. A lower stress-strength In in lower thus creep. of magnitude a result ratio would concrete member subjected to flexural loads, the
the to the is above argument and stress, applied analogous applied moment holds, making reference instead to the moment-strength ratio.
The effect
of compressive reinforcement
on creep deformation
and
Theoretical studies
1.0
108
A
64 0
0.5
S-4 L)
<> & 0 a
a
7 14 28 60 rr nt loadinn ( days 100 300
of creep for
1600
1200
cc;
800
i
400
0KIII
0.2
0.6
0.8
Theoretical
studies
109
containing compressive steel had the same amount of tensile steel and were subjected to the same value of cracking moment. The compressive steel was chosen to be equal or half the amount of tensile steel. Results are expressed and presented as a ratio to the initial reflect the relevant increase or decrease. There is a definite value to
reduction in
compressive steel. This is revealed in Figures 5.25 and 5.26 which show an appreciably lower increase in the neutral axis position and tensile steel stress. Another significant advantage is the larger reduction of the in
concrete compressive stress in the extreme fibre as illustrated Figure 5.27. A theoretical calculation of deflection
540 days under load together with the percentage Increases are shown in Table 5.2. These results are in close agreement with past experimental observations as reported by Faber in the singly reinforced 57 73 and Washa
towards the tensile steel causes an increase in depth of the compression zone and a decrease of the internal moment arm. The latter indicates that
there should be an increase in the compressive concrete force so as to maintain the same internal moment of resistance. Thus even though the stress in the extreme concrete fibre decreases, there is a nett increase in the concrete force due to the larger area subjected to compression. This is substantiated by the increase in tensile steel stress to keep the internal forces in equilibrium. In the doubly reinforced section even though the movement of the
neutral axis is being restrained by the compressive steel, the presence of steel within the compression zone allows the concrete to shed its load
Theoretical
studies
110
Ao
.X
1.
11
0 60 120 180 240 Days efter 300 loading 360 420 480 540
of compressive reinforcement
1.10 0
1.08
1.06
1.04 zP
4'.to
1.02 -
Ir
I. C -
of compressive reinforcement
Theoretical
studies
ill.
pl
4" 1 ly
. o
Arl.0
60
120
180
300 loading
360
420
480
540
on maximum compressive
Section
Steel ratio
Ten. Comp. t-t'
Miximum x0
deflection t-t'
increisse in
deflection W
days
1 2 3
0 0.75 1.5
49 32 23
on creep deflections
directly
onto it and the decrease in concrete stress is substantially section. This accounts for
an overall reduction in the concrete force and as a consequence a smaller increase in the tensile steel stress to maintain equilibrium.
Theoretical
studies
112
5.3.5 Reinforcement
ratio
Influence on
by varying the
the same area of tensile steel. Thus, the influence of the compressive reinforcement over creep effects can be visuallsed further. Each section is
then subjected to its respective cracking moment. As shown in Figure 5.28, an increase in tensile steel in the singly reinforced section resulted in a smaller relative shift on the neutral axis
position. This is accompanied by a smaller change of the concrete stress in the extreme fibre and a higher increase in the tensile steel stress as illustrated in Figures 5.29 and 5.30, respectively.
As mentioned earlier, the movement of the neutral axis downwards would cause an increase in the depth of the compression zone and an overall increase in the compressive concrete force even though the stress in the extreme fibre decreases. A larger percentage of tensile steel would result in an overall stiffer section which tends to resist the movement of the the decrease in concrete stress is the
balance of the internal forces will have to be accommodated by a higher increase in the tensile steel stress. The area of tensile steel for the doubly reinforced at 2.5% whilst increasing the compressive reinforcement section was kept with respect to the of creep
tensile steel. By increasing the compressive steel area the effect within the section is partially suppressed and controlled.
This is apparent
Theoretical
studies
113
1.6
1.2
1.0
0 60 120 180 240 300 Days after loading 360 420 480 540
0.9
a U 0
0.8
U U 0
pjr20%
P. 1-0 %
0.7
PR05%
P. O.? %
0.6
111a 0 60 120
; 80
540
Figure 5.29 Effect of reinforcement ratio on maximum compressive concrete stress in singly reinforced section
Theoretical studies
1.20
114
pirO.5%
1.04
piro-I %
1.0
60
120
180
240
360 loading
420
480
540
from a lesser shift in the neutral axis position coupled with smaller increases in both the tensile and compressive stresses as demonstrated in Figures 5.31,5.32 and 5.33. section, the compression forces which account
steel. The presence of steel in the compression zone allows for redistribution of the stresses from concrete to steel. Consequently, the
compressive steel stress increases and the moment of resistance provided by the compression reinforcement increases. In order to maintain the internal from the concrete should
stress in the extreme fibre and a decrease in the internal moment arm. A section having a high percentage of compressive reinforcement would stiffen
the concrete in the compression zone and would tend to moderate the shift
Theoretical
1.4
studies
115
-O-l
P. to -As
41
0
1.
.2
1.04
60 120 180 240 Days after
300 loading
360
420
480
540
1.c
1.01
0.9 0 60 120 180 240 300 Days after loading 360 420 480 540
Theoretical
4-
studies
116
2
ExO-2S . 2
vros
pI
1 . 80
.
0
6.
261.0 p
1. 4-
1. 2.
--
-a-II-I-
--
360
420
480
540
Figure 5.33 Effect of reinforcement ratio on compressive steel stress in doubly reinforced section
1.0
0.8
45 0
0.6
0.5
to
0.4
60
120
180
240
360
420
480
540
Figure 5.34 Effect of reinforcement ratio on maximum compressive concrete stress in doubly reinforced section
Theoretical
studies
117
of the neutral axis more than a section having a smaller percentage. A smaller shift in the position of the neutral axis implies a smaller decrease in the lever arm and thus would require a larger decrease of the concrete stress in the extreme fibre to maintain the moment balance. This accounts for the higher decrease in the compressive stress in the section having a higher compressive steel percentage, as illustrated Figure 5.34. in
5.4 Summary
The comparative study indicated that results from the EM and AEM methods gave a higher prediction the viscoelastic of the effects of creep in comparison to methods should suffice
in order to obtain an upper bound of the solution to the creep problem in reinforced concrete sections under constant applied moment. By using these
methods, designers would expect a rather conservative estimate of the time dependent deflections of the viscoelastic the theoretical Furthermore, and curvatures of the structural member. The accuracy
aging method, however, can only be judged by comparing data or full-scale measurements.
applicable to conditions of varying moment and this approach will be adopted in a later chapter to investigate full-size structure. and reinforcement ratio have the effects of creep in a
factors have been considered from the aspect of reducing creep rather than from a strength point of view.
118
0nm) .
Desigry charts
of beam sections
and stress
in the form findings the by of surnmarlsed are creep, caused analysis and are
to doubly for working both subjected section reinforced singly and valid loads. The charts along with worked examples are given in Appendix B. Some Codes of Practice structural do include procedures to evaluate the effects of creep in 74 the it is CP 110 One use of recommends and code such members. the long-term curvature
in for the calculating concrete, a modified modulus and def lection. The predicted long-term mid-span deflection
on the worked examples given in Appendix B are compared with the calculated the in As 5.3. Table from 110 expected, values CP and are presented values from CP 110 are much higher than the values obtained using the design is to the effective by 110 CP the similar adopted charts, since approach modulus method.
CHAPTER 6
CASE STUDY I
6.1 Introduction
Time dependent phenomena in concrete such as creep and shrinkage, are often characterised by either the creep coefficient free shrinkage strains respectively. or specific creep, and
need to be determined. This can either be carried out experimentally, is, through control tests on small scale specimens, or computed directly using any of the prediction methods available.
This case study supplements and provides the background work required in the investigation structure 75 of time dependent behaviour of a full-scale being monitored were buried and data obtained from control
waterproofed
order to find out which method gave the best agreement with the data. However, it was necessary to make adjustments and modifications the methods to account for the effects of waterproofing to one of
119
Case study I
120
6.2 Experimental
work
As part of an extensive study into time dependent behaviour of the abutment columns of a bridge, laboratory tests were conducted on small
scale specimens and two large scale beams prepared from the batch Of concrete manufactured on site. The. preparation, instrumentation and control
tests on the specimens are detailed here. Experimental test beams will be described in the following chapter.
6.2.1 Corqmdon
specimens
These are concrete prisms and cylinders prepared from the concrete used f or casting the two test beams and f our instrumented columns on site. The shrinkage and creep specimens are square in cross-section and the length varies according to the type of specimen. in addition to these specimens, modulus prisms, test cubes and cylinders were also prepared. Initially all the specimens with the exception of the modulus prisms and from
samples from the test beams, were stored in an outdoor shed protected the weather. Later, they were moved into the laboratory were conducted. The specified concrete was a 37.5/20 mix suitable for pumping.
Ordinary Portland cement was used and the f lint aggregates were supplied by Streeters from the Laleham Pit. The mix proportions for one cubic metre of concrete are shown in Table 6.1. A pair of shrinkage specimens were prepared for each instrumented column and test beam. These were 100mm square specimens with a length of 500mm. One fr-om each pair of the specimens was painted with
black bituminous paint similar to the type used on the site columns. This task was carried out a day after casting when the moulds were stripped. By
Case study I
21
Cement O. P. C. 1Kg1
[ litres I
335
1033
757
186
keeping one specimen from each pair unpainted, it was possible to establish the effect of the bituminous paint on drying and shrinkage of the concrete.
The creep specimens were slightly shorter than the shrinkage specimens with a length of 250mm. A pair was prepared for each of the site columns and test beams and all were painted with the bituminous paint after the moulds were stripped.
long period of time is that the measuring device should be reliable maintaining
a datum. This has led to the development and use of acoustic resistance strain gauges.
Basically two types of strain measuring devices were employed. Acoustic vibrating wire gauges with a gauge length of 5.51n. (139.7mm) were
used to record the strains on the creep specimens. These gauges have been known to be reliable in measuring long-term changes in strain at the
Case study I
122
in full-scale
structures
76,77,78 .
On the other hand, shrinkage measurements were taken with a 200mm demec gauge. Although the accuracy of this device is much lower than the vibrating wire gauge, It provides a quick and cheap method of especially when a large number of readings are required.
instrumentation
wire gauge
The principle
of the vibrating
been developed and used as surface gauges on buildings and tunnel linings by Mainstone8o and as embedment gauges for concrete by Potockl8l. An acoustic vibrating diameter 0.25mm wire of a consists gauge wire
blocks. Before by in barrel end steel and anchored a steel enclosed Installing the gauges, the wire is tensioned to the required working frequency. This is carried out on a tensioning template by firstly fixing
The by block the other the to screw. a grub of means end end of wire one the is the and once pliers a pair of gradually stretched with wire end of block. the Located is to it is frequency end other clamped reached required is the the an electromagnet, of gauge centre at electrical which when activated by an
the frequency
to a recording device. To measure strain, the gauge is stuck on to the surf ace of the material or member by means of a suitable bonding agent such as plastic that a change in the
distances between the anchorages causes a corresponding strain change in the wire and a change In the frequency of vibration. vibration, The frequency of
Case study I
123
resolving power of the recording device an average accuracy of about 2 microstrains can be easily achieved with this type of gauge. The relevant
formula which is required to compute the strain in the wire is given 82 elsewhere . The various. types of acoustic vibrating wire gauges are shown in Plate 2.
length between two f Ned studs mounted onto a surface. As such, demec the the in tests have been ace of member surf whole where gauges most useful or structure is accessible. The capital outlay for a demec gauge is very
much less than that for a portable acoustic gauge recording equipment and the demec studs which are used as measuring points cost very little compared to acoustic gauges even when recoverable surface gauges are employed. The studs are positioned by means of a reference bar to the required in length the A to then length, change ace. surf glued gauge which -are between the two demec points corresponds to a strain and this is recorded on a calibrated greatly, dial gauge. When the ambient temperature fluctuates
the readings are referred to an invar bar and corrections have to can be
be made. With a 200mm gauge, an accuracy of about 4 microstrains obtained. A typical demec gauge is shown in Plate 3.
6.2.3 Laboratory
tests
Unlaxial compressive creep tests, shrinkage measurements, strength and modulus tests were undertaken to establish the basic and time dependent
124
20
so
0.0
l'i'J'Lu
,-
Hl-uubtic
ViLjI,Lifig
vvirc
'Lldill
qujy2)
Case study 1
125
properties of the concrete. In addition, measurement of the thermal expansion of the concrete was also carried out. The tests were conducted in laboratory conditions with the relative was 200C. Strain
measurements on the creep and shrinkage specimens were taken at regular time intervals. Furthermore, specified periods. strength and modulus tests were carried out at
The determination
of the static
modulus of elasticity
in a compression testing machine. Samples of concrete prisms 61n. (150mm) by 121n. (300mm) long were loaded up incrementally ultimate to about a third of the
to the f aces of each test prism, and strain readings taken by means of an extensometer. The readings were taken at each stage of application of the
load, both during the loading and unloading cycle. The test was carried out 83 in accordance with BS 1881:1970
6.2.3.2 Strength
Cubes and cylinders were tested for compressive and tensile strength, respectively. The compressive strength of the concrete used in this study,
was determined from 6in. (150mm) cubes. Two sets of cubes were prepared from the batch of concrete used to cast the bases and columns of the spill-through bridge abutments and test beams. The first set comprising of
three cubes each from the base slabs and site columns, was tested when the concrete was about 4 months old. After almost a year, the second set of a
cubes was tested. Cubes from the beams were tested at 28 days and after
Case study 1
126
kept the testing, to cylinders were 15 Prior cubes and period of months. tbgether with the shrinkage and creep specimens. Before the cubes were crushed, they were Immersed in water and then tested while they were still wet. According to BS 1881:197084 the load was
applied at the rate of between 0.2 to 0.4 N/mM2 per second until strength no greater
was then
on samples from
at 28 days.
test In order to
three A the of the tensile consisting set concrete. strength of ascertain for (150mm) long, by (150mm) in diameter 61n. prepared 61n. were cylinders, each abutment base, column and test beam. The test was conducted according 84 when the concrete was about a year old. to BS 1881:1970
Measurement of thermal expanslon of the concrete was conducted by heat cycling two unpainted shrinkage specimens in a water bath. Demec studs were to be to faces four readings strain enable of each specimen attached on all taken by means of a demec gauge. Temperature conventional thermometer of the bath was taken with a
temperature coil.
The temperature of the bath was initially raised In five stages until
it reached a maximum of 400C. At each stage readings were recorded after sufficient equilibrium. temperature time was allowed for the concrete to attain thermal
This procedure was repeated as the bath cooled down to the of the water in the laboratory.
Case study I
127
6.2.3.4 Shrinkage
Strains on the unloaded small specimen were measured on all four faces using the demec gauge. Datums were established on the day after casting when the moulds were stripped. The readings were taken initially week and less frequently as the deformations decreased. twice a
These specimens were subjected to shrinkage as well as thermal contraction temperature. and expansion, as a result of variations in the ambient
deducted from the measured strains to yield the free shrinkage strain of the specimen.
6.2.3.5 Creep
Besides providing the basic data on the creep properties of the concrete, primarily control creep tests on the small specimens and test beams were intended to investigate and assess the long-term due to the action of the backfill movements and
deformations
stage on site was completed. The small specimens were loaded up in pairs in the creep rigs shown in Plate 4. They were subjected to a sustained unlaxial compressive stress the to the in root of at strain which resulted a value of strain equivalent the site columns. Due to the orientation of the rigs, only two surface each specimen.
Pressure was applied at the base of the creep rigs, the reaction being provided by steel tension rods restraining the top reaction plate and base. with a
rubber membrane clamped in between. The top half of the base was a hollow
. -I u
CD 0. in
Case study 1
129
circular
a tapering cylindrical
through a small diameter hole in the lower plate, which caused the rubber load this In the deform manner, was to piston upwards. and moved membrane applied to the specimens bearing against the piston. Fluid pressure was maintained by means of an accumulator, up any fluctuations which took
hydraulic the leakage of the to small often unavoidable specimen and also fluid. A stabillsed hydraulic loading system such as this offers the
time. the The rigs same at rigs of creep a number advantage of supporting were connected up to one accumulator and a Bourdon gauge, which registered pump delivered
the applied fluid pressure. A manually operated injection the necessary oil pressure into the system.
the In dependent time requires concrete strains The computation of instant the instantaneous of application the strain at elastic value of load. As a consequence, the modulus of elasticity, of
known. In this study, static modulus tests were carried out at different times in order to arrive at a relationship defining the development of the of painting on E(t'), the
bitumen. beams test from the coated were with modulus prisms 5 The relationships based on the ACI Committee 209 recommendations 85 that proposed by Bazant were used. The growth of E(t1) according to the formula proposed by Bazant Is substantially higher than the ACI relationship. The reason Is that Bazant's
and
Case study I
130
formula content,
is intended for concrete specimens remaining at constant water that is, sealed specimens, while the ACI expression is applicable case, the hydration process is In
rapid and ceases quickly to form products of a poorer physical structure comparison to concrete which is allowed to hydrate slowly. This accounts for the higher values of E(t') for the test beams compared to the other results, as shown in Figure 6.1. In effect, the bitumen helps to reduce
moisture loss from the concrete, thus slowing down the drying process. As can be observed from both Figures 6.1 and 6.2, there Is scatter in the values of E(tI) which suggests variation in the quality of concrete.
This is to be expected from concrete cast on site where consistency is difficult to achieve. Generally, the E(t') values for the A-side specimens
are higher than the E-side specimens and this is possibly due to the ambient temperature during placing. The A-sIde abutment was cast about two
months earlier than E-side abutment during the peak winter period. The lower temperature during placing and setting of concrete on the A-side rate of hydration which resulted in a denser
Bazant also proposed an extension85 to his original formulation, covered concrete subjected to normal as well as partial
which
or retarded drying.
The curves derived from his method, as shown in Figures 6.1 and 6.2, are based on the modification ACI curve. and are much closer to the experimental data than
6.3.2 Strength
The concrete cube strengths obtained from the tests were well above the specified characteristic value and are shown in Table 6.2. With the
Case study I
131
0 0
0 000
ACI 209 drying) RP (Retarded BP (Normal drying) 0 D. 0 Test beams Base E Column E2 Column E3
with time
40
30
C
0 0
to
0.
20
10
* * 0 1 37 14 28 Age of cnncrete 90 ( days ) 180 36n loon ACI 209 OP (Normal drying) Base A Colim A2 Column A3
with time
Case study I
132
showed an increase with age. The 28-day strength values for the base slabs and abutment columns were taken from the Contractors could have been caused by using different values of the modulus of elasticity, and the discrepancy in the
compressive strength in comparison to the E-side and this re-affirms 86 role of the placing temperature in influencing the strength
Base slab Lo ca tl on A E A2 A3
Column E2 E3
Test beam
Note:
for base slabs values of 28 days strength are those obtained by the Contractors
and sIte
columns
Table 6.3. Evidence seems to suggest that these values overestimate actual direct tensile strength of the concrete. The difference due to the method of testing. A review of different out by Hannant 87
is partly
direct tensile strength ranges from 0.95 to 2.4. However the variation included the effects of different different size and shape of specimens as well as value based on
experimental
test overestimates by up to 70% the tensile strength of concretes made with aggregates of a maximum size of 38mm. In order to assess cracking of the
Case study I
133
Locatlpn
Base slab
A E A2 A3
Column
E2 E3
Test beam
Streng Z
m2) (N/MM2)
4.05(364)
3.16(463)
3.41(309)
3.20(315)
3.65(436)
3.17(438)
3.75(446)
2.70(364)
2.11(463) I
2.27(309)
2.13(315)
2.43(436)
2.11(438)
2.50(446)
Note:
*Based on an overestImate of 50% of the measured values. Values In brackets are age In days.
the by those tension shrinkage, restrained caused such as concrete under direct tensile strength must be known and in this study it can only be estimated from the value of splitting strength. The maximum size of
that be it thus, 20mm an assumed can only and aggregate used was overestimation tests. from the The derived in the 50% values occurred of
6.3.3 Coefficient
of thermal expansion
The coefficient
the temperature time the its hygral the of state at mix and composition of 89 fact the that the from the influence The arise proportions of mix change . two main constituents of concrete, cement paste and aggregate, have and the coefficient for concrete is a
The influence of the moisture condition applies to the paste component and is due to the fact that the thermal coefficient is made up of two
Case study I
134
thermal coefficient
and swelling
arises from a decrease in the capillary tension of 90 Is No swelling water held by the paste with an increase in temperature .
possible, however when the concrete is dry, that Is, when It contains no water, or when it is saturated. It follows that at these two extremes the of thermal expansion is lower than when the paste is partially
coefficient saturated.
shrinkage specimens as they were kept under conditions of varying ambient temperature. Furthermore, it was vital to distinguish the true shrinkage
strains and establish the correct datums in the test beams and site columns during the initial significant period after concreting. Thermal effects become more'
The changes in surface strain of the concrete specimens as the temperature of the water bath was raised and lowered, were plotted against These are shown in Figure 6.3. The strain
readings represent the average values taken on the four faces of each of the specimens. The coefficient of thermal expansion is given by the slope
of the line and an average value of 10.7 microstra! rVOC was obtained for concrete in a saturated condition. However, this value has to be corrected to give the actual value for
the unpainted and painted specimens, test beams and site columns. According to Browne9l, suitable correction factors should be added to the saturated expansion for different degrees of
saturation and environmental conditions. An additive factor of 1.3 was proposed for mass concrete pours such as in thick walls, beams and columns particularly where the surface is sealed. Thus, a corrected value of 12 microstrains/OC was arrived at for the painted specimens, test beams and site columns.
Case study I
135
50
40
"i-
30
wo
..
0'. 0-'** . .
...
&.
AY, .
20
13 co p
10
c Q u
40
., *...
I.,
Ar
30
20
10
Specimen No. 2 0 Heating a Cooling Mean line -0IIIII0 50 100 150 Measured strains Figure 6.3 Determination of coefficient 200 ( X10-6 250 300
Case study I
136
The measured free shrinkage strains from both the painted and unpainted column specimens were plotted as a function of log-time and were
then compared with the predicted shrinkage curve from the CEB-Fjp3,4 and 5 7 ACI Codes, and the procedure proposed by Bazant and Panula The predicted . curves are for drying at 60% relative as 200C. humidity and room temperature taken
Readings from the unpainted concrete specimens represent shrinkage under normal drying conditions. A comparison of the predicted curves with experimental data from the unpainted specimens as depicted in Figure 6.4
and 6.5, shows the accuracy of these methods in assessing the development of shrinkage with time and the ultimate Ignoring the vertical value of shrinkage strains.
higher rate of shrinkage than any of the other methods. The shape of the curve which ultimately function determines the slope, is defined by the time is
adopted by the methods. in the ACI procedure, the time function as opposed to Bazant-Panula's
in the form of a simple hyperbolic function (BP) square-root hyperbolic function. function
another. Judging from the plots, the time function adopted by Bazant and the CEB-FIP Codes seem to define the development of the experimental strains closely and thus would be more appropriate of data. for this particular set
Case study I
137
E3
(n 0
_r_ 4-j
cu
0 0 0
4-) C.)
E c 0
cl
00
00
CP
%0 m ab
%0 U., cc '0
1 .- 0
co
co co o a) (3) -c
CP
00 00
CD 09. 0u 01) 00 o ci
%I
so 41 we me so 10
0 00
'o 0
4 cu
ca Y
0 ON
LI) c
'0
C: a) (U (12 16
3:
co C*4
0 . c 0
co CL
E 0 u
0 co rrCl% 41
c
ON (L LL I Ch
0
ho
C)
(L)
(L 1
cu
aco
orl co
4-) c
C',
c =)
0-
. -1 (U
EL
co
Lu u
C13 "
Ld U
L) Cc
n cr
00
o
a
-. 4
(14 93 cr
0 WN 0 -Z CN
(9-01 )
SUTOIIS
0158>IUTIL4s clOIJ
Case study I
138
I
I
CD c CD
T]
0
c 0
'0
00 00
00 so 130 130 00 00 me do do
Co
CD
\C
CO
c "c3 C
"\
\
a lb C: )
m c
so ob
C) .
00
as
1
CD
0 rA
-1
(L) 4-)
ti-
b-4 t LLJ L)
:3 n LLi
Co CL
ij L)
C'4 w
1-4
0 C: )
1410 Ln
CD
-7) C:
C: )
r-
888 n
CD
(14
(9- OIX )
SUTB4S
aaIJ a5L'>IUTILIS
Case study I
139
A matter
in the predicted
largely due to the base shrinkage values which the various methods assumed. As mentioned in Chapter 2, the prediction the influence of the concrete constituents shrinkage but ignore the effect of different methods take into consideration in evaluating the base value of types of aggregate. However,
the BP method is the exception in evaluating the base value not only in terms of the mix proportion, but also the concrete strength. The influence of the aggregate type is therefore included in the prediction through the
In the drying, stored were specimens when period of varied significantly from room
Into drying the by an equivalent period expressing methods time the is, that by scale. change of a considering period,
Since the process of hydration is dependent on temperature, It should 33 34 35 thus obey the rate process theory 1 . Accordingly, Bazant proposed that the correction factor be in the form,
kT TT0R
where L is the activation
exp [
energy
Uh
of
(1_I To
)]
(6.1)
hydration
is gas constant
Case study I
140
T, To
are temperatures in Kelvin (absolute temperature) reference temperature, To, taken as 250C.
and the
drying time Is
tkTt
(6.2)
This proposed dependence has been adopted and applying the correction to the predicted curve of the BP method, a much better f It to the experimental data was obtained as illustrated in Figures 6.6 and 6.7.
In addition to being subjected to varying temperature, were also subjected to varying relative Figures 6.6 and 6.7, after the predicted effects of varying temperature, humidity.
the specimens
the effect
on shrinkage
of the specimens is probably not that significant. During the initial stages of drying, shrinkage Is greatly influenced In
by the relative humidity of the ambient air. This is the dominant factor
determining moisture movements and consequently the magnitude and rate of shrinkage strains. As drying progresses, the moisture content of the concrete drops appreciably and the pore humidity approaches an equilibrium
value. At this stage, the shrinkage process is mainly governed by hydration rather than moisture loss to the ambient air, and the effect humidity activated on hydration is insignificant process. of ambient
Case study I
141
(14
C-, 0
64 4A m (D
.01 10
41
01
I
0 n co
I in ca 0
-4-) (L)
E CD cu
Co
-J.
00 A C2 co r-)
4j U
N c
10 cp
%ID (f) c
\ cp
\00
\00
C:) Co 4 CE) 0 00
'0 C
.,
(D c:n
CCYCh
0 0
10 cc
C',
ow
0\
C3
cc CN
c2
c:
:1
-0-
.0 c: 0
ci
co (U
E 0
c
:IB
:Ic0
c
', -C. -- -.
c U 10 C,
Z0
Ia ,, -. r) 113 0
13
4k
\I
10
p qu IX
IM4
(D I--
11
11
cr
cz
l
Ix
1 00
co
-4
? %0
88
CD
Case study I
142
CD CD
UN
CO c r*)
CD
Co
18
13 c
CD 0n
120 00 ab. CD
CID -0 c:
Co CO c i .A. Co
(L) CY) c
t 1
Co
-4J
D00
00
0
-4
C:) (D c2. CY, c
CC
*t
tu 2E
c2
\ cl\
%
0
ul
c a
cc CY) cr
CX
E 0 U r*-0 (0)
w9%1 (0 gar" %
:1c:
CL c
(D
-, c
'0
0 0
to Et
90
14 t; Co 0
00 ZZ
LL
LLJ
(N LLJ
C:)
C:)
C: ) enC,
(9-OTX
Case study I
143
This probably accounts for the observation in the experimental 60% drying to relative constant of value at a seems exhibit which that is, the ambient humidity during the initial
data humidity,
outdoor shed. The measured shrinkage strains did not respond significantly to the humidity change when moved into the laboratory humidity and neither did the
of varying temperature
This however, is not the case in thick concrete members or mass concrete, for the the quite an appreciable remains wet member core of since reaches an equilibrium value only after a long
be humidity influence the as could drying Thus, ambient of out. period of significant as temperature itself. There is very little of alternating experimental
that than in higher obtained at a shrinkage a results within the given limits.
controlled
A number of the shrinkage specimens in this study were painted with black bituminous paint and as such presented some problems regarding the in the Codes free The the methods strains. shrinkage proper evaluation of of Practice such as the CEB-FIP and ACI, are for normal conditions of
drying and thus not applicable to painted concrete. However, the method include the effects be to Panula by Bazant could modified and proposed painting. The predicted curves for the painted specimens shown in Figures 6.6 and 6.7 were obtained using the modified fotm of the BP method. The paint of
Case study I
144
of the
is the retarded. Plhlajavaara93 Consequently process shrinkage concrete. carried out a study on the effects of painting on drying and shrinkage of
down the drying that based slows painting concluded materials and cement process and reduces shrinkage significantly. 95 94 and Goldbeck and reported by Davis According to the formulation concrete is analogous to diffusion Similar observations were made
in the BP method, the drying process in and thus the parameter which determines
the rate of the shrinkage process is very much dependent on the moisture diffusivity empirically diffusivity the The moisture concrete. of and is a function of the water-cement is evaluated ratio and cement content the rate of in a much
to concrete compared as shrinkage of rate smaller magnitude and slower which is unpainted. This effect can be satisfactorily parameter. which could determine due to painting. modelled by using a
However for concrete specimens coated with a thin layer of epoxy resin, Hilsdorf96 found that the diffusivity factor of 5. Furthermore, parameter was reduced by a
factor of between 0.3 to 0.5 could be easily achieved. In this particular bituminous two paint. the coats of viscous painted with study, concrete was A reduction of the moisture diffusivity justifiable by an order of magnitude seemed data, as
shown in Figures 6.6 and 6.7. It should be noted that the same ultimate value of shrinkage strains
for by both the be However, and specimens. unpainted painted reached would
Case study I
145
the painted specimens this value will only be attained after a longer period of drying in comparison to the unpainted specimen.
The values of creep function obtained from the painted specimens were plotted the log-time various prediction compared with and scale on a are for normal drying at constant humidity in
of 200C.
drying creep
the 1978 the CEB-FIP to the gives method others while values compared lowest values, as shown In Figures 6.8 and 6.9. All the curves exhibit tendency to reach an ultimate value of drying creep. a
in to be 1970 CEB-FIP from the 209 ACI seem The predicted curves and that there f the despite exists no act other close agreement with each correlation between the correction factors adopted by them. For short
the for both the the load, methods creep curves of shape periods under closely resembles the shape of the BP curv'e. However, for periods under load over six months, the deviation Is apparent with the ACI and CEB-FIP curves reaching the ultimate curve. These two simplified experimental the than BP in time a much shorter value fit to the best the also give methods
months. For longer periods of time, these predicted curves tend to an ultimate tendency Increase to the exhibit a values measured value, while methods for creep
Case study I
146
0 Z
G)
E c: 0 ci (U 9-. ci c: c
out
A on
a
Co
to
ON
cu
tn t c2
CC) 04
-0
r0 (n 4Z r-I
E 0
co
(1) CU
ch a1-4 taI M
Ch
I " LAI CO
r
\ , ,
CL
0 C14
l% C14 Pd
w u
L) cx
CL m
cli
Co
%0
-e
CD N
Case study I
147
(A '0 0 C. Aj
cu
4-) -2 0
E C: 0
'0
Mel
so A so we so
1-1 (U W
4-J 4J ccN 4 -i
co
10
n cr c -4 '0 ca
C:) co 4
0 cc
V)
U)
>t M 0
co C14
3: 4-3 CU m
c 0 V) cu
c CU CU
E 0 u cu LL
V10
a-
t0\
011
r0\
co ca " LLcl (I
nw L)
AI
No
ww
co
III
a0111
-4
C, 4 4
C:) 1--4
(0
\o
CD -e
0 (14
Case study I
148
under normal drying conditions gave a reasonable and satisfactory assessment of creep of the painted specimens.
As in the case of shrinkage, the drying process occurring simultaneously with creep in the painted specimens Is affected by painting
and consequently the drying creep curve has to be adjusted to take this
into account.
The approach adopted in the BP procedure to derive the drying creep curve lends itself readily to be modified for the effects of painting. The
actual basic creep curve is unaltered but the other two components which make up the drying creep curve, that is, the drying and predried terms are affected in a manner similar to shrinkage. By appropriately factor, reducing the
moisture diffusivity
the predicted creep curve for the painted in Figures 6.10 and 6.11, data. The variation made
the curves give a much better fit to the experimental in temperature in the laboratory
in ambient temperature
slight, there was however a seasonal variation was found that the relative humidity
a condition of high external humidity prevailed and the laboratory to operate towards the upper limit, lower limit. while in winter
Case study I
149
C14 10
0
W
N (0 Co '0
0c c
r_ Co Co
E co D C: ca
-4-3 Co
C: )
c: -f CD c:
-ca
C: )
(DN
4-
4)
F-4 (1)
(n
(L)
-0
A %
Von
4-J
Co
. -:r
c 0 ul
s' c
co V)
I I
OL
0) 4-)
cc
W CD C -4-J
-H (C (I
--f ct) cl
->,1 . T) 1-0
If% -4 %V L; C)
C. sm 1cl,
Q) cc k 0 0u
CD
0)
-'D1., "D
41 cu I.-
fu f-4
(D 'D W L; co
-4
Nm2
C) Aj C'j a)
in
cn ca
LL
40
-. 4
0000 0
00
co
(gww/N/
OIX 9-
uolqounj
daajo
Case study I
150
n
L) 0
in
If%
'0
'E cu
CD
c CID M Co
4-) ca
ca
0)
0 CC)
C .H 'a CO
0
CY) Co -Y C: .-
C: )
CU 4-) 4-
cu '0 .-
CO
cn >1 c cu ca CU C: )
cu m
Co
c 0
Co CL
(U LL.
m
-I
0 co
ER 00000
8W 4
(zww/N/ 9-
%0
04
OIX ) uoTlounj
daajj
Case study I
151
The effect
of variation in relative
humidity
Is that creep is
increased, only if the load is applied prior to drying out. Tests by Hansen 97 on mortar showed that exposure to relative humidity alternating
between 50-70% gives creep almost as large as at a constant relative humidity of 50% and much larger than at the average constant humidity 3 proposed that, in general, if variations in humidity under of
60%. Nevillel.
load are slow creep tends to a value which it would at the constant lower limit of relative hun-ddity. With rapid variations, creep tends to a value half way between the
humidity
If the load is applied after first drying out, which creep is considerably lower than when the equal to the
for the concrete specimens under test and as mentioned earlier the computations relative for the predicted curves were based on the upper limit of
humidity.
6.4 Summary
This case study has to a certain extent given insight into the effects of painting on the elastic, strength and thermal properties as well as shrinkage and creep characteristics of concrete. Furthermore, it has been
shown that by modifying the predic tion method proposed by Bazant and Panula, the fr ee shrinkage and cree p of painted concrete can be reasonably estimated. Presently the procedures for shrinkage and creep of concrete in lack the capability of predicting these time
dependent phenomena for conditions of drying other than normal such as that imposed by painting. In the next chapter, the BP procedure will be extended to evaluate the free shrinkage strains and creep functions in the test
Case study I
152
beams and site columns. A more extensive experimental concrete specimens of different mix proportions,
programme Involving
the validity
CHAPTER 7
CASE -STUDY 11
7.1 Introduction
Time dependent behaviour of concrete can cause considerable changes in the internal stress distribution in reinforced concrete structures. Lately, of
importance the have have become recognised and aware more engineers design in the of many structures and concrete shrinkage and creep of
have been into dependent time for taking these account phenomena provisions introduced into a number of design codes. of full scale
have been conducted in this country. However, the data available of various prediction and with
performance
field One study such small. current 75 The Surrey from team the University by of a research was undertaken . design methods is still comparatively structure under investigation consisted of the columns of two spill-through
the In bridge. four with conjunction continuous span abutments supporting a field study, creep tests on small specimens and two beams were also carried out under laboratory two Data of years a period over gathered conditions. are presented in this case study.
153
Case study 11
154
One of the primary objectives of the study is to assess the prediction and analytical methods described in the preceding chapters with data monitoring of the test beams and site columns. As method
concluded in the previous chapter, the Bazant and Panula prediction together with the proposed modification
shrinkage and creep properties of the concrete. test beams and site columns and by utilising presented in Chapter 4, the structural could be carried out.
7.2 Full-scale
work
A full-scale
investigation
acting on them. This forms part of a research programme under the auspices of the Department The first lateral of Transport into the design of spill-through abutments.
soil pressures and the internal strains in the abutment columns, events and moments of the abutments. of creep due to sustained lateral
and to relate them to the construction However in this study, only the effects pressures from the embankment fill discussed.
the
A3 London to Portsmouth road at the Wisley Interchange. carrying the A3 over the M25 is constructed
joint along the centre-line.
longitudinal
Case study II
155
four-span continuous voided slab with spill-through of the bridge. The principal details of the structure the elevation
in Figure 7.1. The deck has a skew angle of three degrees and columns and on free
is supported on fixed bearings over the intermediate sliding bearings at the abutments. The M25 motorway passes into cutting lowest level of the three-level
diamond interchange.
D are double the standard height to provide the necessary clearance. However, the spill-through cutting abutments A and E are positioned beyond the
zone and are founded just below previous existing ground level. overall height of 9m and
beams deep 3m supported capping of consist abutments The capping beams are
divided by a 20mm joint- at the bridge centre line to permit independent movements of the deck slabs, but the base slab is structurally continuous
across the full width of the bridge. The elevation of the A-side abutment in Figure 7.2 illustrates the transverse spacing of the columns, while
Figure 7.3 shows the column heights and the base dimensions. The E-side columns are 800mm shorter in height at 4.5m and 4.6m, but the E-side capping beam is 270mm deeper than the A-side. The abutment columns were designed with a constant rectangular 1300mm deep and 800mm wide transversely. backfill, Before being buried in the section
ingress of water into the concrete. At the column root, the main vertical
Case study 11
156
75
Abutment
(?Q
2E0O1
P 9gs 399S
T 19117
T 1997 399s
Q P __
Y 9s
1 0 1863.
BEARING PLINTHS ,,
771
2003
4000
Loan
2000
2000
4000
7000
2* Sf 11'
75
Case study Il
157
I Vibrating wire strain gauges I Concrete strains I Is Vibrating wire strain gauges I Steel strains 1 41, Vibrating wire embedment earth pressure cells 0 0- Vibrating wire boundaryearth pressure cells a Pneumatic boundaryearth pressure cells
75 instrumentation
Case study 11
158
reinforcement
and
150mm bars faces, 32mm H-Y. faces. at the On positioned were side rear centres to control transverse bending moments. Link reinforcement at this
level consisted of 12mm stirrups at 150mm centres. At a height of 3. Om front the bars to 32mm the and on the base main steel reduced slab, above rear faces and 20mm bars at the sides.
7.2.3 Instrumentation
for the bridge selected from were Two columns each abutment of long-term interior two the E2 A2, adjacent two and The columns edge observation. frame to determine action or whether A3, E3 were chosen columns were important. The strains in the columns were recorded by
vibrating
the base slab are included in Figure 7.3. Details of the vibrating gauges are given in Chapter 6. The gauges were located mid-way between the vertical
reinforcing
bars
to record the concrete strains at six positions around the column perimeter. They were supported and protected by a wire mesh between the
bars in the columns. In addition, four gauges were attached to the main 40mm diameter bars on the A-side columns to record the steel strains in the critical design section at the column root. Temperature coils were included
in selected gauges so that changes could be recorded at the top and bottom of each column.
Case study II
159
7.3 Experimental
work
Field measurements on the site columns were supplemented by a series of laboratory tests on small companion specimens and two reinforced test
beams. The 5m long test beams are scaled down versions of the site columns and were constructed in the laboratory using concrete and reinforcement work on the beams are given here,
in cross-section with a width of 230mm and and the steel consisted of 25,
32 and 40mm diameter H. Y. bars with 12mm stirrups. The percentages of longitudinal reinforcement in the sections and spacing, between the bars
were similar to the abutment columns. Details of the beams are illustrated in Figure 7.4. Concreting was carried out in the laboratiry using timber forms and
af ter 4 days the shutterings were removed. The beams were moist cured prior to painting them with bituminous paint sometime between the second and third week after concreting.
7.3.2 Instrunentation
wire gauges were fixed between the bars and clamped on both the tension and compression sides, using the external gauges were attached onto the to check the
Case study 11
160
ED;
2
4 41 co a 64
04
(n C; -1.4 'D
A
9L (n
ozlq
E co a. )
-4 Aj CD
-o .4-)
V)
E 4
CU 4-)
E0
8 0
U'
t.
(U w cc (n
CL) 0
LL.
9zl?
C e
04
-4
I112
Wg
Cos
Case study 11
161
vibrating
wire gauges used are shown in Plate 2. Dernec studs were also of the beams in order to obtain
mounted onto the surface at the mid-section the strain profile across the depth.
Before the beams were positioned and loaded up for the creep test, strain readings from the internal gauges were recorded periodically. These
readings are strains caused by shrinkage in the section provided that there was no large variation in ambient temperature. Unlike the small specimens,
from the thine they were stripped from of the laboratory were
Due to the size of the beams, it was realised that during concreting hydration of the cement would cause large temperature rises. As such a
complete strain history from the time of casting was necessary to define a proper datum for the shrinkage.
The beams were loaded up in bending in a steel frame which has a clear span of 5.25m and a clearance height of 1.64m. Constructed of UB and UC sections, the frame is stiffened with the addition of a lower beam member.
The whole arrangement is bolted down to the floor and the top beam is tied to the nearby wall to ensure stability. positioned longitudinally A pair of light steel frameworks
on each- side of the frame, carry the dial gauges The loading frame together with the beams in
in Plate 5.
16: 1
Now
Case study Il
163
The test beams were placed one on top of the other, the upper beam being inverted. Loads were applied at the third points by means of 20t hydraulic jacks which rested on hydraulic load measuring capsules. The capsule which registered the applied load was essentially a shallow cylinder with its platen bonded to the walls by a flexible underneath the platen was filled platen virtually eccentricity joint. The space
The purpose of the test is to establish and to make a realistic estimate of the effects of creep in bending, under controlled intended to investigate loading
the abutments. Thus, the creep tests in the laboratory after the final backfilling effect
of this major load case on the structure. The beams were subjected to strain levels similar to that recorded at
the root of the abutment columns after the capping beams were backfilled. Due to the nature of the construction in three increments. The first applying the next increment. sequence, the load had to be applied
Case study Il
164
test beams and site columns presented herein are based on the work described in the preceding chapter. These basic parameters need to be evaluated before determining the shrinkage and effects reinforced sections. Both experimental of creep in the
results and discussed. Although creep and shrinkage are the prevailing assumption is that they can be treated the analysis. This approximation 98 . appears
The size and shape of a member definitely moisture moves to and from the concrete,
and therefore
shrinkage991100. Moreover in a large mass, there' exists greater variations of moisture content from point to point as compared to a small specimen and these variations will produce a non-uniform mass. However, for practical as a homogeneous material approach has having different volume/surface state of shrinkage within the
purposes it is often useful to treat concrete and avoid the complications. This form of
and these have been included into the Codes of Practice. Applying the modified version of the BP prediction that the size dependence according to the formulation unpainted and painted concrete, then the resulting method and assuming
for the test beams are as shown in Figure 7.5. As In the case of the small
Case study 11
165
0 c
I co 1-4 4-)
0) cu -Y c c V) u
1-4 4(U
u (D w IL
Go
UN
(U
rl
-4
Case study Il
166
diffuslvlty by derived these a reduced moisture using are curves specimens, factor to account for the effect of the bitumen paint. The values of free to account for the delay
in painting the beams. This was achieved by transposing the curve attributed to the painted beam onto the unpainted curve at the times the
beams were painted. The free shrinkage curves for the site columns were also derived in the buried, for the beams. Since the columns were procedure accordance with the ambient conditions were different average ambient temperature column root and the relative value in the laboratory. compared to the laboratory. An
the from 100C at measurements obtained was of humidity was assumed to be higher than the it was expected that
for free in The be shrinkage curve magnitude. small shrinkage strains would the site columns are shown in Figure 7.5 along with the predicted curves for the small specimens and test beams to indicate the relative of the strains. magnitude
The shape and size of a concrete member will also influence the
magnitude of creep. However only a portion of the creep referred to as drying creep is affected while basic creep itself remains Independent of of these factors lies
significance
shrinkage, the influence of size and shape of a concrete element Is commonly represented by the volume/surface and for practical section. ratio or the effective thickness
across the
Case study 11
167
at simultaneous drying Is
expressed in terms of basic creep and two additional drying creep terms. The size dependence for shrinkage is assumed to hold true for these two drying creep components. Thus the creep functions for large concrete elements such as the test beams and site columns, can be readily evaluated and by similarly earlier reducing the moisture diffusivity factor as carried out
relieved by creep. Thus, in order to compute the shrinkage strains and stresses under the restraining effect of the reinforcing bars in the beams
for the test beams and site columns as shown in Figure 7.6 are derived for both the case of applied loading and restrained shrinkage.
Once the free shrinkage strains and creep functions were established, the shrinkage strains in the test beams and site columns can be calculated. The method of Elvery and Shafi 103 was utillsed. The analysis was developed
bond between the concrete and steel, and uniform shrinkage of the strains at the
throughout the member. By considering compatability levels of the reinforcement and equilibrium
strains and stresses in the steel and concrete caused by shrinkage can be derived. The modifying effect effective of creep Is taken Into account by using an reinforced section such
as in a column, the resultant shrinkage force acts at the centre and causes a direct stress to act on the section. This simplifies the analysis and the
Case study 11
166
U) > m cc 'D
PIN rCN m
'0 (U
c
c 0 u
8
co
u m '0 ct) 0 c
70 ca
n (. (U V)
0 Co-
ul c 0
C Ch
cu 4-) 4m W >% co
u c qcl (1) (U
C) co C%4 "0 cu
.1-A
(n
CD .0
(U
-C) 7: ) CD CL '0
4 c 4
co
0)
in
gn 9
0
c: -H
0 U c
4m N, Ln NN (a (D
cl
CD cm (a .Y c
M -Y c
a) a)
41
fA (U
cl 03 k
-4 -1
-q rl -Y (a
C4 m ctl C13
U-.
is
I II
I' II
(n
III
CN
I
-4 0C0 Q
Ln
-4
0
daajj
U-%
(zLDW/N/9-OTX ) uoTqounj
Case study II
169
the the across mean represent value stress and strain computed shrinkage
section.
The initial contraction in concrete is the resultant of thermal 104 and depending on the thickness of
the concrete member the cooling period can extend from a few hours to in the implies that which days. measurements This shrinkage early several thermal effects in the disregarded case of especially and are erroneous are and should not be profile across
large concrete members the error can be quite significant overlooked. Furthermore in large members, the temperature
the section varies and causes one part of the section to restrain the movement of another part of the same section. The coefficient expansion for steel is of the order of 10 microstraln/OC sufficiently differential temperature and is of thermal
in to the that any problems to cause not concrete of close thermal movements between the two materials over normal ranges. was
During casting of the test beams and site columns, the formwork removed while the surface of the concrete was relatively the surface zone to cool and contract interior.
Consequently, the surface zones were subjected to tensile to compressive stresses. At this early
the to have the overcome enough strength acquired concrete might not stage tensile stresses, thus making the surface zone susceptible to cracking. As the core cooled, a reversal in the state of affairs occurred and any
In the developed have surface earlier might microcracks or cracks which 105 zones would then tend to close up The measured strains on the steel bars in the test beams and at
the root of site columns are shown in Figures 7.7 to 7.10. A distinguishing feature exhibited in the plots is the sudden rise In compressive strains few days after concreting followed by a gradual Increase.
Case study 11
170
cl,
CL
(D 0)
cl
ra
C: )
0 uj c
cn
Q)
F
-4
'0 CL
00
. 4
00
In
(I
c c -0 (n cu (L)
000
000
E c:
00 0.0 .a0
00
-0
oo
00
0
0 00
000
00
\0
00r
.2
-D m
99R8
()TX ) suTells
"0 0
00
Co
9068>1uTlqs
Case study 11
171
-V
C, 4 11032 R ct
00
0
m m as ION
ION.
00
00
am
r
loo00
mI.
00
ON ION I
c X
mo low
130 00
!F cli
p-I
-0
Do
on
In (V
Co
00
00
00
On
(U w
E
001 1301 00 on
40 0 0
00 00
ON ON 00
on ON
00130 00 111
so 2*
0:
2
0
Co
DO Do
LL
ma
ED
0
00
8 -4
( 9-
Go
()TX ) SuTBJIS a6e>juTjt4s
Case study II
172
%0 L) tv 40.4 (U Ix
441 c 0
co
OL 0 41
4 CII cl
ca (D -4 L). D
-4 -4
E %0
CN 04
c
-4 Au :3 0, m
c (1 r= -4 co -4 (L) 4
C-) .0 -. q-
E D
kn -I
c
N
cu
Ul (U (U
04) ZID
c 0
(D
0 CL L) 00) VVU
1
-
co Y
0
tu M 2
.. 4
L)
(D Z:
C164 cis. co
Lli (h
L) CL
CD -I
LL
. 1 L) U
-4 . 4-
0)
-4 co
041 L) U)
L--Omdo
v II
.
C: )
Uai
Ii
II
BJJS OTX ) UT
1,0
9-
Case study II
173
Ar
CD CJ a
$4 (a cu ct:
W 0 19 4
c 0 $.4
1
CL 0
I
j
c
0 u c C14 m
Cyb
CL E ci (0 -4 um L --IMC Ic 4 E -f CL CL (0 -4
tu w -4
-4 Q(a
& S-4
tu 4.3
(U 0 :)D> CO U
(U CY) M (I
k a) -4 La
ci. DO
ku -Y 0 (U Z'v
cu cu
c 0 Im 0) ca
0 U)'D L-430e (a -Y (U -Y Ln 40
4 CL cl to L)
$4 :3 0 CL
4Lli 0
(U
L)
ci
LL
-9
0 L)
Ra00
RR
Ckuoo 9OTX ) uTei4s
.t. .0 R
Case study II
174
As mentioned earlier, this sharp increase is caused by both thermal cooling and drying shrinkage of the concrete. initial contraction There is no means of resolving the
into its two components. However, since the test beams after stripping, initial
and site columns were kept moist immediately shrinkage would be reduced or even eliminated. total initial contraction
is caused solely by the cooling of the concrete. would have to be added to the value of the
contraction
shrinkage strains derived from the analysis to yield the theoretical curves. These curves are also included in Figures 7.7 to 7.10. The difference between free shrinkage and the measured shrinkage
strain is termed the restrained shrinkage strain. it is this restrained movement which induces stresses in the concrete and the!e stresses can cause cracking. The computed concrete tensile stresses adjacent to the reinforcing bars in the test beams and site columns are shown In factor influencing section
the development of shrinkage strains and stresses in a reinforced is the magnitude of restraint. With increasing stiffness of restraint
there is a decrease In
the measured shrinkage strains. Consequently, the restrained movement Is greater which accounts for a higher concrete tensile stress. The estimated concrete tensile strength of the beams as given in Table 6.3 is 2.5 WMM2 at 446 days. After a period of a year, the concrete
stress adjacent to the 40mm diameter bars in one of the beams f ar exceeded the estimated tensile strength and would have caused cracking within that zone. Another factor which influences the development of shrinkage stresses is the magnitude of free shrinkage itself. This is apparent by comparing
the mean concrete tensile stresses in the columns and the concrete tensile stresses adjacent to the reinforcement Figure 7.11. in the beams as shown in
Case study II
175
(1) Ol M -Y c
"0 (L) c m
lu
64
64
w .0 .0
co
'0 cu
Aj c Q)
ca u
E
0 4CL) 0
c (D C-) cu 0
(U V) (D V) cu (n c cu
CL) a) L) c 0 u -.4
LL
C3
cl
69%
1"1
U%
r1l
0 C;
Case study 11
176
The free shrinkage strains in the columns were minimal and as a result the concrete tensile stresses arising from restrained shrinkage were relatively small and would not be sufficient for the formation of cracks.
However these computed stresses and strains represent the mean values the to the the reinforcing adjacent across values section and higher than the mean due to the greater amount of bars would be
restraint.
The viscoelastic aging method developed in Chapter 4 was used to from the applied dependent the time resulting strains stresses and compute loads on the test beams and the backfilling stage on the site columns. The
method as previously discussed could take into account creep under variable moment histories and makes it more viable and suited for the loading conditions encountered in both cases. One of the major assumptions made in the analysis was that the is the tension thus below the all and concrete neutral axis was cracked carried by the reinforcement. This conveniently confined the phenomenon of
creep within the compressive zone. in the case of the test beams under due in to the tensile the load, concrete stresses of sustained magnitude restrained shrinkage coupled with the application of the loads would have and
cracked the concrete in the areas adjacent to the tensile reinforcement consequently making conditions closer to the assumption in the analysis.
However in the site columns the magnitude of the tensile stresses due to restrained shrinkage were relatively strains from the backfilling small and together with the measured to crack would
Case study Il
177
loads applled
in three stages to resemble the build up of strains In the The applied moment
are shown in Figure 7.12 and these values of moment are utillsed of the creep strains and stresses. In order to obtain
in the computation
the total time dependent strains, creep and shrinkage are assumed to be additive. As such the time dependent strains in the compressive steel are derived by summing up both the predicted shrinkage and
reinforcement
be in a cracked state, shrinkage of the concrete should have little on the strains in the tensile reinforcement6l.
A comparison between the measured and computed total time dependent strains In the steel reinforcement of the test beams are shown In
Figures 7.13 and 7.14. The analysis based on a cracked section gave higher values of instantaneous strains which is more evident on the tension side as shown in the plots. Although there is an indication beams that concrete adjacent to the reinforcement in both the test
not necessarily imply that the tensile strength of the concrete can be totally ignored. There is some contribution from the concrete In tension the reinforcement In
tension. As such a cracked analysis could not be expected to yield accurate predictions of the instantaneous strains under such circumstances.
Nevertheless, it should be emphasized that such an analysis Is simpler to perform and can readily and easily be accommodated In the computation time dependent effects In structural of
of a non-zero concrete tensile cracking stress. moreover It Is shown In this study, that under service load conditions and where long term deformations are concerned the analysis results in predicted values which
Case study 11
178
so
6C z c (U E 40
36.24 49.41
0 E
'0
CLcl
20
0 350 4CO 450 500 Days after 80 32/32 test beam casting 550 600 6 5C
60
46.76
4 c
cu E
0 E 'D cu CL cl
40 34.36'
CE
20
21.19
0 350 400
Case study 11
179
400
300
200
100
Iz 0
x
ca w j Ln
100 9
200
60 I
120
180
240
300
360
420
480
540
60C
660
Time (days)
Case study 11
180
400
300
200
c cu
100
Iz
-I
0
Co
k 100 u
200 0 60 120 ISO 240 300 360 420 480 540 600 660
1Ime (days)
Case study 11
181
A comparison between measured and predicted the test beams are illustrated of superposition
on
Chapter 4, to predict the creep deflections The measured time dependent deflections theoretical confirms predictions
under varying applied moment. from the 25/40 beam along with the
from both methods are given in Table 7.1 and analysis are much closer to the
the effects
of creep under increasing moment. feature of the strain plots in Figures 7.13 and curve closely
Another significant
7.14 is that the corresponding slopes of the strain-time matched that of the experimental basic time dependent characteristics satisfactorily.
results which seems to suggest that the used in the analysis are modelled of relief on the strains in
apparent from the plot of the measured strains for the 32/32 test beam in Figure 7.14. The magnitudes of applied moment in this beam are less than the other beam and consequently the extent and degree of cracking in this in
beam is small, rendering the concrete below the neutral axis effective shrinkage. However, the apparent relief first two stages of loading. Following adjacent to the reinforcement is no longer significant
in strains occurred only during the a further increase in load, the and
The abutment columns which are subjected to sustained lateral from the backfill
would bend and creep with time when unequal pressures act act as would
on the front and back faces of the columns. They would effectively cantilevers resisting the resultant moment and the maximum effect
occur at the column root. Prior to backfilling the columns was only significant
Case study II
182
cu ca . -q -4 (U (U L) V) . -I
10 W J) m
'0 (L)
00
>>
u (f) ., f
V)
:i ioll
:I
E cu .0
V, cu
IIoo
E 0
0 0 Ln c0 u 0 0 cu
LO 0 0 0 0
E c
-ID c cu cl cu 0
a)
E
76 0
D
r cu
\
R
Ln .2 cli
LL
rb
(WLU) UOTIOaTJO(I
Case study II
183
Days after 1st loading increment ist increment 8 28 53 2no increment 68 89 119 3ro increment 126 133
13 5W
measured
Effective
modulus
(0 days) 1.58 1.64 1.65 (60 days) 2.08 2.29 2.34 (120 cays) 2.34 2.90
2.95
5.81 3.84
3.87
4.13
4.2 (1
4.24
196 ')91
1472
3.02 3.25
3.62
3.96 4.05
4.12
4.38 4.56
4.7 u
538
3.73
4.13
4.76
Case study Il
184
beam
of thermal
effects.
However,
as shown in
Figures 7.16 and 7.17, these effects term bending caused primarily
Unlike the case of the test beams, the bending moment acting on the
section at the level of the measured strains is unknown. By applying the
at estimating
and construction
operations involving
in the the the then strains measured compared with of moment were reinforcement. Thus the approach to the computation was slightly different
to that of the test beams. The moment histories presented in Figure 7.18 are estimates of the resultant bending moments and the corresponding creep strains inclusive of
shrinkage are shown in Figures 7.19 and 7.20. The comparison between measured and computed time dependent strains revealed certain with the results from the test beams. Overall the prediction in the compressive reinforcement similarities
of the strains
As observed and concluded before, the computed strains in the tensile reinforcement are much higher due to an overestimation of the instantaneous of three
is almost a factor
and such a discrepancy occurred as a result of assuming a fully cracked section when in actual fact the concrete was probably partially Initially resultant sufficiently substantiated design calculations of the lateral cracked.
bending moments acting at the root of the columns would be large to cause cracking of the concrete. by an independent computation This was further
measurements of earth pressures on the faces of the columns inferred that an analysis of creep effects
Case study II
185
CO
c7)
4.CD
%0
LL
R
-4
CR
. Ual
) UT814S OTX 9-
C3 in
R
*clwoo
Case study Il
186
CD
N C14 1 A
E 0 u 0 ) .b.. cl (D E
C, m
V) c 0 0)
CL)
.0 4C) cu u C: m u
cu
LL
SR
000
C14 C14
Lo
-4
8 -4
WN
) UTOIIS OTX 9dwo: )
-uaj
Case study Il
187
be justified.
base the to the is linked this relieving rotation such conditions and 75 cantilever action of the columns . Long term strain predictions were further complicated by the fact that the columns were bending sideways and
that cyclic
700
Column A21-2 Column A3-2
600
500
400
11 Z A
9-
3,90
CD
/I w
1) U
200
cp C
100
n
AIi till lilt) I 11 1 11111 (day,; ) I 1111o
11 IIA t
Figure 7.18 Bending moment histories for site columns after backfilling
capping beam
7.5 Summary
from both the small scale tests and measurements of shrinkage on the larger
Case study 11
188
R8
4
-4
. U21 9-
OTX ) UTBJIS
-dwoo
Case study 11
189
8
vl
0 C. Ln
-Y
co
C:
C) Q)
E 76
4-1 0
LL
RF8R0R8R Ual
9OTX ) UTeIIS
*dwoo
Case study 11
190
test beams and site columns. The study has also re-affirmed painting or waterproofing the concrete,
that by
stresses due to restrained shrinkage and minimising cracking as the concrete develops and attains
seem to produce high values of instantaneous strains and are well within the safe working stress of the steel. The strains and the corresponding stresses in the reinforcement the due to but time increase creep with would
to too be increase the cause any adverse small probably would magnitude of effects in the long term. However a matter which may be of concern is the of of
magnitude of tensile concrete stresses near the surface. The effect restraint would be high within this region due to the concentration and as the free shrinkage strain progressively
reinforcement
increases with
time, cracks would begin to develop and the durability would be affected.
of the structure
This study has also revealed that there exists two kinds of error in the prediction originates of shrinkage and creep effects in structural members. One
in the representation
and the other is due to the assumptions and approximate method of analysis. Comparing the predictions structures,
does not provide a way of separating the two kinds of error, no information on the accuracy of either the method of
or analysis is obtained. However, it should not be inferred, are of little value. For the degree
of accuracy of the analysis. For example, if the time dependent parameters are adequately represented, and yet a large discrepancy with the
Case study 11
191
In the then
CHAPTER 8
CONCLUSIONS
of this study was to ascertain the relevance and to prediction methods, for the in an
of the concrete
it should be remembered that prediction and all are based on the fitting tests. of the concrete or environmental prediction
methods are
of data obtained in
laboratory
controlled
conditions
level as demonstrated
small scale tests are invaluable and can either be used to the required basic time dependent parameters or to check
directly
on the accuracy of a chosen prediction involved in the prediction the concrete in a full-scale the difference be minimised.
of the free shrinkage or creep coefficient structure, after taking due consideration conditions,
In the environmental
There is close agreement between the measured free shrinkage strains and creep functions from the painted companion specimens and the
192
Conclusions
193
theoretical
prediction
version of the Bazant and Panula of the proposed modification of painting the concrete.
to the BP method, to account for the effect Furthermore, judgingJrom the configuration
time dependent strains in the test beams and abutment representation of the
basic time dependent parameters has been made using this method. It should be reallsed that according to the prediction, retarded due to painting but the same magnitude of ultimate strain would be attained painted concrete, shrinkage is shrinkage However in
shrinkage is not only retarded but also prolonged and as shrinkage value will be reached only after a
very long period of drying. From the design point of view, this could well prove to be a viable solution in delaying the formation especially in heavily reinforced most likely of shrinkage cracks The
early-age thermal
columns of the abutments occur at a slow and steady rate and that the bulk of these def6rmations the backfill are being caused by sustained lateral pressures from
acting on the columns. The axial loads on the columns such as of the columns and capping beam, have for they have a
not been considered in the analysis for creep effects relatively strain-time
small influence on the behaviour and this was observed in the plots. Cyclic creep caused by traffic loads on the bri.dge and have also been neglected In In the
seasonal variations
the analysis. Thus for a proper assessment of time dependent effects abutment columns all themfactors a better
Concluslons
194
lateral
earth
pressures.
which the member is subjected to a constant stress or moment. The solution for a variable loading case is arrived at by applying superposition and ignoring the stress or moment histories. However, the build up of the earth
pressures behind the columns was in stages and as such the reinforced section of the columns were subjected to variable moment histories. general viscoelastic aging analysis developed in this study catered for in the formulation. to the The
such a loading case and included the moment histories Thus, this method is more appropriate effective
shows that values from the viscoelastic measured time dependent deflections.
The charts which have evolved
from
this
study
are prepared
using the
viscoelastic
aging analysis and are based on concrete loaded at 28 days, and environmental conditions are specified. They
tension. The aim of the charts is to assist in the prediction deflections and stresses in the concrete
of creep which
A comparison of predicted
charts and CP 110 as given In Table 5.3, indicates that the difference magnitude is relatively large. However, the modified
which the method in CP 110 is based, Is known to overestimate of creep under a constant moment and when the concrete early age. Thus a lower value such as that calculated would produce a more reliable and closer estimation
Is loaded at an
Conclusions
195
deflections
in reinforced
concrete members.
8.2 Recommendations
Is by
far the most comprehensive compared to the recommendations and ACI Codes. Although slightly account for unusual conditions more involved, it can
of the CEB-FIP to of
be modified
of drying. Furthermore,
of different
aggregate which could be invaluable in instances where the concrete consists of less commonly used aggregates. Therefore, circumstances in such special
it is recommended that the BP method should be adopted. work on the test beams represents a stage nearer to in comparison to the small specimens. check on the as well as
method adopted to represent the shrinkage and creep properties, on the analysis for time dependent effects.
scale testing should always be considered in the experimental time dependent studies involving Whenever the deformation case of shrinkage in a reinforced actual structures. in the concrete is restrained,
readings in the concrete should always be carried out with caution. The strain gauges embedded in the concrete section and will not reflect record the overall deformation of the Thus
it is not surprising
when in actual fact the concrete is under tensile stress. The family different inserting of design charts developed here can be prepared for conditions by simply into
Conclusions
196
proportions
suggested that the charts should be used to minimise carrying out a creep analysis.
One of the assumptions made in analyses for time dependent effects developed in this study was that concrete in the tension zone has cracked and cannot sustain any tension. Although estimate of the creep deformation this would tend towards a safer
the magnitude of the error can be high especially when analysing a section which is only partially can cater for this. It is now common practice which acts as a waterproofing in construction to use bituminous paint and of cracked. Thus there is a need for an analysis which
foundations of buildings or bridges. Thus to assist in the prediction the free shrinkage strain and the creep coefficient
been painted, a more detailed and extensive research programme Is required. This would also allow for a better assessment of the correction to the BP prediction factor
method.
REFERENCES
of cement mortars and concrete through ASTM Proc. 11,1911, recommendations pp. 531-555.
of concrete structures,
Association, London.
4. CEB-FIP 1978, International construction London 5. ACI Commitee temperature 209, Subcom. Ii Prediction in concrete structures, of creep, shrinkage and in Designing for effects recommendations for the design and Association,
of concrete structures,
effects
prediction
Shrinkage, Materials
197
References
198
beam, ASCE
1968,
5764, Jan.
pp. 283-308.
10. Verbeck, G. J., Hardefied concrete - pore structure, ASTM Sp. Tech. Publ.
No. 169,1955,
11. Powers, T. C., The physical structure concrete, pp. 39. Portl.
12. Feldman, R. F. and Sereda, P. J., A model for hydrated Portland paste as deduced from sorption-length properties, pp. 509-519. 13. Neville, A. M., Creep of concrete
North-Holland, Amsterdam, 1970.
cement
Materials
and Structures,
: plain, reinforced
and prestressed,
14. Rusch, H., Jungwirth, D. and Hilsdorf, H. K., Creep and shrinkage : Their effect
York.
Sprlnger-Verlag,
New
15. England, G. L. and Illston, J. M., Methods of computIng stress In concrete from a history of measured strain, Review 60, Pt. 1, April Civil Engineering and Public Works
17. Neville, A. M., Ward, M. A. and Kwe!, G. C-S., Basic and drying creep of concrete, Materials and Structures, Vol. 2, No. 8,1969, pp. 117-124.
du beton, Paris,
19. Bazant, Z. P., Theory of creep and shrinkage In concrete structures precis of recent developments, Nemat-Nasser, Mechanics Today, Vol. 2, S.
:A
References
199
creep, Cement and Concrete Research, Vol. 14, No. 6,1984, 21. Ross,A. D., Concrete creep data, The Structural Aug. 1937, pp. 314-326. 22. Lorman, W.R., The theory of concrete 1940, pp. 1082-1102.
23. Thomas, F. G., A conceptlon of the creep of unreinforced
concrete
and an
estimation
of the limiting
1933, pp. 69-73. 24. McHenry, D., A new aspect of creep in concrete and its application design, Proceedings, ASTM, 43,1943, 25. Ulltskil, I. I., A method of computing of concrete for practical
pp. 174-180. Translation
to
Industrial
Research Organisation,
Research, Vol. 11, No. 33, Nov. 1960, pp. 143-150. 27. Wallo, E. M. and Kesler, C. E., Prediction concrete, Nov. 1968. 28. Shank,J. R., The mechanics of plastic Proceedings, 32,1936, 29. Bureau of Reclamation, concrete,
1953.
University
of Illinois
flow of concrete,
ACI,
U. S. Bureau of Reclarnatlon
30. Troxell, G. E., Raphael, J. M. and Davis, R. E., Long-time shrinkage tests of plain and reinforced pp. 1101. 31. Kasperklewicz, J., Some aspects of water dIffuslon Materlaux Et Constructions, Vol. 5, No. 28,1972, concrete,
creep and
References
200
A. H., Influence
on
water
diffusion
in nonsaturated
concrete,
Materlaux
Et Constructions,
34. Brophy, J. M., Rose, R. M. and Wulff, J., The thermodynamIcs Vol. II of The Structure I
New York,
and Properties
of Materials,
ed. by Wulff,
Wiley,
1964.
of matter,
36. Pirtz, D., Creep characteristics Report No. 65-2, Structural California, October 1968.
study
of mass concrete
Engineering Laboratory,
of creep
properties
of concrete,
Bureau
38. Freudenthal, A. M. and Roll, F., Creep and creep recovery of concrete under high compressive stress, ACI Journal, Proc., V. 54, No. 12, June 1958, pp. 1111-1142. 39. Davis, R. E. and Davls, H. E., Flow of concrete under the action of sustained loads, ACI Journal, Proc., V. 27, No. 7, Mar. 1931, pp. 837-901. 40. Harboe, E. M., Creep of concrete under high Intensity Laboratory
1956.
42. Ross, A. D., Creep of concrete under varlable stress, ACI Journal, Proc., Vol. 54,1957-58, pp. 739-758.
Ref erences
201
of the superposition
principle
in creep and
Beton-un-Stahlbetonbau,
pp. 261-269. 44. Bazant, Z. P., Prediction
Berlin-Wilmersdorf,
of concrete
creep
effects
using age-adjusted
effective
1972,
pp. 212-217. 45. Volterra, V., Theory of functionals integro-differential and of integral and
equations, Dover Publ., New York, 1959. of creep of concrete with T and AM Report
46. Vaishnav, R. N. and Kesler, C. E., Correlation its dynamic properties, N0.603, Oct. 1961. University
of Illinois,
47. Ross, A. D., Creep and shrinkage in plain, reinforced concrete :A general method of calculation,
and prestressed
1943, pp. 38-57. 48. Flugge, W., Mechanical models of plasticity interpretation
University,
of creep in concrete,
California, Tech. Rpt.
No. 8, Nov.
of the inelastic
deformation
of concrete
by
means of a mechanical model, Nature, London, 178, No. 4257,1956, pp. 278-279. 50. Gopalakrishnan, K. S., Neville, A. M. and Ghali, A., A hypothesis on mechanism of creep of concrete with reference compression, ACI Journal, Proc., 67,1970, 51. Riesz, F. and Nagy, B. Sz., Functional 52. Lockett, F. J., Nonlinear viscoelastic London, 1972. 53. Glanville, W.H., Studies in reinforced concrete III - Creep or flow of to multiaxial
pp. 29-35.
concrete under load, Building Res. Tech. Paper No. 12, Dept. of Scientific and Industrial Research, London, 1930.
References
202
Precontrainte,
Dunod,
55. Arutyunian, N. K., Some problems in the theory of creep, Pergamon Press,
New York, 1966.
57. Faber, O., Plastic yield, shrinkage and other problems of concrete and their effect on design, ICE Proc., London, Vol. 225, Pt. 1,1927, pp. 27.
58. Glanville, W.H. and Thomas, F. G., Studies in reinforced Further investigations
Tech. Paper No. 21, Building Res. Dept. of Scientific Research, London, 1939 59. Torroja, E. and Paez, A. Set and reinforced : their elasticity North-Holland, concrete,
and industrial
Building materials
60. Bazant, Z. P. and Najjar, L. J., Comparison of approximate concrete creep, ASCE Proc., J. of Struct. pp. 1851-1874. 61. Gesund,H., Shrinkage and creep influence reinforced on deflections
and moments of
May 1962, pp. 687-704. 62. Davles, R. D., Some experiments on the applicability of the principle of
superposition to the strains of concrete subjected to changes of stress with particular reference to prestressed concrete, pp. 161-172. Mag. of
63. Kimishima, H. and Kitahara, Y., Creep and creep recovery of mass concrete,
Tokyo,
Power Industry,
Sept.
References
203
64. BackstromS.,
Creep
and creep
recovery
of cement
mortar,
IABSE,
Fifth
Congress, Preliminary
65. England, G. L., Methods of computing stress in concrete of measured strain, Civil June 1965, pp. 846-847.
66. Mamillan, M., A study of the creep of concrete, RILEM,
Bulletin
No. 3,
July 1959, pp. 15-31. 67. Yashin, A. V., Creep of young concrete, pp. 18-73. 68. Davis, R. E., Davls, H. E. and Hamilton, J. S., Plastic flow of concrete under sustained stress, ASTM, Proc. 34, Pt. 2,1934, 69. Glanville, W.H., Creep of concrete
V. 11, No. 2,1933, pp. 54-73.
and creep of
71. Ruetz, W. A hypothesis for creep of hardened cement paste and the influence Concrete, of simultaneous shrinkage, Int. Conf. on the Structure of
72. Reichard, T. W., Creep and drying shrinkage of lightweight normal-weight concretes, National
Mar. 1964, pp. 30. 73. Washa,G. W. and Fluck, P. G., Effect plastic flow of reinforced concrete of compressive reinforcement on the
V-49, No. 8, Oct. 1952, pp. 89-108. 74. CP 110, Code of Practice Standards Institution, 75. Lindsell, P., Monitoring submitted Oct. 1984. for the structural use of concrete, British
Engineering,
University
Ref erences
204
76. Tyler, R. G., Creep, shrinkage and elastic strain in concrete bridges in the United Kingdom, 1963-71, Magazine of Concrete Research, Vol. 28, No. 95, June 1976, pp. 55-84. 77. Parrott, L. J., A study of some long term strains measured in two concrete structures, structures, Proc. lst. Symp. on testing in situ of concrete
78. Swamy, R. N. and Potter, M. M. A., Measurement of in situ stresses in reinforced April concrete columns, The Structural Engineer, Vol. 55, No. 4,
79. Malhak, A. G., Messaite fur einen akustischen Dehnugsmesser, Reichs patentschrift, No. 550,721,1932. wire strain gauges for use in long term , 1953,176, pp. 153-156.
Engineering
in concrete,
Acoustic
vibrating
Wiltshire.
83. British Standards Institution, for other than strength, 84. British
Standards institution,
strength,
85. Bazant, Z. P. and Osman, E., Double power law for basic creep of concrete, Materials and Structures, Vol. 9, No. 49,1976, pp. 3-11. of
cement paste, Proc. 5th Int. Symp. on the Chemistry 1968, Part III, pp. 1-32.
87. Hannant, D. J., The tensile strength of concrete :a
of Cement, Tokyo,
review
paper,
The
Structural
Ref erences
205
in tension and
compression, Mag. of Concrete Research, Vol. 20, No. 65, December 1968, pp-221-228.
89. Neville, A. M., Properties of concrete, Pitman, London, 1977.
90. Powers, T. C. and Brownyard, T. L., Studies of the physical properties hardened Portland cement paste (Nine parts), ACI Journal, 43, Oct. 1946-Apr. 1947. 91. Browne, R. D., Thermal movement of concrete, Concrete, November 1972, pp. 51-53. Current practice
of
sheets,
92. Fintel, M. and Khan, F. R., Effects buildings - prediction Proc., 66, (12), 1969. 93. Pihlajavaara, S.E., Effect carbonation evaluation of inelastic
of painting
surface of
mortars and concretes produced by causes other than stress, Proc. ASTM, V01.30, Pt. 1,1930, pp. 668-685. of concrete
95. Goldbeck, A. T. and F. H. Jackson, Expansion and contraction and concrete roads, U. S. Dept. Agr. Bulletin 96. Hilsdorf, H. K., A method to estimate
shields, Nuclear Engineering and Design, 6,1967, 97. Hansen, T. C., Creep of concrete. Institute, Stockholm, Bulletin
No. 33,1958,
98. Fenwick, R. C., Creep and shrinkage on concrete bridges, RRU Bulletin,
No. 28, National Roads Board, wellington, New Zealand.
Ref erences
206
and Constructional
Engineering (London), V. 39, No. 8, Aug. 1944, pp. 193-199. 101. Pickett, G., Shrinkage stresses in concrete,
V. 42, No. 3, Jan. -Feb.
102. Ross, A. D. et al, Short and long term deformations influenced by its physical structure
of concrete as
of reinforced
APPENDIX
A. 1 Program I:
Computation
of relaxation
function
C C
C--------------------------------------------------------------------function C Calculation of relaxation C--------------------------------------------------------------------DIMENSION T(200), DDER(200), ER(200), CF(200), TF(200) C CMMD N EC2 8, COEFF READ(5, * ) TO) T(2) , AGE1 READ(5, *) EC28, COEFF READ(5, *) N DDER(1)=YMOD(AGEl) ER(1)=DDER(l) AGE1 WRIM6,602) ER(1)=DDER(l) WRIIE(6,603) ERM WRITE (6,6 00)
21
/FLOAT(L-1))
20 22
GOTO 23
23
C-------------------------------------------------------------------C Set up iteration procedure DO 3 I=2, NINT TIM E=AGE1 +T (I) CF(I)=CREEP(T3ME, AGE1) I 1=I- 1 Z1 =-2. / (CREEP(T IM E, TIME)+C REEP(TIM E, AGE1+T(I 1) Z2=0. DO 4 J=1, Il 4 Z2=Z2+(DDER(J)*0.5)*(CREEP(TIME, AGE1+T(J))+ *CREEP(TIME, AGE1+T(J-1))-CREEP(AGEl+T(Il), AGE1+T(J))*CREEP(AGE1+T(I1), AGE1+T(J-1))) DDER(I)=Zl*Z2 E R(I) =E R(I- 1) +D DER(I) C CF=(1 REENAGE1 AGE1 *C F(I)1. ./C , AGE=(l. -(ER(I)/ER(l) ))**(-l. )-(l. /CCF) 3 NF=I 1 WRITE(6,6O1)NF, TIMEj(NF) CF(NF) ER(NF) AGE, CCF , STOP 600 FOFMAT(8X, 4H11me, 6X, 16Hume after loading, 4X, 14HOr eep function, 207
4X,
Appendix
208
*19i%laxation 4X, 1 7HAging coefficient function, 4X, 17HCreep coeffic *i ent) 601 FOEMAT(I3,2X, F9.3,5XF9.3,10X, E15.4,3X, E15.4,15X, F5.3,13X, F6.3) 602FO114AT(17HAgeat loading E8.2) 603 FOR4AT(1 8HYoung's modulus J8.2) END C------------------------------------------------------------------C Function for calculating the development FUNCTION YMOD(X) C CMM3 N EC2 8, COEFF YM T(X/ (4. +. 8 5*X) OD=E C28* SQR R ETUR N END C-------------------------------------------------------------------C Function for calculating creep function FUNCTION CREEP(X, Y) C CMMO N EC2 8, COEFF TCON= 1.2 5/ (Y* *. 1 18) T FUN= (X-Y) * *. 6 (T (10. FIJ C COEFF OEFF*T CON* NCI +T FUNQ =C CREEP=(1 +C COEFF)/YMOD(Y) . R ETUR N END C-------------------------------------------------------------------C of modulus of concrete
Appendix A
209
A. 2 Program 2:
modulus
C
C--------------------------------------------------------------------the effects C Program to calculate of creep in cracked reinforced C concrete members. C **** Ef fective modulus method C------------------------------------------------------------------MR(200) DN(200) AGCF(200) MBD(200), DIMENSION T(200) PHI(200) F CC(2 00) , FST(2 00) , FSC(2 00) REAL*8 MR, MBD EC28, COEFF CCMMON C-- - ----------------------------------------------------------------C Read in values of p, pI and dI/d (5, * )S RT, SRC,DR READ C-------------------------------------------------------------------C Read in steel modulus at loading and concrete READ(5, *)ES, EC28 C------------------------------------------------------------------C Read in age of concrete at loading (5, * )AGE1 READ C--------------------------------------------------------------------to determine C Read in function creep coefficient of age at loading (5, * )C OEFF READ C--------------------------------------------------------------------C Read in number of data
READ(5 * )NT WRITE(6,600)AGE1 EOD=YMOD(AGEl) WRITE(6,601)ECO MR(l) =ES/ECO ACON=S RT+SRC*DR RT+SRC BCON=S CCON=DR IF(CCON. EQ. 1.0)GOTO 5 WRITE (6,6 03) GOTO 6 5 WRITE(6,606) 6 CONrINUE DO 1 I=1, NT c -------------------------------------------------------------------C Read in times after loading, moment and applied aging coefficient AGCF(I), MBD(I) READ(5, *)T(I), T IM E=TM+AGE1 1) PHIM =CREENT IM E, AGE IF(I. GT. 1)GOTO 2 *MR(I)*ACON)-MR(I)*BCON DN(I)=SQRT(MR(I)*MR(I)*BCON*BCON+2. )+SRC*MR(1)*(l. DCON=DN(1)*(l. -DN(l)/3. -CCON/DN(l))*(l. -CCON) G070 3 2E OD N=MR( 1 (1 +PHI (I ) *AGCF( I . A=2. *B CON/3. B= (2. *B CON) + (2. *A CON/3-) -S RC+(S RC*CCON) + (DC ONI ECON) (S C= (2. *A CON) (S RC* CON) C RC*CCON* CCON) + D N(I) (B-S QR T(B* B- 4. * A* C) )/ (2. * A) MRU ) (D N(I) *D NM )/ (2. * (S RT+SRC*CCON-D N(I ) *B CON) ) 3 FC1=0-5*DN(I)*(1. -DN(I)/3. FC2=S RC*MR(I) *(1. -C CON/DN(I) )*(1. -C CON) FCC(I)=MBD(I)/(FC1. +FC2) c
Appendix A
210
FSTM =FCC(I) *MR(I) *(l. /DN(I)1. ) MCCON. EQ. 1.0)GOTO 4 FSC( I) =FCC( I) *MR(I) *0. -CCON/DN(I) WRITE(6,604)T M, PHI(I), DN(I), MBD M, FCC(I) FST(I), FSC( I) GOTO 1 4W RITE (6,6 05) T (I PHI (I D N(I (I MBD FCC(I FST(I 1C ONTINUE STOP 600 FOf? 4AT(17HPge at loading = F7.1) 601 FOIMAT(21Hbdulus at loading = E10.2) 603 F ORAAT(5X, 4HTime, 6X, 3H Phi, 5X, 4Hk (t) 7X, 5HM/bdd, 9X, 3HFcc, 1OX, 3HFst, *1 OX,3HFs c) 604 MAT (F1 0.1,20 X, F6-3) 4(3 X, El 0.4) ) 'F 605 FOfHAT(FlO. 1,2(3X, F6.3), 3(3X, E10.4)) 606 FOR4AT(5X, 4Hllme, 6X, 3HFhi, 5X, 4Hk(t), 7X, 5HM/bdd, 9X, 3HFcc, 10X, 3HFst) END C--------------------------------------------------------------------C C Calculation of the development of modulus of concrete FUNZTION YMOD(X) C CMMO N EC2 8, COEFF YM OD=E C28*SQRT(X/ (4. +. 8 5*X) R ETUR N END C-------------------------------------------------------------------C C Function for calculating to ACI 209 creep function according FUNCTION CREEP(X, Y) C CMMO N EG2 8, COEFF TCON= 1.2 5/ (Y* *. 1 18) T FIJW= (X -Y) **. 6 C REEP=C OEFF*T CON*(T FUNC/ 0 0. +T FUNO R ETUR N END C-------------------------------------------------------------------C
Appendix
211
A. 3 Program 3:
aging
C
C--------------------------------------------------------------------C Program to calculate the effects of creep in cracked reinforced C concrete members. **** C Viscoelastic analysis C------------------------------------------------------------------DIMENSION T(200), FI(200), GI(200), HI(200), HJ(200), V(200), W(200) DIMENSION DN(200), M(200), ALPHA(200), BETA(200), C(17) D IMENSION ER(2 00) D DER(2 00) DIMENSION MBD(200), MBAR(200) R EAL*4 M13D, MBAR,M,MR REAL*8 F, FX, TOLX, X, Y, C INTEGER IFAIL, IND, IR EXTERNAL F COMMON AI, BI, CI, DI C-------------------------------------------------------------------C Read in values of p, pl and d1/d READ(5, *) SRT, SRC, DR C-------------------------------------------------------------------C Read in steel modulus READ(5, *) ES C--------------------------------------------------------------------C time values and age at loading Read in initial (5, *) AGE1 READ C--------------------------------------------------------------------C Read in number of time ordinates (5, *)N INT READ C--------------------------------------------------------------------C Read in values of time and M/bdd E C28=YM 0 D(2 8. ) DO 22 I=1, NINT READ(5, *) T(I), MBD(I) (I) =MBD(I) /E C28 22M BAR C-------------------------------------------------------------------C Evaluation of constants use in analysis MR--ESIEC28 CON1=0. -DR) C ON2=S RC/S RT CON3=SRT*MR CON4=S RC*MR C-------------------------------------------------------------------C Set up iteration procedure and equations WRITE(6,5 90) WRITE(6,591) SRTSRC, DR, AGE1 IF(DR. EQ. 1. ) GOTO 13 WRITE(6,592) GOTO 19 13 WRITE(6,593) 19 DO 3 I=1, NINT TIME=AGE1+T(I) I 1=I- 1 IF(Il. EQ.0) T(Il)=O. O HI(Il)=(CF(TIME, TIt4E)-CF(TIME, AGE1+T(Il)))/2. HO=(CF(TIME, AGE 1+T (11) )-CF(T3ME, AGE1) )12. IF(TIME. EQ. AGE1) GOTO 4 FI(Il)--'ALPHA(Il)*(l. /DN(Il))) -(l.
Appendix A
212
GIM)
-(DR/DN(I1)
))
-(1. -(DR/DNM /DN0)
*0. *0.
4 FIM)=0.0 G1(11) =0.0 M 1) =0.0 W(I 1) =0.0 5 Hl=(CF(TIME, TIME)-CF(TIME, AGEl))/2. M(I) = (H1 -C F(T IM E, TIM E) )*MBAR(I)+Hl *MBAR(I) C ON5 =H1(11) *FI(I 1) C ON6 =H1(11) *G M 1) C OUT =W(I 1) +C ON6 CON8=H1(11) -C F(T IM E, TIM E) A I=V U 1) +C ON5 +CON2*W(I 1) +C ON2*CON6 *C ON2*CON1*CON7)-(3 *V (11) )-(3. *CON5) B I= (3 *MM VC ON3 +(3. * -(3. *CON2*W(Il) )-(3. *CON2*CON6) * (M(I)+CON4*CON1*CON7)*CON8) CI=(6. *V(Il)*CON4*CON8*CON1)-(6. * +(6. *CON5*CON4*CONB*CON1)+(6. *CON2*W(Il)*CON4*CON8*CON1) * -(6. *CON2*(M(I)+CON4*CON1*CON7)*CON8) * +(6. *CON2*CON6*CON4*CON8*CON1) *(M(I)+CON4*CON1*CON7)*CON8) DI=(-CI*DR)+(l-DR)*(6. C-------------------------------------------------------------------C C05AZF references Subroutine TOLX=l. E-5 X=0.0 Y=1.0 IR=O I ND= 1 IFAIL=1 C-------------------------------------------------------------------Calculate C depth of neutral axis 8 CALL C05AZF(X, Y, FX, TOLX, IR, C, IND, IFAIL) IF(IND. EQ.0) GOTO 6 IF(IND. LT. 2 OR. IND. GT. 4) GOTO 7 FX=F(X) GOM 8 6D NM =X C-------------------------------------------------------------------C Calculate deformation parameters ) CCN9=1 -(DR/DN(I) . APH1=6. * (M(I)+CON4*CON1*CON7) C
Appendix A
213
C------------------------------------------------------------------C Calculate the steel stresses ETA( I)) 11 STRST=ES* (ALPHA(I)+B IF(DR. EQ. 1. ) GOTO 12 STRSC=ES*(ALPHA(I)+DR*BETA(I) DN(I), T]24EMBD(I), WRITE(6,610) GOTO 3 TIME, MBD(I), DN(I) 12 WRITE(6,620) 3C ONrINUE WRITE (6,5 90) 7S TOP 590 FOf?4AT(77(
ALPHA(I), ALPHA(I),
BETA(I), BETA(I),
. 591 FOR4AT(lX, 2Hp=, lX, F6.4/lX, 3Hpl=, lX, F6.4/lX, 5Hd'/d=, lX, F6.4/ *lX, 1(HAgeat loading =, lX, F6.1) 592 FOHIAT(25( 1*1 )//2X, 3HAge, 7X, 4HM/bd, 5X, 8HNA Depth, 5X, 5HAlpha, 7X, 4H *Beta, 8X, 3HFc c, 9X, 3HFs t, 9X, 3HFsc) 593 FOR4AT(25( 1*1 )//2X, 3HAge, 7X, 4HM/bd, 5X, 8HNA Depth, 5X, 5HAlpha, 7X, 4H *Beta, 8X, 3HFbc, 9X, 3HFst) 600 FORIAME12-3) 610 FOfMAT(lXF7.1,3X, E10.4,3X, F5.3,3X, 5El2.4) E10.4,3X, F5.3,3X, 4El2.4) 620 FOIMAT(lXF7.1,3X, END C--------------------------------------------------------------------C for calculating C Function the development of modulus of concrete C to BP Method according FUNCTION YMODW R EAL*4 MN DATA TO/1. /
DATA EO/5 9. EY, PHI/2.8/, ALPHA/.045/ 9M/ 12/ N/. -3/, 9083-51 DATA BH/. 0 59/, TSH14
XB=X-TO X1 =XB/ (1. +(XB/TSH) X2: 0. -B H) *X 1 XE=TO+(BH*XB)+X2 C ON1=1./EO CON2=PHI/EO C ON3=1 (10. * *N) -/ (XE* *M) )+ALPHA / 1. C ON4 =( CON5---C ON1 +C ON2*CON3*CON4 YMOD= 1C ON5 R ETUR N END C-------------------------------------------------------------------C Function for calculating the creep function according F=TION CREEP(X, Y) REAL*4 M, N, KH1, KH2 DATA TO/4. / DATA EO/59. E3/, PHI/2.8/, ALPHA/0.045/#M/0-3/, N/0.12/ DATA CD1/1.9/, TSW49083.5/, PHID/0-0326/, KH1/0.284/, DATA CP 1/0.8 3/, KH2/0-3 6/ IF(X. EQ.Y) GOTO 22 CO=(PHI/EO) *(Y**(-M)+ALPHA) *(X-Y) **N SD=(1. +(l 0. *TSH/ (X-Y) ))**(-CD1*N) PHID1=( (1. +(Y-TO) /0 0. *TSH) )**(-. 5) )*PHID CD=(PHIDl/EO) *(Y**(-M/2. ) )*KH1*ESH*SD SP=(l. +(l 00. *TSH/ (X-TO) ))**(-N) CP=CP1*KH2*SP*CO C to BP
ESH/450. /
Appendix A
214
CD-CP CREEP= 1. /EO 4C0+. GOTO 23 22 CREEP=1. /YMOD(Y) 23 RETURN END C--------------------------------------------------------------------for calculating Function C specific creep non-dimensionalised FUW,TION CF(X, Y) CF=YMOD(28. )*CREEP(X, Y) R ETUR N END C-------------------------------------------------------------------function the relaxation for calculating Function C FUW,TION RF(X, Y) DIMENSION T(200), ER(200), DDER(200) DATA TO) T(2) /0. . 1/ I F(X. EQ.Y) GOTO 14 OD(Y) R1) =YM D EE E RO =DDER(1 L=1 0 /FLOAT(L- 1) 21 R= ( (X-Y) /T(2) )**(l. 0. * *(1. /16. ) RMAX=1 IF(R. LE. RMAX) GOTO 20 L=L+ 1 GOTO 21 20 C ONTINUE K=3 15 TW =T (K- 1) *R IF( (T(K)+Y) GT. X) GOTO 18 K=K+ 1 GOTO 15 18 N INT=K -1 DO 16 I=2, NINT X 1=Y+TU) Il=I-1 (X 1, Xl)+CREEP(X 1, Y+T (11) )) Zl=-2. / (CREEP Z2=0. DO 17 J=1, Il Y+T(J-1)) 17 Z2=Z2+(DDER(J)*. 5)*(CREEP(Xl, Y+T(J))+CREEP(Xl, Y+T(J-1))) Y+T(J))-CREEP(Y+T(Il), -CREEP(Y+T(Il), DDER(I)=Zl*Z2 16 E RM =E R(I- 1) +D DERM RF=ERM N R ETUR 0 DW 14 RF=YM N R ETUR END --------------------------------------------------------------------Function for subroutine C C05AZF references REAL*8 FUNCTION F(X) R EAL*8 X C 04MON Al, BI, CI, DI F=(AI*X*X*X)+(BI*X*X)+(CI*X)+DI RETURN END --------------------------------------------------------------------C function
APPENDIX B
B. 1 Design charts
The shift of the neutral axis and changes in the concrete and steel stresses caused by creep, can be conveniently charts shown in Figures B. 1 to BA computed from the design concrete
beam sections under service Loads. The ACI 209 method was used to derive the creep functions and the parameters charts were, adopted in the preparation of the
Age at loading = 28 days Characteristic Characteristic concrete cube strength strength = 30 N/m M2 = 410 N/mm'
of reinforcement
of concrete
at loading = 25 kN/m M2
The corresponding values of k, OC 7,360 and at and as as' f max 10000 days after application of the load, are expressed as ratios to the values at instantaneous load application. These ratios are given in the
ckarts for various values of p and pl, and are defined as such,
K,
= ratio
for for
neutral maximum
axis depth
K2 = ratio
compressive
215
Appendix B
216
*3= *4=
B. 2 Worked examples
The worked examples are based on the use of the design charts and illustrate the calculation of the mid-span creep deflections relationships for two
are used,
1 rb
Es
(B. 1)
6b
(B. 2)
where L=
1= rb
effective
mid-span
span
curvature for beams
6b=
mid-span deflection
Example 1:
Singly reinforced
section
b=
100mm, d=
250mm, pl = 0, p=1.5%
and L=
3m
128.3 N/mM2 = of resistance kNrn 10.47 = from Figures table. B. 1 and B. 3, and the
moment
of K,
are surnmarlsed
in the following
Appendix B
217
(t
t') 7
k(t)d
cy s
(N/MM2)
(x10-6 )
(mm)
360 10000
Example 2:
Doubly reinforced
section
b=
250mm, d=
and L=
5m
Based on a cracked section analysis, k(to)d = 166.6mm max. compressive concrete stress = 10 N/mM2 Tensile steel stress = 124.1 N/mM2 Cracking moment of resistance = 98.13 kNm The values of K1 and K3 are obtained from Figures B. 1 and B. 3 and the calculations are surnmarised in the following table.
(N/MM2)
(mm)
Appendix B
218
r4l,
CC cj co C c (LI
cn >
.0 tu 0 cu C)
,U
co
.H 'n
14 D L)
L) w aj
cp
cD '0
81
I'l
II
I'l
a, 3
4J co V) k0 1 cu
V)
11)
Ul)
-Z) Z
.Z i
2s
2 .x m 76
z
9 18 IIIII
cu
LL
/ /
'0 '-4
/
(N C
6 CL
CD
Appendix
219
C
r4 N
cu
(U W CM cu L) CU
4j c (1, m
C\ U)
ll
(U cu CL
In
CU 1-4 (U cu 4j ca
V)
(D
0
4- Ln
4-)
a)
u c 0 u
ci
C
tn
tn r
6 0
C CD
C 0
r CD
ZA
1 CD
n CD
n 0
Appendix B
220
1 fl,
1 CN
C Z Co
Ln
1
c Ln c
C)
C
cw
co
w a, (L M S.
co
I'D
CD
-1
C14
CD
Appendix B
221
U-' (N
Ln 0
C (N
cu
LO
(D V)
%n 41
>1 'D
C -4 C
Z CD
cu Lt -
-4 G c') C 0 -4 C
'o
n wck s9 C6
V) (D
-4
I. CD-
Ln
E C) u 14:
4J
D oi
0 -4
LL
cli
CD
0
l
CN
c2
Co
"0
r --;
r4
CD
17 A