High Performance Concrete and Applications by Shah and Ahmed
High Performance Concrete and Applications by Shah and Ahmed
High Performance Concrete and Applications by Shah and Ahmed
CONCRETES AND
APPLICATIONS
Edited by
S P Shah
Walter P Murphy Professor of Civil Engineering, and
Director of NSF Center for Science and Technology
of Advanced Cement Based Materials
North Western University, Evanston, IL, USA
S H Ahmad
Professor of Civil Engineering
North Carolina State University
Raleigh, NC, USA
Edward Arnold
A member of the Hodder Headline Group
LONDON MELBOURNE AUCKLAND
© 1994 P Shah and S H Ahmad
P Acker
Head, Division: 'Betons et Ciments pour Ouvrages d'Art', Laboratoire
Central des Fonts et Chaussees, Paris, France
S H Ahmad
Professor, Department of Civil Engineering, North Carolina State
University, Raleigh, NC, USA
P N Balaguru
Professor, Civil Engineering Department, Rutgers University, Piscataway,
N J, US A
T W Bremner
Professor of Civil Engineering, University of New Brunswick, Fredericton,
Canada
Fde Larrard
Senior Scientist, Division: 'Betons et Ciments pour Ouvrages d'Art',
Laboratoire Central des Fonts et Chaussees, Paris, France
A S Ezeldin
Assistant Professor, Department of Civil, Environmental and Coastal
Engineering, Stevens Institute of Technology, Hoboken, NJ, USA
R Gettu
Senior Researcher, Technical University of Catalunya, Barcelona, Spain
S K Ghosh
Director, Engineered Structures and Codes, Portland Cement
Association, Skokie, IL, USA
O E Gj0rv
Professor, Division of Building Materials, Norwegian Institute of
Technology - NTH, Trondheim - NTH, Norway
TA Holm
Vice President of Engineering, Solite Corporation, PO Box 27211,
Richmond, VA, USA
R Le Roy
Research Engineer, Division: 'Betons et Ciments pour Ouvrages d'Art',
Laboratoire Central des Ponts et Chaussees, Paris, France
S Mindess
Professor, Department of Civil Engineering, University of British
Columbia, Vancouver, Canada
S Nagataki
Professor of Civil Engineering, Tokyo Institute of Technology, O-
okayama, Meguru-ku, Tokyo 152, Japan
A H Nilson (Professor)
162 Round Pound Road, HC-60, Box 162, Medomak, Maine, USA
(formerly of Cornell University)
H G Russell
Vice President, Construction Technology Laboratories Inc., 5420 Old
Orchard Road, Skokie, IL, USA
M Saatcioglu
Associate Professor of Civil Engineering, University of Ottawa, Canada
E Sakai
Manager, Special Cement Additives Division, Denki Kagaku Kogyo Co.
Ltd, Yuraku-cho, Chiyoda-ku, Tokyo 100, Japan
S P Shah
Walter P Murphy Professor of Civil Engineering; Director, NSF Center for
Science and Technology of Advanced Cement-Based Materials; and
Director, Center for Concrete and Geomaterials, Northwestern
University, Evanston, IL, USA
Contents
Preface .................................................................................... ix
1.1 Introduction
High performance concretes (HPC) are concretes with properties or
attributes which satisfy the performance criteria. Generally concretes with
higher strengths and attributes superior to conventional concretes are
desirable in the construction industry. For the purposes of this book, HPC
is defined in terms of strength and durability. The researchers of Strategic
Highway Research Program SHRP-C-205 on High Performance Concrete1
defined the high performance concretes for pavement applications in terms
of strength, durability attributes and water-cementitious materials ratio as
follows:
• It shall have one of the following strength characteristics:
4-hour compressive strength ^25OO psi (17.5 MPa) termed as very
early strength concrete (VES), or
24-hour compressive strength ^50OO psi (35 MPa) termed as high
early strength concrete (HES), or
28-day compressive strength ^10,0OO psi (70 MPa) termed as very
high strength concrete (VHS).
• It shall have a durability factor greater than 80% after 300 cycles of
freezing and thawing.
• It shall have a water-cementitious materials ratio =$0.35.
High strength concrete (HSC) could be considered as high performance
if other attributes are satisfactory in terms of its intended application.
Generally concretes with higher strengths exhibit superiority of other
attributes. In North American practice, high strength concrete is usually
considered to be a concrete with a 28-day compressive strength of at least
6000 psi (42MPa). In a recent CEB-FIP State-of-the-Art Report on High
Strength Concrete2 it is defined as concrete having a minimum 28-day
compressive strength of 8700 psi (6OMPa). Clearly then, the definition of
'high strength concrete' is relative; it depends upon both the period of time
in question, and the location.
The proportioning (or mix design) of normal strength concretes is based
primarily on the w/c ratio 'law' first proposed by Abrams in 1918. At least
for concretes with strengths up to 6000 psi (42MPa), it is implicitly
assumed that almost any normal-weight aggregates will be stronger than
the hardened cement paste. There is thus no explicit consideration of
aggregate strength (or elastic modulus) in the commonly used mix design
procedures, such as those proposed by the American Concrete Institute.3
Similarly, the interfacial regions (or the cement-aggregate bond) are also
not explicitly addressed. Rather, it is assumed that the strength of the
hardened cement paste will be the limiting factor controlling the concrete
strength.
For high strength concretes, however, all of the components of the
concrete mixture are pushed to their critical limits. High strength concretes
may be modelled as three-phase composite materials, the three phases
being (i) the hardened cement paste (hep); (ii) the aggregate; and (iii) the
interfacial zone between the hardened cement paste and the aggregate.
These three phases must all be optimized, which means that each must be
considered explicitly in the design process. In addition, as has been pointed
out by Mindess and Young,4
'it is necessary to pay careful attention to all aspects of concrete
production (i.e. selection of materials, mix design, handling and
placing). It cannot be emphasized too strongly that quality control is
an essential part of the production of high-strength concrete and
requires full cooperation among the materials or ready-mixed sup-
plier, the engineer, and the contractor'.
In essence then, the proportioning of high strength concrete mixtures
consists of three interrelated steps: (1) selection of suitable ingredients -
cement, supplementary cementing materials, aggregates, water and che-
mical admixtures, (2) determination of the relative quantities of these
materials in order to produce, as economically as possible, a concrete that
has the desired rheological properties, strength and durability, (3) careful
quality control of every phase of the concrete-making process.
Portland cement
There are two different requirements that any cement must meet: (i) it
must develop the appropriate strength; and (ii) it must exhibit the
appropriate rheological behaviour.
High strength concretes have been produced successfully using cements
meeting the ASTM Standard Specification C150 for Types I, II and III
portland cements. Unfortunately, ASTM C150 is very imprecise in its
chemical and physical requirements, and so cements which meet these
rather loose specifications can vary quite widely in their fineness and
chemical composition. Consequently, cements of nominally the same type
will have quite different rheological and strength characteristics, particu-
larly when used in combination with chemical admixtures and sup-
plementary cementing materials. Therefore, when choosing portland ce-
ments for use in high strength concrete, it is necessary to look carefully at
the cement fineness and chemistry.
Fineness
Increasing the fineness of the portland cement will, on the one hand,
increase the early strength of the concrete, since the higher surface area in
contact with water will lead to a more rapid hydration. On the other hand,
too high a fineness may lead to rheological problems, as the greater
amount of reaction at early ages, in particular the formation of ettringite,
will lead to a higher rate of slump loss. Early work by Perenchio5 indicated
that fine cements produced higher early age concrete strengths, though at
later ages differences in fineness were not significant. Most cements now
used to produce high strength concrete have Elaine finenesses that are in
the range of 1467 to 1957 ft2/lb (300 to 400 m2/kg), though when Type III
(high early strength) cements are used, the finenesses are in the range of
2201 ft2/lb (450 m2/kg).
Chemical composition of the cement
The previously cited work of Perenchio5 indicates that cements with higher
C3A contents leads to higher strengths. However, subsequent work6 has
shown that high C3A contents generally leads to rapid loss of flow in the
fresh concrete, and as a result high C3A contents should be avoided in
cements used for high strength concrete. Aitcin7 has shown that the C3A
should be primarily in its cubic, rather than its orthorhombic, form.
Further, Aitcin7 suggests that attention must be paid not only to the total
amount of SO3 in the cement, but also to the amount of soluble sulfates.
Thus, the degree of sulfurization of the clinker is an important parameter.
In addition to commercially available cements conforming to ASTM
Types I, II and III, a number of cements have been formulated specifically
for high strength concrete. For instance, in Norway, Norcem Cement has
developed two special cements for high strength concrete, in addition to
their ordinary portland cement. The characteristics of these cements are
given in Table 1.1.s Note that for the two special cements (SP30-4A and
SP30-4A MOD), the C3A contents were held to 5.5%.
Table 1.1 Composition of special cements for high strength concrete (developed
by Norcem Cement8)
SP30* SP30-4A SP30-4A MOD
C2S (%) 18 28 28
C3S (%) 55 50 50
C3A (%) 8 5.5 5.5
C4AF (%) 9 9 9
MgO (%) 3 1.5-2.0 1.5-2.0
S0 3 (%) 3.3 2-3 2-3
Na2O equivalent (%) 1.1 0.6 0.6
Elaine fineness (m2/kg) 300 310 400
heat of hydration (kcal/kg) 71 56 70
setting time (min): initial 120 140 120
final 180 200 170
* Ordinary portland cement, for comparison
Im 2 /kg = 4.89ft 2 /lb
Silica fume
It is possible to make high strength concrete without silica fume, at
compressive strengths of up to about 14,000psi (98MPa). Beyond that
strength level, however, silica fume becomes essential, and even at lower
strengths 9000-14,000 psi (63-98 MPa), it is easier to make HSC with silica
fume than without it. Thus, when it is available at a reasonable price, it
should generally be a component of the HSC mix.
Silica fume14 is a waste by-product of the production of silicon and
silicon alloys, and is thus not a very well-defined material. Consequently, it
is important to characterize any new source of silica fume, by determining
the specific surface area by nitrogen adsorption, and the silica, alkali and
carbon contents. In addition, it is desirable to minimize the content of
Table 1.4 Some Canadian specifications for silica
fume (taken from CSA Standard A23.59)
Chemical requirements
Si0 2 ,min(%) 85
SO 3 ,max(%) 1.0
Loss in ignition, max (%) 6.0
Physical requirements
Accelerated pozzolanic activity index, min, (%) of control 85
Fineness, max, (%) retained on 45 jxm sieve 10
Soundness - autoclave expansion or contraction (%) 0.2
Relative density, max variation from average (%) 5
Fineness, max variation from average (%) 5
crystalline material. The acceptance limits for silica fume, taken from
CSA-A23.59 are given in Table 1.4.
Silica fume is available in several forms. In its bulk form, its unit weight
is in the range of 118 to 147.5 pcf (200-250 kg/m 3 ), which makes handling
difficult. More commonly now, silica fume is available in a densified form,
in which the bulk densities are about twice as great as those of the bulk
form (i.e. 400-500 kg/m 3 ). In general, this makes it easier to handle. In
addition, silica fume is available in slurry form (often in conjunction with
superplasticizers in the liquid phase), with a solids content of about 50%.
This form of silica fume requires special equipment for its use. Finally,
silica fume is available already blended with portland cement (at percen-
tages of the total mass of cementitious material in the range of 6.7 to 9.3%)
in Canada, France and Iceland. In spite of this apparently wide selection,
however, in any one location the choice of silica fumes will be very limited,
and one must use what is locally available.
Fly ash
Fly ash has, of course, been used very extensively in concrete for many
years. Fly ashes are, unfortunately, much more variable than silica fumes
in both their physical and chemical characteristics. Any fly ash which works
well in ordinary concrete mixes is likely to work well in high strength
concrete as well. However, most fly ashes will result in strengths of not
much more than 10,000 psi (70 MPa), though there have been a few reports
of high strength concretes with strengths of up to 14,000 psi (98MPa) in
which fly ash has been used. For higher strengths, silica fume must be used
in conjunction with the fly ash, though this practice has not been common
in the past.
In general, for high strength concrete applications, fly ash is used at
dosage rates of about 15% of the cement content. Because of the
variability of the fly ash produced even from a single plant, however,
quality control is particularly important. This involves determinations of
the Elaine specific surface area, as well as the chemical composition (in
particular the contents of SiO2, Al2O3, Fe2O3, CaO, alkali, carbon and
sulfates). And, as with silica fume, it is important to check the degree of
crystallinity; the more glassy the fly ash, the better.
Superplasticizers
In modern concrete practice, it is essentially impossible to make high
strength concrete at adequate workability in the field without the use of
Superplasticizers. Unfortunately, different Superplasticizers will behave
quite differently with different cements (even cements of nominally the
same type). This is due in part to the variability in the minor components of
the cement (which are not generally specified), and in part to the fact that
the acceptance standards for Superplasticizers themselves are not very
tightly written. Thus, some cements will simply be found to be incompati-
ble with certain Superplasticizers.
There are, basically, three principal types of superplasticizer: (i) ligno-
sulfonate-based\ (ii) polycondensate of formaldehyde and melamine sul-
fonate (often referred to simply as melamine sulfonate; and (iii) polycon-
densate of formaldehyde and naphthalene sulfonate, (often referred to as
naphthalene sulfonate).
In addition, a variety of other molecules might be mixed in with these
basic formulations. It may thus be very difficult to determine the precise
chemical composition of most Superplasticizers; certainly manufacturers
try to keep their formulations as closely guarded secrets.
It should be noted that much of what we know about Superplasticizers
comes from tests carried out on normal strength concretes, at relatively low
superplasticizer contents. This does not necessarily reflect their perform-
ance at very low w/c ratios and very high superplasticizer addition rates.
Lignosulfonate-based Superplasticizers
In high strength concrete, lignosulfonate Superplasticizers are generally
used in conjunction with either melamine or naphthalene Superplasticizers.
They tend not to be efficient enough for the economic production of very
high strength concretes on their own. Sometimes, lignosulfonates are used
for initial slump control, with the melamines or naphthalenes used
subsequently for slump control in the field.
Superplasticizer dosage
There is no a priori way of determining the required Superplasticizer
dosage; it must be determined, in the end, by some sort of trial and error
procedure. Basically, if strength is the primary criterion, then one should
work with the lowest w/c ratio possible, and thus the highest Superplasticiz-
er dosage rate. However, if the rheological properties of the high strength
concrete are very important, then the highest w/c ratio consistent with the
required strength should be used, with the Superplasticizer dosage then
adjusted to get the desired workability. In general, of course, some
intermediate position must be found, so that the combination of strength
and rheological properties can be optimized. Typical Superplasticizer
dosages for a number of high strength concrete mixes are given below, in
Tables 1.5 to 1.10.
Table 1.5 Mix proportions for Interfirst Plaza, Dallas (adapted from Cook15)
/ cm max 25 cm max
size aggregate size aggregate
water (kg/m3) 166 148
cement, Type I (kg/m3) 360 357
fly ash, Class C (kg/m3) 150 149
coarse aggregate (kg/m3) 1052 1183
fine aggregate (kg/m3) 683 604
water reducer L/m3 1.01 1.01
Superplasticizer L/m3 2.54 2.52
w/cementitious ratio 0.33 0.29
fc 28-day (MPa)- moist cured 79.5 85.8
fc 91-day (MPa)- moist cured 89.0 92.4
1 lb/yd3 = 0.59 kg/m3 or lkg/m 3 = 1.69pcf
1 in. =25. 4 mm or 1 in. = 0.0393 mm
Table 1.6 High strength concrete mix design guidelines (after Peterman and
Carrasquillo16)
H-H-OO H-H-Ol H-H-IO H-H-Il
3
water (kg/m ) 195 143 173 134
cement (kg/m3) 558 474 391 335
fly ash (kg/m3) 167 144
coarse agg./fine agg. ratio 2.0 2.0 2.0 2.0
superplasticizer yes* yes*
w/cementitious ratio 0.34 0.30 0.31 0.27
fc' 56-day (MPa) 66 72 69 76
* Use highest dosage of superplasticizer which will not lead to segregation or
excessive retardation.
1 lb/yd3 = 0.59 kg/m3 or 1 kg/m3 = 1.69 pcf
lin. =25. 4 mm or 1 in. =0.0393 mm
Table 1.7 Mix proportions for high strength concrete at Pacific First Center,
Seattle (adapted from Randall and Foot17)
water (kg/m3) 131
cement - Type II (kg/m3) 534
fly ash -Type F (kg/m3) 59
silica fume (kg/m3) 40
coarse aggregate - 1 cm max. size (kg/m3) 1069
fine aggregate - F. M. =3.2 (kg/m3) 623
water reducer I (LIm3) 1.77
water reducer II (L/m ) 7.39
w/cementitious ratio 0.21
fc' 56-day (MPa) 124
1 lb/yd3 = 0.59 kg/m3 or 1 kg/m3 = 1.69 pcf
1 in. = 25.4 mm or 1 in. = 0.0393 mm
Table 1.8 Five examples of commercially produced high strength concrete mix
designs (after Aitcin, Shirlaw and Fines18)
water (kg/m3) 195 165 135 145 130
cement (kg/m3) 505 451 500 315 513
fly ash (kg/m3) 60
slag (kg/m3) 137
silica fume (kg/m3 ) 30 36 43
coarse aggregate (kg/m3) 1030 1030 1110 1130 1080
fine aggregate (kg/m3) 630 745 700 745 685
water reducer (L/m3 ) 0.975 0.9
retarder (L/m3) 4.5 1.8
superplasticizer (L/m3 ) 11.25 14 5.9 15.7
w/cementitious ratio 0.35 0.37 0.27 0.31 0.25
fc' 28-day (MPa) 64.8 79.8 42.5 83.4 119
fc' 91-day (MPa) 78.6 87.0 106.5 93.4 145
1 lb/yd3 = 0.59 kg/m3 or 1 kg/m3 = 1.69 pcf
1 in. = 25.4 mm or 1 in. = 0.0393 mm
Table 1.9 High strength mixtures in the Chicago area (adapted from Burg and
Ost19)
Mix number
1 2 3 4 5
3
water (kg/m ) 158 160 155 144 151
cement (kg/m3) 564 475 487 564 475
3
fly ash (kg/m ) 59 104
silica fume (kg/m3 ) 24 47 89 74
coarse aggregate, SSD 12 mm 1068 1068 1068 :L068 1068
max size
fine aggregate, SSD (kg/m3) 647 659 676 593 593
superplasticizer - Type F (L/m3) 11.61 11.61 11.22 20.12 16.45
3
retarder - Type D (L/m ) 1.12 1.04 0.97 1.47 1.51
w/cementitious ratio 0.281 0.287 0.291 0.220 0.231
fc 28-day (MPa) - moist cured 78.6 88.5 91.9 118.9 107.0
fc' 56-day (MPa) - moist cured 81.4 97.3 94.2 121.2 112.0
fc 91-day (MPa) - moist cured 86.5 100.4 96.0 131.8 119.3
Table 1.10 Mix design for a high strength concrete designed for a low heat of
hydration (adapted from Burg and Ost )
water (kg/m3 ) 141
cement - Type I (kg/m3 ) 327
fly ash - Type F (kg/m3) 87
silica fume (kg/m3 ) 27
coarse aggregate - 25 mm max. size (kg/m3) 121
fine aggregate (kg/m3 ) 742
superplasticizer, ASTM Type F (L/m3) 6.31
superplasticizer, ASTM Type G (L/m ) 3.25
water/cementitious ratio 0.32
fc 28-day (MPa) - moist cured 3.1
fc 91-day (MPa) - moist cured 88.6
Retarders
At one time retarders were recommended for some high strength concrete
applications, to minimize the problem of over rapid slump loss. However,
it is difficult to maintain a compatibility between the retarder and the
superplasticizer, i.e. to minimize slump loss without excessively reducing
early strength gain. In modern practice, retarders are recommended only
as a last resort; the rheology is better controlled by the use of the
appropriate supplementary cementing materials described above.
Aggregates
The aggregate properties that are most important with regard to high
strength concrete are: particle shape, particle size distribution, mechanical
properties of the aggregate particles, and possible chemical reactions
between the aggregate and the paste which may affect the bond. Unlike
their use in ordinary concrete, where we rarely consider the strength of the
aggregates, in high strength concrete the aggregates may well become the
strength limiting factor. Also, since it is necessary to maintain a low w/c
ratio to achieve high strength, the aggregate grading must be very tightly
controlled.
Coarse aggregate
It goes without saying that, for high strength concrete, the coarse aggregate
particles themselves must be strong. A number of different rock types have
been used to make high strength concrete; these include limestone,
dolomite, granite, andesite, diabase, and so on. It has been suggested1 that
in most cases the aggregate strength itself is not usually the limiting factor
in high strength concrete; rather, it is the strength of the cement-aggregate
bond. As with ordinary concretes, however, aggregates that may be
susceptible to alkali-aggregate reaction, or to D-cracking, should be
avoided if at all possible, even though the low w/c ratios used will tend to
reduce the severity of these types of reaction.
From both strength and rheological considerations, the coarse aggregate
particles should be roughly equi-dimensional; either crushed rock or
natural gravels, particularly if they are of glacial origin, are suitable. Flat or
elongated particles must be avoided at all costs. They are inherently weak,
and lead to harsh mixes. In addition, it is important to ensure that the
aggregate is clean, since a layer of silt or clay will reduce the cement-
aggregate bond strength, in addition to increasing the water demand.
Finally, the aggregates should not be highly polished (as is sometimes the
case with river-run gravels), because this too will reduce the cement-
aggregate bond.
Not much work has been carried out on the effects of aggregate
mineralogy on the properties of high strength concrete. However, a
detailed study by Aitcin and Mehta,20 involving four apparently hard
strong aggregates (diabase, limestone, granite, natural siliceous gravel)
revealed that the granite and the gravel yielded much lower strengths and
E-values than the other two aggregates. These effects appeared to be
related both to aggregate strength and to the strength of the cement-
aggregate transition zone. Cook15 has also pointed out the effect of the
modulus of elasticity of the aggregate on that of the concrete. However,
much work remains to be done to relate the mechanical and mineralogical
properties of the aggregate to those of the resulting high strength concrete.
It is commonly assumed that a smaller maximum size of coarse aggregate
will lead to higher strengths,1'2'5'6'21 largely because smaller sizes will
improve the workability of the concrete. However, this is not necessarily
the case. While Mehta and Aitcin6 recommend a maximum size of
10-12 mm, they report that 20-25 mm maximum size may be used for high
strength concrete. On the other hand, using South African materials,
Addis22 found that the strength of his high strength concrete increased as
the maximum size of aggregate increased from 13.2 to 26.5 mm. This, then,
is another area which requires further study.
Fine aggregate
The fine aggregate should consist of smooth rounded particles,2 to reduce
the water demand. Normally, the fine aggregate grading should conform to
the limits established by the American Concrete Institute3 for normal
strength concrete. However, it is recommended that the gradings should lie
on the coarser side of these limits; a fineness modulus of 3.0 or greater is
recommended,1'6 both to decrease the water requirements and to improve
the workability of these paste-rich mixes. Of course, the sand too must be
free of silt or clay particles.
Proportions of materials
Water/cementitious ratio
For normal strength concretes, mix proportioning is based to a large extent
on the w/c ratio 'law'. For these concretes, in which the aggregate strength
is generally much greater than the paste strength, the w/c ratio does indeed
determine the strength of the concrete for any given set of raw materials.
For high strength concretes, however, in which the aggregate strength, or
the strength of the cement-aggregate bond, are often the strength-
controlling factors, the role of the w/c ratio is less clear. To be sure, it is
necessary to use very low w/c ratios to manufacture high strength concrete.
However, the relationship between w/c ratio and concrete strength is not as
straightforward as it is for normal strength concretes.
28-day compressive strength (MPa)
w/c ratio
Fig. 1.1 Compressive strength versus w/c material ratio: (1) after Aitcin7; (2) after Fiorato25;
(3) after Cook15; (4) normal strength concrete from CPCA24
w/c ratio
Fig. 1.2 Compressive strength versus w/c material ratio: (1) high early strength cement, after
Addis and Alexander23; (2) after Hattori26; (3) ordinary Portland cement, after Suzuki27; (4)
high early strength cement after Suzuki24
Superplasticizers
With very careful mix design and aggregate grading, it is possible to
achieve strengths of about 14,000 psi (98MPa) without Superplasticizers.
However, as they are readily available they are now almost universally
used in high strength concrete, since they make it much easier to achieve
adequate workability at very low w/cementitious ratios.
Age at test
Traditionally, the acceptance standards for concrete involve strength
determinations at an age of 28 days. Although there is, of course, nothing
magical about this particular test age, it has been used universally as the
reference time at which concrete strengths are reported. However, for high
strength concretes, it has become common to determine compressive
strengths at 56 days, or even 90 days. The justification for this is that
concrete in structures will rarely, if ever, be loaded to anything
approaching its design strength in less than 3 months, given the pace of
construction. The increase in strength between 28 and 56 or 90 days can be
considerable (10% to 20%), and this can lead to economies in construc-
tion. It is thus perfectly reasonable to measure strengths at later ages, and
to specify the concrete strength in terms of these longer curing times.
There are, however, two drawbacks to this approach. First, it can be
misleading to compare the compressive strengths of normal and high
strength concretes, if these are measured at different times. Of more
importance, there is a certain margin of safety when concrete strengths are
measured at 28 days, since the concrete will generally be substantially
stronger when it finally has to carry its design loads, perhaps at the age of
one year for a typical high-rise concrete building. If strengths are specified
at later ages, this margin is reduced (by an unknown amount), and hence
there is an implicit reduction in the factor of safety. And, of course, finding
higher strengths at later test ages does not in any way imply that the
concrete has somehow become 'better' than a concrete whose strength was
measured in the conventional way at 28 days.
Curing conditions
In general, the highest concrete strengths will be obtained with specimens
continuously moist cured (at 100% relative humidity) until the time of
testing. Unfortunately, the available data on this point are ambiguous.
Carrasquillo, Nilson and Slate28 found that high strength concrete, moist-
cured for 7 days and then allowed to dry at 50% relative humidity till 28
days showed a strength loss of about 10% when compared to continuously
moist-cured specimens. However, in subsequent work, Carrasquillo and
Carrasquillo29 found that up to an age of 15 days, specimens treated with a
curing compound and allowed to cure in the field under ambient conditions
yielded slightly higher strengths than moist-cured specimens. At 28 days,
moist-cured specimens and field-cured specimens (with or without curing
compounds) yielded approximately the same results. Only at later ages (56
and 91 days) did the strengths of the moist-cured specimens surpass those
of the field-cured specimens treated with a curing compound. Similarly, for
the mixes shown in Table 1.9, Burg and Ost19 found that, when specimens
that had been moist cured for 28 days were then subjected to air curing,
their strengths at 91 days exceeded those of continuously moist-cured
specimens; however, by 426 days, the continuously moist-cured specimens
were from about 3% to 10% higher in strength than the air-cured ones.
On the other hand, several investigators have reported that, as long as a
week or so of moist curing is provided, subsequent curing under ambient
conditions is not particularly detrimental to strength development. Peter-
man and Carrasquillo16 have stated that 'the 28-day compressive strength
of high strength concrete which has been cured under ideal conditions for 7
days after casting is not seriously affected by curing in hot or dry conditions
from 7 to 28 days after casting.'
Finally, contrary results were reported by Moreno30 who indicated that
air-cured specimens were about 10% stronger than moist-cured specimens
at all ages up to 91 days.
Specimen size
For most materials, including concrete, it has generally been observed that
the smaller the test specimen, the higher the strength. For high strength
concrete, however, though this effect is often observed, there are contra-
dictory results reported in the literature. The results of a number of studies
are compared in Table 1.11. It may be seen that the observed strength
ratios of 4 x 8 i n . (100x200 mm) cylinders to 6x12 in. (15Ox
300 mm)cylinders range from about 1.1 to 0.93. These contradictory results
may be due to differences in testing procedures amongst the various
investigators.
It must be noted that while for a given set of materials and test
procedures, it may be possible to increase the apparent concrete strength
by decreasing the specimen size, this does not in any way change the
strength of the concrete in the structure. One particular specimen size does
not give 'truer' results than any other. Thus, one should be careful to
specify a particular specimen size for a given project, rather than leaving it
as a matter of choice.
Table 1.11 Effect of specimen size on the compressive strength of high strength
concrete
1.5 Conclusions
In conclusion, then, it has been shown that the production of high strength
concrete requires careful attention to details. It also requires close
cooperation between the owner, the engineer, the suppliers and producers
of the raw materials, the contractor, and the testing laboratory.32 Perhaps
most important, we must remember that the well-known 'laws' and
'rules-of-thumb' that apply to normal strength concrete may well not apply
to high strength concrete, which is a distinctly different material. Nonethe-
less, we now know enough about high strength concrete to be able to
produce it consistently, not only in the laboratory, but also in the field. It is
to be hoped that codes of practice and testing standards catch up with the
high strength concrete technology, so that the use of this exciting new
material can continue to increase.
Acknowledgements
This work was supported by the Canadian Network of Centres of Excell-
ence on High-Performance Concrete.
References
1 SHRP-C/FR-91-103 (1991) High performance concretes, a state of the art
report. Strategic Highway Research Program, National Research Council,
Washington, DC.
2 FIP/CEB (1990) High strength concrete, state of the art report. Bulletin
d'Information No. 197.
3 ACI Standard 211.1 (1989) Recommended practice for selecting proportions for
normal weight concrete. American Concrete Institute, Detroit.
4 Mindess, S. and Young, J.F. (1981) Concrete. Prentice Hall Inc., Englewood
Cliffs.
5 Perenchio, W.F. (1973) An evaluation of some of the factors involved in
producing very high-strength concrete. Research and Development Bulletin,
No. RD014-01T, Portland Cement Association, Skokie.
6 Mehta, P.K. and Aitcin, P.-C. (1990) Microstructural basis of selection of
materials and mix proportions for high-strength concrete, in Second Interna-
tional Symposium on High-Strength Concrete, SP-121. American Concrete
Institute, Detroit, 265-86.
7 Aitcin, P.-C. (1992) private communication
8 Ronneburg, H. and Sandvik, M. (1990) High Strength Concrete for North Sea
Platforms, Concrete International, 12, 1, 29-34
9 CSA Standard A23.5-M86 (1986) Supplementary cementing materials. Cana-
dian Standards Association, Rexdale, Ontario.
10 ASTM C618 Standard specification for fly ash and raw or calcined natural
pozzolanfor use as a mineral admixture in portland cement concrete. American
Society for Testing and Materials, Philadelphia, PA.
11 ASTM C989 Standard specification for ground iron blast-furnace slag for use in
concrete and mortars. American Society for Testing and Materials, Phila-
delphia, PA.
12 Malhotra, V.M. (ed) (1987) Supplementary cementing materials for concrete.
Minister of Supply and Services, Canada.
13 Aitcin, P.-C., Sarkar, S.L., Ranc, R. and Levy, C. (1991) A High Silica
Modulus Cement for High-Performance Concrete, in S. Mindess (ed.),
Advances in cementitious materials. Ceramic Transactions 16, The American
Ceramic Society Inc., 102-21.
14 Malhotra, V.M., Ramachandran, V.S., Feldman, R.F. and Aitcin, P.-C.
(1987) Condensed silica fume in concrete. CRC Press Inc., Boca Ratan,
Florida.
15 Cook, I.E. (1989) 10,000 psi Concrete. Concrete International, 11, 10, 67-75.
16 Peter man, M. B. and Carrasquillo, R. L. (1986) Production of high strength
concrete. Noyes Publications, Park Ridge.
17 Randall, V.R. and Foot, K.B. (1989) High strength concrete for Pacific First
Center. Concrete International: Design and Construction, 11, 4, 14-16.
18 Aitcin, P.-C., Shirlaw, M. and Fines, E. (1992) High performance concrete:
removing the myths, in Concrescere, Newsletter of the High-Performance
Concrete Network of Centres of Excellence (Canada), 6, March.
19 Burg, R.G. and Ost, B.W. (1992) Engineering properties of commercially
available high-strength concretes. Research and Development Bulletin
RD104T, Portland Cement Association, Skokie.
20 Aitcin, P.-C. and Mehta, P.K. (1990) Effect of coarse aggregate type or
mechanical properties of high strength concrete. ACI Materials Journal,
American Concrete Institute, Detroit, 87, 2, 103-107.
21 ACI Committee 363 (1984) State-of-the-art report on high strength concrete
(ACI 363R-84). American Concrete Institute, Detroit.
22 Addis, B.H. (1992) Properties of High Strength Concrete Made with South
African Materials, Ph.D. Thesis, University of the Witwatersrand, Johannes-
burg, South Africa.
23 Addis, BJ. and Alexander, M.G. (1990) A method of proportioning trial
mixes for high-strength concrete, in ACI Sp-121, High strength concrete,
Second International Symposium, American Concrete Institute, Detroit, 287-
308.
24 Canadian Portland Cement Association (1991) Design and control of concrete.
Edition CPCA, Ottawa.
25 Fiorato, A.E. (1989) PCA research on high-strength concrete. Concrete
International, 11, 4, 44—50.
26 Hattori, K. (1979) Experiences with mighty superplasticizer in Japan, in ACI
SP-62, Superplasticizers in concrete, American Concrete Institute, Detroit,
37-66.
27 Suzuki, T. (1987) Experimental studies on high-strength superplasticized
concrete, in Utilization of high strength concrete, Symposium proceedings.
Stavanger, Norway: Tapis Publishers, Trondheim, 53-4.
28 Carrasquillo, R.C., Nilson, A.H. and Slate, P.O. (1981) Properties of high
strength concrete subject to short-term loads. Journal of American Concrete
Institute, 78, 3, 171-8.
29 Carrasquillo, P.M. and Carrasquillo, R.L. (1988). Evaluation of the use of
current concrete practice in the production of high-strength concrete. ACI
Materials Journal, 85, 1, 49-54.
30 Moreno, J. (1990) 225 W. Wacker Drive. Concrete International, 12, 1, 35-9.
31 Hester, W.T. (1980) Field testing high-strength concretes: a critical review of
the state-of-the-art. Concrete International, 2, 12, 27-38.
32 Howard, N.L. and Leatham, D.M. (1989) The production and delivery of
high-strength concrete. Concrete International, 11, 4, 26-30.
33 Aitcin, P.-C. (1989) Les essais sue les betons a tres hautes performances, in
Annales de L'Institut Technique du Batiment et des Travaux Publics, No. 473.
Mars-Avril. Serie: Beton 263, 167-9.
34 Carrasquillo, R.L., Slate, P.O. and Nilson, A.H. (1981) Microcracking and
behaviour of high strength concrete subjected to short term loading. American
Concrete Institute Journal, 78, 3, 179-86.
35 Richardson, D.N. (1990) Effects of testing variables on the comparison of
neoprene pad and sulfur mortar-capped concrete test cylinders. ACI Material
Journal, 87, 5, 489-95.
36 Carrasquillo, P.M. and Carrasquillo, R.L. (1988) Effect of using unbonded
capping systems on the compressive strength of concrete cylinders. ACI
Materials Journal, 85, 3, 141-7.
37 Boulay, C. (1989) La boite a sable, pour bien ecraser les betons a hautes
performances. Bulletin de Liaison des Laboratoires des Ponts et Chausses,
Nov/Dec.
38 Boulay, C,, Belloc, A., Torrenti, J.M. and De Larrard, F. (1989) Mise au
point d'un nouveau mode operatoire d'essai de compression pour les betons a
haute performances. Internal report, Laboratoire Central des Ponts et Chaus-
sees, Paris, December.
39 CTL Review (1992) Construction Technology Laboratories, Inc., Skokie,
Illinois, 15, 2.
2 Short term mechanical
properties
S H Ahmad
2.1 Introduction
Chapter 1 discussed the production of concrete and the effects of a large
number of constituent materials - cement, water, fine aggregate, coarse
aggregate (crushed stone or gravel), air and other admixtures on the
production process. Some quality control issues were also addressed. In
the present chapter, the mechanical properties of hardened concrete under
short term conditions or loadings are discussed.
Concrete must be proportioned and produced to carry imposed loads,
resist deterioration and be dimensionally stable. The quality of concrete is
characterized by its mechanical properties and ability to resist deteriora-
tion. The mechanical properties of concrete can be broadly classified as
short-term (essentially instantaneous) and long-term properties. Short-
term properties include strength in compression, tension, modulus of
elasticity and bond characteristics. The long-term properties include creep,
shrinkage, behavior under fatigue, and durability characteristics such as
porosity, permeability, freeze-thaw resistance and abrasion resistance. The
creep and shrinkage characteristics are discussed in Chapters, the be-
havior under fatigue and the bond characteristics is addressed in Chapter 4.
The important aspect of durability is presented in Chapter 5.
While information on high performance concretes (HPC) as defined in
Chapter 1 is scarce, there is a substantial body of information on the
mechanical properties of high strength concrete and additional information
is being developed rapidly. One class of high performance concretes are
the early strength concretes. The mechanical properties of these types of
high performance concretes are being investigated under the Strategic
Highway Research Program SHRP C-205 which is in progress at North
Carolina State University. Since high performance concretes typically have
low water/cementitious materials (w/c) ratios and high paste contents,
characteristics will in many cases be similar to those of high strength
concrete. Much of the discussion in this chapter will therefore concentrate
on high strength concretes.
A significant difference in behavior between the early strength and the
high strength concretes is in the relationship of compressive strength to
mechanical properties. Strength gain in compression is typically much
faster than strength gain in aggregate-paste bond, for instance. This will
lead to relative differences in elastic modulus and tensile strength of early
strength concretes and high strength concretes, expressed as a function of
compressive strength. The relationships of mechanical properties to 28-day
compressive strength developed in other studies cannot necessarily be
expected to apply to early strength concretes. The information developed
under the SHRP program will be useful to fill this knowledge gap.
2.2 Strength
The strength of concrete is perhaps the most important overall measure of
quality, although other characteristics may also be critical. Strength is an
important indicator of quality because strength is directly related to the
structure of hardened cement paste. Although strength is not a direct
measure of concrete durability or dimensional stability, it has a strong
relationship to the w/c ratio of the concrete. The w/c ratio, in turn,
influences durability, dimensional stability and other properties of the
concrete by controlling porosity. Concrete compressive strength, in par-
ticular, is widely used in specifying, controlling and evaluating concrete
quality.
The strength of concrete depends on a number of factors including the
properties and proportions of the constituent materials, degree of hydra-
tion, rate of loading, method of testing and specimen geometry.
The properties of the constituent materials which affect the strength are
the quality of fine and coarse aggregate, the cement paste and the
paste-aggregate bond characteristics (properties of the interfacial, or
transition, zone). These, in turn, depend on the macro- and microscopic
structural features including total porosity, pore size and shape, pore
distribution and morphology of the hydration products, plus the bond
between individual solid components. A simplified view of the factors
affecting the strength of concrete is shown in Fig. 2.1.
Testing conditions including age, rate of loading, method of testing, and
specimen geometry significantly influence the measured strength. The
strength of saturated specimens can be 15% to 20% lower than that of dry
specimens. Under impact loading, strength may be as much as 25% to 35%
higher than under a normal rate of loading (10 to 20 microstrains per
second). Cube specimens generally exhibit 20% to 25% higher strengths
than cylindrical specimens. Larger specimens exhibit lower average
strengths.
CONCRETE STRENGTH
Fig. 2.1 An oversimplified view of the factors influencing strength of plain concrete53
Coarse aggregate
The important parameters of coarse aggregate are its shape, texture and
the maximum size. Since the aggregate is generally stronger than the paste,
its strength is not a major factor for normal strength concrete, or in early
strength concrete. However, the aggregate strength becomes important in
the case of higher-strength concrete or lightweight aggregate concrete.
Surface texture and mineralogy affect the bond between the aggregates
and the paste and the stress level at which microcracking begins. The
surface texture, therefore, may also affect the modulus of elasticity, the
shape of the stress-strain curve and, to a lesser degree, the compressive
strength of concrete. Since bond strength increases at a slower rate than
compressive strength, these effects will be more pronounced in early
strength concretes. Tensile strengths may be very sensitive to differences in
aggregate surface texture and surface area per unit volume.
The effect of different types of coarse aggregate on concrete strength has
been reported in numerous articles. A recent paper12 reports results of
four different types of coarse aggregates in a very high strength concrete
mixture (w/c = 0.27). The results showed that the compressive strength
was significantly influenced by the mineralogical characteristics of the
aggregates. Crushed aggregates from fine-grained diabase and limestone
gave the best results. Concretes made from a smooth river gravel and from
crushed granite that contained inclusions of a soft mineral were found to be
relatively weaker in strength.
The use of larger maximum size of aggregate affects the strength in
several ways. Since larger aggregates have less specific surface area, the
bond strength between aggregates and paste is lower, thus reducing the
compressive strength. Larger aggregate results in a smaller volume of paste
thereby providing more restraint to volume changes of the paste. This may
induce additional stresses in the paste, creating microcracks prior to
application of load, which may be a critical factor in very high strength
concretes.
The effect of the coarse aggregate size on concrete strength was
discussed by Cook et al.22 Two sizes of aggregates were investigated: a
3/8 in. (10 mm) and a 1 in. (25 mm) limestone. A superplasticizer was used
in all the mixes. In general, the smallest size of the coarse aggregate
produces the highest strength for a given w/c ratio, see Figs 2.2-2.6. It may
be noted that compressive strengths in excess of 10,000 psi (70 MPa) can be
produced using a 1 in. (25 mm) maximum size aggregate when the mixture
is properly proportioned.
Although these studies12'22 provide useful data and insight, much more
research is needed on the effects of aggregate mineral properties and
Compressive strength, psi
w/c - 0.32
No superplasticizer
Water-cementitious ratio
Fig. 2.3 Effect of aggregate type on 56 day strength for concrete for different w/c materials
ratio22
Paste characteristics
The most important parameter affecting concrete strength is the w/c ratio,
Water-cementitious ratio
Fig. 2.4 Relationship of w/c materials ratio with and without a high range water-reducing
admixture for coarse aggregate size not exceeding | in. (10 mm)22
56-day compression strength, psi
Superplasticizer Fly ash
Water-reducer 1 inch limestone
Water-cementitious ratio
Fig. 2.5 Relationship of w/c materials ratio with and without a high range water-reducing
admixture for coarse aggregate size not exceeding 1 in. (25.4 mm)22
sometimes referred to as the w/b (binder) ratio. Even though the strength
of concrete is dependent largely on the capillary porosity or gel/space ratio,
these are not easy quantities to measure or predict.The capillary porosity
of a properly compacted concrete is determined by the w/c ratio and degree
of hydration. The effect of w/c ratio on the compressive strength is shown
in Fig. 2.7. The practical use of very low w/c ratio concretes has been made
possible by use of both conventional and high range water reducers, which
permit production of workable concrete with very low water contents.
Supplementary cementitious materials (fly ash, slag and silica fume)
have been effective additions in the production of high strength concrete.
Although fly ash is probably the most common mineral admixture, on a
volume basis, silica fume (ultra-fine amorphous silica, derived from the
production of silicon or ferrosilica alloys) in particular, used in combina-
56-day compression strength, psi
with superplasticizer
Water-cementitious ratio
Fig. 2.6 Effect of aggregate type on strength at different ages for a constant w/c materials
ratio, with superplasticizer22
silica fume Notes:
• Test ages 28 to 105 days
• 4 x 8 or 6 x 12-in. (102 x 203 or 152 x 305-mm) cylinders
• Moist curing at least 1 day
• All non-air-entrained concretes
• 1. Oksi = 6.895 MPa Ref. 15
Compressive strength, ksi
Ret. 13
Ref. 2
Ref. 3
Ref. 4
Ref. 9
Ref. 16
Ref. 7
'zero' slump Ref. 5
8% CSF
High performance
concrete
Compressive strength, MPa
16% CSF
Reference concrete
w/c
Fig. 2.8 28-day compressive strength versus w/c materials ratio for concrete with different
condensed silica fume contents28
required is about 0.35 if no silica fume is used; however, with 8% silica
fume, the w/c needed is about 0.50, and with 16% silica fume content the
w/c ratio requirement increases to about 0.65. This indicates that higher
compressive strength can be achieved very easily with high silica fume
content at relatively higher w/c ratios.
The efficiency of silica fume in producing concrete of higher strength
depends on water/cement + silica fume ratio, dosage of silica fume, age and
curing conditions. Yogenendram et al.S5 investigated the efficiency of silica
fume at lower w/c ratio. Their results indicated that the efficiency is much
lower at w/c ratio of 0.28 as compared to the efficiency at w/c ratio of 0.48.
The performance of chemical admixtures is influenced by the particular
cement and other cementitious materials. Combinations which have been
shown to be effective in many cases may not work in all situations, due to
adverse cement and admixture interaction (see Fig. 2.9). Substantial
testing should be conducted with any new combination of cements, and
mineral or chemical admixtures prior to large scale use.
Test age-days
Fig. 2.9 Effect of varying dosage rates of normal retarding water-reducing admixtures on the
strength development of concrete22
Compressive strength at 95 days
Compressive strength
High strength
Medium strength
Normal strength
Age, days
Fig. 2.10 Normalized strength gain with age for limestone concretes moist-cured until
testing16
Notes
Moist curing unless noted
1,000 psi = 6.895 MPa
Age, days
Fig. 2.11 Compressive strength development for concretes with and without high range
water reducers29
28-day strength
0% CSF
5% CSF
10% CSF
months
Age
Fig. 2.12 Compressive strength development of concrete cured at 20 0C with different
dosages of condensed silica fume48
P30
MP3010%flyash
MP3025%flyash
MP3015%slag
months
Age
Fig. 2.13 Development of compressive strength in reference concrete cured in water at 5 0C
for 28 days then at 20 0C. 100% represents 28-day strength at 20 0C for each cement type48
P30
MP3010%flyash
MP3025%flyash
MP3015%slag
months
Age
Fig. 2.14 Development of compressive strength in concrete containing 10% condensed silica
fume and cured in water at 5 0C for 28 days then at 20 0C. 100% represents 28-day strength at
20 0C for each cement type48
Relative compressive strength, %
MP3015%slag
28-day strength
at 2O0C
months
Age
Fig. 2.16 Development of compressive strength in concrete containing 10% condensed silica
fume and water-cured at 35 0C for 28 days then at 20 0C. 100% represents 28-day strength at
20 0C for each cement type48
Table 2.2 Effect of drying on compressive strength16
Compressive strength
Conventionally, in the USA, concrete properties such as elastic modulus,
tensile or flexural strength, shear strength, stress-strain relationships and
bond strength are usually expressed in terms of uniaxial compressive
strength of 6x12 in. (15Ox 300 mm) cylinders, moist cured to 28 days.
Compressive strength is the common basis for design for most structures,
other than pavements, and even then is the common method of routine
quality testing. The terms 'strength' and 'compressive strength' are
used virtually interchangeably. The discussion above generally applies
equally well to all measures of strength, although most results and con-
clusions were based either primarily or exclusively on compressive strength
results.
Maximum, practically achievable, compressive strengths have increased
steadily in the last decade. Presently, 28 day strengths of up to 12,000 psi
(84MPa) are routinely obtainable. The trend for the future has been
examined in a recent ACI Committee 363 article3 which identified develop-
ment of concrete with compressive strength in excess of 20,000psi
(138 MPa) as one of the research needs.
Testing variables have a considerable influence on the measured com-
pressive strength. The major testing variables are: mold type, specimen
size, end conditions and rate of loading. The sensitivity of measured
compressive strength to testing variables varies with level of compressive
strength.
Since the compressive strength of early strength concretes are at
conventional levels, conventional testing procedures can be used for the
most part, although curing during the first several hours can affect test
results dramatically. Testing of very high strength concretes is much more
demanding. However, in all concretes, not just high performance concrete,
competent testing is critical.
The effect of mold type on strength was reported in a recent paper by
Carrasquillo and Carrasquillo.18 Their results indicated that use of
6 x 12 in. (150 x 300 mm) plastic molds gave strengths lower than steel
molds, and use of 4 x 8 in. (102 x 203 mm) plastic molds gave negligible
difference wiih steel molds. They concluded that steel molds should be
used for concrete with compressive strengths up to 15,000 psi (103 MPa). It
seems appropriate that steel molds should also be used for concrete of
higher strengths. The specimen size effect on the strength is shown in
Fig. 2.17, which shows the relationship between the compressive strength
of 4 x 8 in. (102 x 203 mm) cylinders and 6 x 12 in. (150 x 300 mm) cylin-
ders. The figure indicates that 4 x 8 i n . (102x203 mm) cylinders exhibit
approximately 5% higher strengths than 6x12 in. (15Ox 300 mm) cylin-
ders. Similar results were also obtained in a recent study at North Carolina
Compressive strength, psi - 6x12 in. cylinders
Y = 0.965x-103
r = 0.995
n = 52
Sy = a y Vl -r 3
Sy = 292 psi
Age, days
Fig. 2.19 Compressive strength development for 17,000 psi (117 MPa) concrete54
Table 2.3 Comparison of compressive strength test results of 12,000 psi (83 MPa)
concrete at 56 days as obtained by 6 X 12 in. (152 x 304 mm) and 4 x 8 in. (102 x 203
mm) cylinders54
Cylinder size 6 x 12 in. 4 x 8 in.
Mean 13,444 13,546
Standard deviation 463 515
Coefficient of variation, percent 3.4 3.8
Number of tests 29 29
applicable to high strength concrete. This study also confirms the belief
that job cured specimens do not give accurate measurements of the
in-place strength. The reason for lower strength in the middle portion of
the columns is probably due to temperature rise, i.e. 1000F (38 0C) for high
strength mixtures. In a recent study,11 it was shown that the strength of
4 in. (100mm) cores taken from a mock column at two and a half years
after casting was nearly identical to that of specimens cured for 28 days in
lime-saturated water at room temperature. The strength of the concrete
tested was 12,300 psi (85 MPa).
The effect of end conditions on the compressive strength of concrete is
summarized in a recent paper.18 More than five hundred 6x12 in.
(150 x 300 mm) cylinders from concretes having compressive strengths
from 2500 to 16,500 psi (17MPa to 114MPa) were tested with either
unbonded caps (two types) or sulfur mortar caps. It was concluded that use
of unbonded caps (with a restraining ring and elastomeric insert) could
provide a cleaner, safer and more cost-effective alternative to sulfur mortar
for capping concrete cylinders. For concretes between 4000 and 10,000 psi
(28 and 69 MPa), the use of polyurethane inserts with aluminium restrain-
ing rings in testing concrete cylinders yielded average test results within 5%
of those obtained using sulfur mortar. For concrete strengths below
11,000 psi (76 MPa), the use of neoprene inserts with steel restraining rings
in testing concrete cylinders yielded average test results within 3% of those
obtained using sulfur mortar. For higher strength concrete, the use of
either unbonded capping system is questionable. Substantial differences in
compressive strength test results were obtained when two sets of restrain-
ing rings obtained from the same manufacturer were used. It was recom-
mended that prior to acceptance, each set should be tested for correlation
to results obtained from cylinders capped according to ASTM C617 for all
strength levels of concrete for which the unbonded caps are to be used.
(Equipment now exists for parallel grinding the ends of concrete cylinders
prior to compression testing, thereby eliminating the need for any type of
end cap.)
Measured compressive strength increases with higher rates of loading.
This trend has been reported in a number of studies14'15'26'38'49'83 for
concrete with strengths in the range of 2000 to 5500 psi (14 to 49MPa).
However, only one study8 has reported the effect of strain rate on
concretes with compressive strengths in excess of 6000 psi (41 MPa). Based
Table 2.4 Column core strengths versus 6 x 12 in. (152 x 304 mm) cylinders* 22
Test age 7-Days 28-Days 56-Days ISO-Days !Year
Maximum size stone Jin lin iin lin iin lin iin lin iin lin
Tensile strength
The tensile strength governs the cracking behavior and affects other
properties such as stiffness, damping action, bond to embedded steel and
durability of concrete. It is also of importance with regard to the behavior
of concrete under shear loads. The tensile strength is determined either by
direct tensile tests or by indirect tensile tests such as flexural or split
cylinder tests.
Strain rate
Fig. 2.20 Effect of strain rate on tensile strength of concrete59
Direct tensile strength data is not available for concrete with strengths in
excess of 8000 psi (55 MPa).
The effect of rate of loading on the tensile strength has been the focus of
some studies by Hatano,34 Suaris and Shah,80 and Zielinski et al.87 The
effect of fast strain rate on the tensile strength of concrete as observed by
these studies is shown in Fig. 2.20. Also shown in the figure is a comparison
of the predictions per a constitutive theory for concrete subjected to static
uniaxial tension59 and the experimental results.
The effect of sustained and cyclic loading on the tensile properties of
concrete was investigated by Cook and Chindaprasirt.21 Their results
indicate that prior loading of any form reduces the strength of concrete on
reloading. Strain at peak stress and the modulus on reloading follows the
same trend as strength. This behavior can be attributed to the cumulative
damage induced by repetitive loadings. Saito69 investigated the microc-
racking phenomenon of concrete understatic and repeated tensile loads,
and concluded that cumulative damage occurs in concrete due to reloading
beyond the stage at which interfacial cracks are formed.
The effect of uniaxial impact in tension was investigated by Zielinski et
al.87 Their results indicated an increase in the tensile strength similar to the
phenomenon generally observed under uniaxial impact in compression.
t;p = 7.40VfJ(Ref.4)
Tensile strength fJp (psi)
the splitting tensile strength (fct) for normal weight concrete can be
estimated by
fa = 7.4V/7 psi 3000 ^/c'sc 12,000 psi ( .
( }
(21 ^/c^ 83 MPa)
In 1985, based on the available experimental data of split cylinder tests
on concretes of low-, medium-,32-37'81 and high strengths,7'16'25 an empiri-
cal relationship was proposed by Ahmad and Shah4 as
fcf = 4.34 (£') psi fc' ^ 12,000 psi
}
(/ c '^83MPa) ^
Figure 2.21 shows the experimental data, with the predictions using the
above equation and the recommendations of the ACI Committee 363. The
latter appears to overestimate values of tensile strength. Recommenda-
tions of ACI Committee 363 were based on work performed at Cornell
University.16
Figure 2.22 shows the aging effect on splitting tensile strength, which is
similar to that under compressive loading. In an investigation by Ojdrovic
on cracking modes,60 it was concluded that at early ages, tensile strength of
concrete is the property of the matrix which governs the cracking mode.
The effect of prior compressive loading on the split tensile strength was
investigated by Liniers46 and the results are shown in Fig. 2.23. From this
figure, he concluded that limiting the compressive stresses to 60% of the
strength is essential if only tolerable damage is to be accepted.
The tensile strength of condensed silica fume (CSF) concrete is related
to the compressive strength in a manner similar to that of normal concrete.
Normalized toughness and strength
2a/D - 0.0
2a/D - 0.2
2a/D = 0.5
2a = Length of crack
D = Depth of the section
Days
Fig. 2.22 Normalized splitting tensile strength as a function of the age at testing60
However if CSF concrete is exposed to drying after one day of curing in the
mold, the tensile strength is reduced more than the control concrete.28
(b) Flexural strength or modulus of rupture Flexural strength or mod-
ulus of rupture is measured by a beam flexural test and is generally taken to
be a more reliable indicator of the tensile strength of concrete. The
modulus of rupture is also used as the flexural strength of concrete in
pavement design. It is often assumed that flexural strength of concrete is
about 15% of the compressive strength.
In the absence of actual test data, the modulus of rupture may be
estimated by
fr = kVJ7 (2.5)
typically in the range of 7.5 to 12. For high strength concrete, the ACI
Committee 363 State-of-the-Art Report2 recommends a value of k = 11.7 as
T = 1 min
T= 100 min
Tensile strength loss (%)
T= 10 000 min.
(7 days)
appropriate for concrete with compressive strength in the range of 3000 psi
to 12,000 psi (21 MPa to 83 MPa).
Based on the available data of beam flexural tests on concretes of low,
medium32'37'81 and high strengths,7'16'25 an empirical equation to predict
the flexural strength (modulus of rupture) was proposed4 as
/r = 2.30(/c')2/3 (2.6)
where fc' is the compressive strength in psi.
The above equation is of the same form as proposed by Jerome,39 which
was developed on the basis of data for concretes of strengths up to 8000 psi
(56MPa). Figure 2.24 shows the plot of the experimental data and the
proposed equation4 for predicting the modulus of rupture of concretes
with strengths up to 12,000 psi (83MPa). Also shown in the figure is the
expression recommended by Carrasquillo and Nilson.16
The results of uniaxial and biaxial flexural tests86 indicated that the
tensile strength was 38% higher in the uniaxial stress state than in the
biaxial stress state.
Flexural strength is higher for moist-cured as compared to field cured
specimens.19 However, wet-cured specimens containing condensed silica
fume (CSF) exhibit a lower ratio of tensile to compressive strength than
dry-stored concrete specimens with silica fume.47 For all concretes,
allowing a moist cured beam to dry during testing will result in lowered
measured strength, due to the addition of applied load and drying
shrinkage stresses on the tensile face. The flexural strength of condensed
silica fume (CSF) concrete is related to the compressive strength in a
manner similar to that of concrete without silica fume; however, if CSF
concrete is exposed to drying after only one day of curing in the mold, the
flexural strength reduces more than the control concrete.28
2.3 Deformation
The deformation of concrete depends on short-term properties such as the
static and dynamic modulus, as well as strain capacity. It is also affected by
time dependent properties such as shrinkage and creep.
Static modulus
The static modulus of elasticity can be expressed as secant, chord or
tangent modulus. According to the ACI Building Code (ACI-318-89),1 Ec,
the static, secant modulus of elasticity, is denned as the ratio of the stress at
45% of the strength to the corresponding strain. Static, chord modulus of
elasticity, as determined by ASTM C469, is defined as the ratio of the
difference of the stress at 40% of the ultimate strength and the stress at 50
millionths strain to the difference in strain corresponding to the stress at
40% of ultimate strength and 50 millionths strain.
At present there are two empirical relationships that can be used for
design when the static modulus of elasticity has not been determined by
tests. They are the ACI Code formula1
Ec = 33(O1-5 VfJ psi (2.7)
where o> = unit weight in pounds per cubic foot (pcf) and the formula
recommended by the ACI Committee 363 on High Strength Concrete2 for
concrete with unit weight of 145 pcf.
Ec= 1.0 XlO 6 + 40,000VfJ psi (2.8)
16
This formula is based on work performed at Cornell University.
Figure 2.25 shows the range of scatter of data with the predictions of the
ACI equation and the ACI Committee 363 equation. A third equation was
f£, MPa
Range for which ACI code
formula was derived
*№)"*«*
CO / x 10- MPa
3
Ec = o)2-5(v%)°-65psi
15
proposed
EC \(~)
equation (7a) €c = (40,000 V£ + i.oxio 6 )
(co/145)15psi
'Point spread
VfJ, psi
Fig. 2.25 Secant modulus of elasticity versus concrete strength4
Normal wt.
Static modulus of elasticity, Ex 106, psi
ff 50 ,psi
Fig. 2.26 Secant modulus of elasticity versus concrete strength for normal weight concrete22
Modulus of elasticity, million psi
Note:
•6x12-in. (152x305-mm) cylinders
Ref.6 • 1.0 ksi = 1,000 psi = 6.895 MPa
Ref.9
Ref.4
Material service
225W.Wacker
6
EC = (40,000 VfI+1.0x10
15
)
(wc/145) ' psi
Vfc.psi
Fig. 2.28 Secant modulus of elasticity variation with square root of the compressive
strength54
Maximum stress = 9400 psi Maximum stress = 9200 psi
at 7 days test age at 7 days test age
Stress (psi)
Dynamic modulus
The measurement of dynamic modulus corresponds to a very small
instantaneous strain. Therefore the dynamic modulus is approximately
equal to the initial tangent modulus. Dynamic modulus is appreciably
higher than the static (secant) modulus. The difference between the two
moduli is due in part to the fact that heterogeneity of concrete affects the
two moduli in different ways. For low, medium and high strength
concretes, the dynamic modulus is generally 40%, 30% and 20% respec-
tively higher than the static modulus of elasticity.53
Popovics66 has suggested that for both lightweight and normal weight
concretes, the relation between the static and dynamic moduli is a function
of density of concrete, just as is the case with relation between the static
modulus and strength.66 Popovics expressed Ec as a linear function of
EdlA/p where p is the density of concrete, and Ed is the dynamic modulus.
The ratio of static to dynamic modulus is also affected by the age at
testing as shown by Philleo65 in Fig. 2.30. The figure indicates that at early
ages (up to 6 months) the ratio of the two moduli increases from 0.4 to
about 0.8 and becomes essentially constant thereafter.
A typical relationship between the dynamic modulus determined by the
vibration of the cylinders and their compressive strength is shown in
Fig. 2.31. It has been reported by Sharma and Gupta77 that the rela-
tionship between the strength and the dynamic modulus is unaffected by
air entrainment, method of curing, condition at test, or type of cement.
Ratio at moduli
106psi
Compressive strength - MPa
Fig. 2.31 Relation between the dynamic modulus of elasticity, determined by transverse
vibration of cylinders, and their compressive strength77
Diabase fc range
Limestone,
GC range
Gravel
Stress (psi)
Stress (MPa)
'Granite
Complete stress/strain
at28d
E range
200 microstrain
Microstrain
Fig. 2.32 Effect of the aggregate type on the ascending portion of the stress-strain curves of
concrete at 28-days12
analysis and that it can have considerable influence on the predicted
structural behavior and strength.67
To obtain the descending part of the stress-strain curve, it is necessary to
avoid specimen-testing machine interaction. One approach is to use a
closed-loop system with a constant rate of axial strain as a feedback signal
for closed-loop operation. The difficulties of obtaining the postpeak
behavior experimentally and methods of overcoming these difficulties are
Present study
Karr, Hanson and Capell
Nilson and Slate
Ahmad and Shah (Ref. 5)
Wischers
Stress (ksi)
Stress (MPa)
A = E0^7 (2.13)
JC
the application of the load, steel platens were glued to the top and bottom
of the specimens. In order to provide plane-parallel and axial connection of
these platens, a special gluing press was designed. Some 500 direct tensile
tests, 300 compressive and 300 splitting tests were performed. A typical
stress-strain curve with a 95% confidence region for concrete subjected to
direct tension is shown in Fig. 2.34. The stress-strain curve shown in the
figure is for dry specimens. The results may vary slightly for specimens
tested in moist conditions.
A study at Northwestern by Gopalaratham and Shah31 points out that
due to the localized nature of the post-cracking deformations intension, no
unique tensile stress-strain relationship exists. According to this study, the
uniaxial tensile strength can be estimated by V^7, and the tangent
modulus of elasticity is identical in tension and compression. The stress-
strain relationship in tension before peak is less nonlinear than in com-
pression.
Laser speckle interferometry was employed in a recent study,13 to
investigate the behavior of concrete subjected to uniaxial tension. Unique
post-peak stress-strain and stress-deformation behavior were not observed.
The stress-strain response of concrete was found to be sensitive to
gauge-length. Strains measured within a gauge length inside the microc-
racking zone were two orders of magnitude higher than values previously
reported.27
In a recent study,20 it was shown that while the use of strain gauges
would lead to non-objective constitutive stress-strain relations, interfer-
ometric measurements on notched specimens allow an indirect determina-
tion of the local stress-strain and stress-separation (deformations) rela-
tions. Guo and Zhang33 tested 29 specimens in direct tension and obtained
complete stress-deformation curves. Based on the experimental results an
equation was also derived for the stress-displacement curves.
Flexural tension
While the information on the stress-strain behavior in tension is severely
limited, virtually no data are available regarding the strain capacity in
flexural tension. This is an area for which research is sorely needed to
provide a basis for design where flexural cracking is an important consid-
eration.
Acknowledgements
This work was supported by the Strategic Highway Research Program
(SHRP) at North Carolina State University, Raleigh, NC.
References
1 Committee 318 (1989) Building code requirements for reinforced concrete.
American Concrete Institute.
2 Committee 363 (1984) State-of-the-art report on high strength concrete. ACl
Journal, 81, 4, July-Aug, 364-411.
3 Committee 363 (1987) Research needs for high-strength concrete. ACI
Materials Journal, November-December, 559-61.
4 Ahmad, S.H. and Shah, S.P. (1985) Structural properties of high strength
concrete and its implications for precast prestressed concrete. PCI Journal,
Nov-Dec, 91-119.
5 Ahmad, S.H. (1981) Properties of confined concrete subjected to static and
dynamic loading. Ph.D. Thesis, University of Illinois, Chicago, March.
6 Ahmad, S.H. and Shah, S.P. (1982) Stress-strain curves of concrete confined
by spiral reinforcement. ACI Journal, 79, 6, Nov-Dec, 484-90.
7 Ahmad, S.H. (1982) Optimization of mix design for high strength concrete.
Report No. CE 001-82, Department of Civil Engineering, North Carolina
State University, Raleigh, NC.
8 Ahmad, S.H. and Shah, S.P. (1985) Behavior of hoop confined concrete under
high strain rates. ACI Journal, 82, 5, Sept-Oct, 634-47.
9 Ahmad, S.H. and Shah, S.P. (1979) Complete stress-strain curves of concrete
and nonlinear design. Progress Report PFR 79-22878 to the National Science
Foundation, University of Illinois, Chicago, August and also Non-linear design
of concrete structures. University of Waterloo Press, 1980, 222-30.
10 Aitcin, P.-C. (1983) Condensed silica fume. University of Sherbrooke, Sher-
brooke, Quebec.
11 Aitcin, P.-C., Sarkar, S.L. and Laplante, P. (1990) Long-term characteristics
of a very high strength concrete. Concrete International, Jan, 40-4.
12 Aitcin, P.-C. and Metha, P.K. (1990) Effect of coarse-aggregate characteris-
tics on mechanical properties of high-strength concrete. ACI Materials Jour-
nal, March-April, 103-107.
13 Ansari, F. (1987) Stress-strain response to microcracked concrete in direct
tension. ACI Materials Journal, 84, 6, Nov-Dec, 481-90.
14 Atchley, B.L. and Furr, H.L. (1967) Strength and energy absorption capabili-
ties of plain concrete under dynamic and static loadings. ACI Journal, 64, 11,
Nov, 745-56.
15 Bresler, B. and Bertero, V.V. (1975) Influences of high strain rate and cyclic
loading on behavior of unconfined and confined concrete in compression.
Proceedings, Second Canadian Conference on Earthquake Engineering, Mac-
Master University, Hamilton, Ontario, Canada, June.
16 Carrasquillo, R.L., Nilson, A.H. and Slate, P.O. (1981) Properities of high
strength concrete subject to short-term loads. ACI Journal, May-June, 171-8.
17 Carrasquillo, R.L., Slate, P.O. and Nilson, A.H. (1981) Microcracking and
behaviour of high strength concrete subjected to short term loading. ACI
Journal, 78, 3, May-June, 179-86.
18 Carrasquillo, P.M. and Carrasquillo, R.L. (1988) Effect of using unbonded
capping systems on the compressive strength of concrete cylinders. ACI
Materials Journal, May-June, 141-7.
19 Carrasquillo, P.M. and Carrasquillo, R.L. (1988) Evaluation of the use of
current concrete practice in the production of high-strength concrete. ACI
Materials Journal, 85, 1, Jan-Feb, 49-54.
20 Cedolin, L., PoIi, S.D. and Iori, I. (1987) Tensile behavior of concrete.
Journal of Engineering Mechanics, ASCE, 113, 3, March, 431-49.
21 Cook, DJ. and Chindaprasirt, P. (1981) Influence of loading history upon the
tensile properties of concrete. Magazine of Concrete Research, 33, 116, Sept,
154-60.
22 Cook, J.E. (1989) Research and application of high-strength concrete:
10,000 PSI concrete. Concrete International, Oct, 67-75.
23 Cornelissen, H. A.W. (1984) Fatigue failure of concrete in tension. Heron, 29,
4, 1-68.
24 Cornelissen, H.A.W. and Reinhardt, H.W. (1984) Uniaxial tensile fatigue
failure of concrete under constant-amplitude and programme loading. Maga-
zine of Concrete Research, 36, 129, Dec, 216-26.
25 Dewar, J.D. (1964) The indirect tensile strength of concrete of high compress-
ive strength. Technical Report, No. 42.377. Cement and Concrete Association,
Wexham Springs, England, March.
26 Dilger, W.H., Koch, R. and Andowalczyk, R. (1984) Ductility of plain and
confined concrete under different strain rates. ACI Journal, 81, 1, Jan-Feb,
73-81.
27 Evans, R.H. and Marathe, M.S. (1968) Microcracking and stress-strain curves
for concrete in tension. Materials and Structures, Research and Testing,
(RILEM, Paris), 1, 1, Jan-Feb, 61-4.
28 FIP Commission on Concrete (1988) Condensed Silica Fume in Concrete. State
of Art Report, Federation Internationale de Ia Precontrainte, London.
29 Fiorato, A.E. (1989) PCA research on high-strength concrete. Concrete
International, April, 44-50.
30 Freedman, S. (1970/71) High strength concrete. Mocern Concrete, 34, 6, Oct
1970, 29-36; 7, Nov 1970, 28-32; 8, Dec 1970, 21-24; 9, Jan 1971, 15-22; and
10, Febl971, 16-23.
31 Gopalaratham, V.S. and Shah, S.P. (1985) Softening response of concrete
in direct tension. Research Report, Technological Institute, Northwestern
University, Chicago, June 1984 also ACI Journal, 82, 3, May-June 1985,
310-23.
32 Grieb, W.E. and Werner, G. (1962) Comparison of splitting tensile strength of
concrete with flexural and compressive strengths. Public roads, 32, 5, Dec.
33 Guo, Z-H. and Zhang, X-Q. (1987) Investigation of complete stress-
deformation curves for concrete in tension. ACI Materials Journal, 84, 4,
July-Aug, 278-85.
34 Hatano, T. (1960) Dynamic behavior of concrete under impulsive tensile load.
Technical Report, No. C-6002, Central Research Institute of Electric Power
Industry, Tokyo, 1-15.
35 Helland, S. et al (1983) Hoyfast betong. Presented at Norsk Betongdag,
Trondheim, Oct. (In Norwegian).
36 (1977) High-strength concrete in Chicago, high-rise buildings. Task Force
Report No. 5, Chicago Committee on High-Rise Buildings, Feb.
37 Houk, H. (1965) Concrete aggregates and concrete properties investigations.
Design Memorandum, No. 16, Dworshak Dam and Reservoir, U.S. Army
Engineer District, Walla, WA.
38 Hughes, B.P. and Gregory, R. (1972) Concrete subjected to high rates of
loading and compression. Magazine of concrete Research, 24, 78, London,
March.
39 Jerome, M.R. (1984) Tensile strength of concrete. ACI Journal, 81, 2,
March-April, 158-65.
40 Johansen, R. (1979) Silicastov Iabrikksbetong. Langtidseffekter. Report,
STF65 F79019, FCB/SINTEF, Norwegian Institute of Technology, Tron-
dheim. (In Norwegian), and Johansen R, (1981) Report 6: Long-term effects.
Report, STF65 A81031, FCB/SINTEF, Norwegian Institute of Technology,
Trondheim, 1981.
41 Kaar, P.M., Hanson, N.W. and Capell, H.T. (1977) Stress-strain characteris-
tics of high strength concrete. Research and Development Bulletin, RD051-
01D, Portland Cement Association, Skokie, Illinois, 11 pp. also Douglas
McHenry International Symposium on Concrete and Concrete Structures, ACI
special publication, SP-55, Detroit 1978, 161-85.
42 Klieger, P. (1958) Effect of mixing and curing temperatures on concrete
strength. ACl Journal, 54, 12, June, 1063-81.
43 Klink, S.A. (1985) Actual elastic modulus of concrete. ACI Journal, Sept-
Oct, 630-3.
44 Kotsovos, M.D. (1983) Effect of testing techniques on the post-ultimate
behaviour of concrete in compression. Materiaux et Constructions (RILEM,
Paris), Jan-Feb, 16, 91, 3-12.
45 Leming, M.L. (1988) Properties of high strength concrete: an investigation of
high strength concrete characteristics using materials in North Carolina.
Research Report FHWA/NC/88-006, Department of Civil Engineering, North
Carolina State University, Raleigh, N.C., July.
46 Liniers, A.D. (1987) Microcracking of concrete under compression and its
influence on tensile strength. Materiaux et Constructions (RILEM, Paris), 20,
116, Mar, 111-16.
47 Loland, K.E. and Gj0rv, O.E. (1981) Silikabetong. Nordisk Belong, 6, 1-6 (In
Norwegian).
48 Maage, M. and Hammer, T. A. (1985) Modifisert Portlandsement. Detrapport
3. Fasthetsutvikling Og E-Modul. Report, STF65 A85041, FCB/SINTEF,
Norwegian Institute of Technology, Trondheim (In Norwegian).
49 Mainstone, RJ. (1975) Properties of materials at high rates of straining or
loading. Materieaux et Constructions (Rilem, Paris), 8, 44, March-April.
50 Malhotra, H.L. (1956) The effect of temperature on the compressive strength
of concrete. Magazine of Concrete Research, V. 8, No. 23, Aug, 85-94.
51 Malhotra, V.M. (ed.) (1983) Fly ash, silica fume, slag and other mineral
by-products in concrete, SP 79, Vol. II, American Concrete Institute.
52 Malhotra, V.M. (ed.) (1986) Fly ash, silica fume, slag and naturalpozzolans in
concrete, SP 91, Vol. II, American Concrete Institute.
53 Metha, P.K. (1986) Concrete structures, properties and materials. Prentice
Hall, Inc, Englewood Cliffs, New Jersey.
54 Moreno, J. (1990) The state of the art of high-strength concrete in Chicago:
225 W. Wacker Drive. Concrete International, Jan, 35-9.
55 Nawy, E.G. Reinforced concrete: a fundamental approach, Second edition.
Prentice Hall, New Jersey.
56 Neville, A.M. (1981) Properties of concrete, Third edition. Pitman Publishing
Ltd., London.
57 Ngab, A.S., Nilson, A.H. and Slate, F.O. (1981) Shrinkage and creep of high
strength concrete. ACI Journal, 78, 4, July-Aug, 255-61.
58 Nilson, A.H. and Slate, P.O. (1979) Structural design properties of very high
strength concrete. Second Progress Report, NSF Grant ENG 7805124, School
of Civil and Environmental Engineering, Cornell University, Ithaca, New
York.
59 Oh, B.H. (1987) Behavior of concrete under dynamic tensile loads. ACI
Materials Journal, Jan-Feb, 8-13.
60 Ojorv, O.E. (1988) High strength concrete. Nordic Betong, 32, 1, Stockholm,
5-9.
61 Parrot, LJ. (1969) The properties of high-strength concrete. Technical Report
No. 42.417, Cement and Concrete Association, Wexham Springs, 12 pp.
62 Pentalla, V. (1987) Mechanical properties of high strength concretes based on
different binder compositions. Proceedings Symposium on Utilization of High
Strength Concrete, Norway, June 15-18, 123-34.
63 Perenchio, W.F. (1973) An evaluation of some of the factors involved in
producing very high-strength concrete. Research and Development Bulletin,
No. RD014.01T, Portland Cement Association, Skokie.
64 Perenchio, W.F. and Klieger, P. (1978) Some physical properties of high
strength concrete. Research and Development Bulletin, No. RD056.01T,
Portland Cement Association, Skokie, IL.
65 Philleo, R.E. (1955) Comparison of results of three methods for determining
Young's Modulus of elasticity of concrete. ACl Journal, 51, Jan, 461-9.
66 Popovics, S. (1975) Verification of relationships between mechanical prop-
erties of concrete-like materials. Materials and Structures (RILEM, Paris), 8,
45, May-June, 183-91.
67 Pramono, E. (1988) Numerical simulation of distributed and localized failure
in concrete. Structural Research Series, No. 88-07, C.E.A.E. Department,
University of Colorado, Boulder.
68 Ronne, M. (1987) Effect of condensed silica fume and fly ash on compressive
strength development of concrete. American Concrete Institute, SP-114-8,
175-89.
69 Saito, M. (1987) Characteristics of microcracking in concrete under static and
repeated tensile loading. Cement and Concrete Research, 17, 2, March,
211-18.
70 Sargin, M. (1971) Stress-strain curves relationships for concrete and analysis of
structural concrete sections. Study No. 4, Solid Mechanics Division, University
of Waterloo, Ontario, Canada.
71 Saucier, K.L. (1984) High strength concrete for peacekeeper facilities. Final
Report, Structures Laboratory, U.S. Army Engineer Waterways Experimental
Station, March.
72 Saucier, K.L., Tynes, W.O. and Smith, E.F. (1965) High compressive strength
concrete-report 3, Summary Report. Miscellaneous Paper No. 6-520, U.S.
Army Engineer Waterways Experiment Station, Vicksberg, Sept.
73 Sellevold, EJ. and Radjy, F.F. (1983) Condensed silica fume (microsilica) in
concrete: water demand and strength development. Publication SP-79, Amer-
ican Concrete Institute, Vol. II, 677-94.
74 Shah, S.P., Gokos, U.N. and Ansari, F. (1981) An experimental technique for
obtaining complete stress-strain curves for high strength concrete. Cement,
Concrete and Aggregates, CCAGDP, 3, Summer.
75 Shah, S.P., Fafitis, A. and Arnold, R. (1983) Cyclic loading of spirally
reinforced concrete. Journal of Structural Engineering, ASCE, 109, ST7, July,
1695-710.
76 Shah, S.P. and Sankan, R. (1987) Internal cracking and strain suffering
response of concrete under uniaxial compression. ACI Materials Journal, 84,
May-June, 200-12.
77 Sharma, M.R. and Gupta, B.L. Sonic modulus as related to strength and static
modulus of high strength concrete. Indian Concrete Journal, 34, 4, 139-41.
78 Skalny, J. and Roberts, L.R. (1987) High-strength concrete. Ann. Rev. Mater.
ScL, 17, 35-56.
79 Smith, S.S., William, K.J., Gerstle, K.H. and Sture, S. (1989) Concrete over
the top, or: is there life after peak? ACI Materials Journal, 86, 5, Sept-Oct,
491-7.
80 Suaris, W. and Shah, S.P. (1983) Properties of concrete subjected to impact.
Journal of Structural Engineering, ASCE, 109, 7, July, 1727-41.
81 Walker, S. and Bloem, D.L. (1960) Effects of aggregate size on properties of
concrete. ACI Journal, 32, 3, September, 283-98.
82 Wang, P.T., Shah, S.P. and Naaman, A.E. (1978) Stress-strain of normal and
lightweight concrete in compression. ACI Journal, 75, 11, November, 603-11.
83 Watstein, D. Effect of straining rate on the compressive strength and elastic
properties of concrete. ACI Journal, 49, 8, April, 729-44.
84 Wischers, G. (1979) Application and effects on compressive loads on concrete.
Betontechnische Berichte, 1978, Betonverlag Gmbh, Dusseldorf, 31-56.
85 Yogenendran, V., Langan, B.W. and Ward, M.A. (1987) Utilization of silica
fume in high strength concrete. ACI Materials Journal, 84, 2, March-April,
85-97, also Proceedings Symposium on Utilization of High Strength Concrete,
Norway, June 15-18, 85-97.
86 Zielinski, Z.A. and Spiropoulos, I. (1983) An experimental study on the
uniaxial and biaxial flexural tensile strength of concrete. Canadian Journal of
Civil Engineering ,10,104-15.
87 Zielinski, A.J., Reinhardt, H.W. and Kormeling, H.A. (1981) Experiments
on concrete under uniaxial impact tensile loading. Materiaux et Constructions
(RILEM, Paris), 14, 80, Mar-Apr, 103-12.
88 Zielinski, A.J., Reinhardt, H.W. and Kormeling, H.A. (1981) Experiments
on concrete under repeated uniaxial impact tensile loading. Materiaux et
Constructions (RILEM, Paris), 14, 81, May-June, 163-9.
3 Shrinkage creep and
thermal properties
F de Larrard, P Acker and R Le Roy
3.1 Introduction
Shrinkage and creep are time-dependent deformations that, along with
cracking, provide the greatest concerns for the designers because of
uncertainty associated with their prediction. Concrete exhibits elastic
deformations only under loads of short duration and, due to additional
deformation with time, the effective behavior is that of an inelastic and
time-dependent material. A quantitative knowledge of mechanical be-
havior, including delayed deformations and thermal effects, is necessary
for a number of structures: bridges, buildings etc. In other cases, control of
short- and long-term cracking requires an accurate modelling of strains and
stresses at all ages of the structure.
High performance concretes (HPC) are concretes with attributes (in-
cluding creep, shrinkage and thermal effects) that are superior to conven-
tional concretes. It is generally recognized that the higher the compressive
strength of concrete, the better the other attributes. For the purpose of this
chapter, high strength concrete (HSC) is defined as concrete with 28 day
compressive strength in excess of 6000 psi (42MPa). With the increasing
use of concretes of higher strengths, the knowledge of time dependent
behavior and the thermal properties is becoming important to ascertain the
long term behavior of structures utilizing high strength concrete.
In this chapter, the mechanisms of shrinkage and creep are presented,
along with experimental data which enunciate the effects of different
parameters on the time dependent properties of concrete. The thermal
properties are summarized. Also some field case histories are presented in
which concretes to meet the higher performance requirements with special
attention to shrinkage, creep and strength characteristics were designed
and used. In these field cases, thermal effects and long-term strains have
been monitored, and compared with predictions based on laboratory data.
3.2 Shrinkage
Shrinkage is the decrease of concrete volume with time. This decrease is
due to changes in the moisture content of the concrete and physio chemical
changes, which occur without stress attributable to actions external to the
concrete. Swelling is the increase of concrete volume with time. Shrinkage
and swelling are usually expressed as a dimensionless strain (in./in. or
mm/mm) under given conditions of relative humidity and temperature.
Shrinkage is primarily a function of the paste, but is significantly influenced
by the stiffness of the coarse aggregate. The interdependence of many
factors creates difficulty in isolating causes and effectively predicting
shrinkage without extensive testing. The principal variables that affect
shrinkage are summarized in Table 3.1.1 The key factors affecting the
magnitude of shrinkage are:
Aggregate
The aggregate acts to restrain the shrinkage of cement paste; hence
concrete with higher aggregate content exhibits smaller shrinkage. In
addition, concrete with aggregates of higher modulus of elasticity or of
rougher surfaces is more resistant to the shrinkage process.
Member size
Both the rate and the total magnitude of shrinkage decrease with an
increase in the volume of the concrete member. However, the duration of
shrinkage is longer for large members since more time is needed for
shrinkage effects to reach the interior regions.
Cement type
The effects of cement type are generally negligible except as rate-of-
strength-gain changes. Even here the interdependence of several factors
make it difficult to isolate causes. Rapid hardening cement gains strength
more rapidly than ordinary cement but shrinks somewhat more than other
types, primarily due to an increase in the water demand with increasing
fineness. Shrinkage compensating cements can be used to minimize or
eliminate shrinkage cracking if they are used with restraining reinforce-
ment.
Carbonation
Carbonation shrinkage is caused by the reaction between carbon dioxide
(CO2) present in the atmosphere and calcium hydroxide (CaOH2) present
in the cement paste. The amount of combined shrinkage varies according
to the sequence of occurrence of Carbonation and drying process. If both
phenomena take place simultaneously, less shrinkage develops. The
process of carbonation, however, is dramatically reduced at relative
humidities below 50%.
The effect of the aggregate content and the w/c ratio on the shrinkage
Aggregate
content by
Shrinkage x 10 6(in./in.)
volume percent
Water/cement ratio
Fig. 3.1 Effect of w/c materials ratio and aggregate cement on shrinkage23
Sedimentation
Before setting, concrete constituents are in suspension and, in certain cases
nonoptimized packing of granular skeleton. A vertical displacement of the
constituents occurs by gravity:10 downward for the larger grains, upward
for the water which entraps the elements having a lower sedimentation
velocity; a film of very clean water appears at the top surface (bleeding).
Le Chatelier's contraction
The volume of hyrdrates formed in the hydration reaction is substantially
smaller than the sum of the volumes of the two components (anhydrous
cement and water) entering into the reaction. The range is generally
between 8 to 12%, depending on the properties of the cement. The
potential lineic shrinkage of the cement paste is assumed to be about 3 to
4% (these values cannot be produced experimentally, due to the mechanic-
al stiffness of the hardened paste itself).
Heat of hydration
The hydration reaction is highly exothermic (150 to 350 joules per gram of
cement) which elevates the temperature under adiabatic conditions to
values between 25 and 55 K in the concrete.
Self-desiccation
Only a fraction of the hydration is completed during the setting process.
The hydration continues inside a rigid porous skeleton, resulting in a
reduction in the water content in the pore space, a reduction that has the
same mechanical effect as drying.11
Desiccation
This occurs in ordinary concrete, when about twice as much water as is
strictly necessary for the hydration of the cement is used for reasons of
workability, whereas this is not the case for high performance concretes
(HPC). After demolding, a drying process begins from the surfaces in
contact with the ambient atmosphere.12
Heat development
The setting of cement is a highly exothermic chemical process which
generates about 150 to 350 Joules per gram of cement. This generates to a
temperature rise of between 25 and 55 K in the concrete under adiabatic
conditions. Setting may then occur (in the core of massive parts, for
instance) at a higher temperature and the differential contractions that
occur during the subsequent cooling may lead to very significant
cracking.18 Hydration heat in concrete is roughly proportional to the
amount, fineness and chemical composition (C3A content) of the cement
used. It increases with the amount, the fineness of the cement and the C3A
content in the cement. For the same cement content, less water leads to a
reduction in the degree of hydration and an increase in the tensile strength
of the hardened concrete.
Heat of hydration is affected not only by the cement content, but also by
the water/cementitious ratio w/(c + s) and the silica fume content.19
Smeplass and Maage investigated the heat of hydration for a number of
high strength concrete mixes by the means of a so-called semi-adiabatic
calorimeter test. The results of Smeplass and Maage19 indicate that the
heat of hydration can be affected within a relatively wide range by the
utilization of traditional mix-design parameters (Figs. 3.2 and 3.3). The
heat of hydration per cement unit decreases approximately by 9% when
the w/(c + s) ratio is reduced from 0.36 to 0.27. At a constant paste/
aggregate ratio, the resulting temperature rise increases by 6% (30C).
Accumulated heat (semi-adiabetic)(kJ/kg cm)
ND 65
ND 95
ND115
Maturity (hours)
Fig. 3.2 Semi-isothermal heat development of three high strength concretes with different
values of w/(c + s) ratio: 0.50 (ND65), 0.36 (ND95) and 0.27 (ND115)
Consequently, the effect of an increase in cement content normally is
stronger than the effect of a reduced degree of hydration.
Smeplass and Maage19 also found that at w/(c + s) ratio of 0.50, the
replacement of cement by silica fume on a 1:1 basis induces a signifcant
increase in the heat evolved per cement unit. The use of silica fume reduces
the retarding effect of the dispersing agent, but also the magnitude of the
secondary accelerating effect. The effect is observed for w/(c + s) ratios
ranging from 0.50 to 0.27. At w/(c + s) ratios about 0.50 the effect of the
silica fume on the total heat evolvement is just as strong as the effect of the
cement. The effect weakens with decreasing w/(c + s) ratio. At w/(c + s)
ratio 0.27, silica fume does not affect the heat evolvement significantly.
Below w/(c + s) ratio 0.40, cement replacements by silica fume result in a
reduced 'final' semi-adiabatic temperature.
ND115
Semi-adiabetic temperature (0C)
ND 95
ND 65
Time (hours)
Fig. 3.3 Semi-adiabatic temperature development for three high strength concretes with
different values of w/(c + s) ratio: 0.50 (ND65), 0.36 (ND95) and 0.27 (NDl 15)
From these results, it appears that HSC is not necessarily more sensitive
to thermal cracking under all conditions. Furthermore, it should be noticed
that thermal effects are not significant for concrete thickness less than 1.6 ft
(40 cm). This is because of a drastic size effect due to the fact that the rate
of heat diffusion increases like the square of the thickness (and this effect is
accentuated by thermo-activation).20 For specific industrial applications
more sensitive to thermal cracking, like reservoirs, appropriate HSC
mix-designs may be utilized with relatively low heat of hydration. 17
Figure 3.4 shows that for a given temperature rise, the setting occurs
sooner in high strength concretes than in normal strength concretes (NSC).
Hardening kinetics
In order to obtain a parametric description of the mechanical behavior
such as strength, modulus of elasticity, autogenous shrinkage and basic
creep of high strength concrete, a total of ten mixes were examined by de
Larrard and Le Roy.21 The variables were the water/cement ratio (w/
c = 0.28, 0.33, 0.38, 0.42), the paste volume ratio (Vp = 0.269, 0.286,
0.313, 0.325), and the silica fume ratio (s/c - 0.00, 0.05, 0.10, 0.15). The
mix proportions and the results are given in Table 3.2.21
The hardening kinetics during early age and up to 28 days were
investigated by studying the effect of aging (between 1 and 28 days) on the
compressive strength and the modulus of elasticity for the ten HSC mixes
and control reference mix [Figs. 3.5 and 3.6]. From these figures, it clearly
appears that the hardening kinetics of high strength concretes significantly
differs from that of ordinary concrete during the early age (up to 1 day),
because of the reduction of water/cement ratio, and after one day, due to
the hydration process of silica.
Autogenous shrinkage
The high autogenous (or self-desiccation) shrinkage of silica-fume HSC
was first pointed out by Paillere, Buil and Serrano.22 A very early
shrinkage, even for non-drying specimens, was monitored in free speci-
mens, leading to early cracking in restrained ones. More recently, de
Larrard and Le Roy21 have carried out measurements of autogenous
shrinkage on ten HSC mixes (Table 3.2), and they proposed a relationship
between mix-design and autogenous shrinkage, for HSC in the range of 60
to 100 MPa.23 The final form of the equation can be written as
^(l + EJEg}(l-glg*) + [4(l-g*}g/g*}l[(l-2g*}+g*EgIEd]
\+glg* + (l-glg*)EdIEg
(l-0.65exp[-lls/c])(l-1.43w/c)7C c
^c ^ j
Table 3.2 Compositions and results of tests on 10 HSC mixes. Concrete C0 is a control concrete, without admixture nor silica fume
Formulae lb/yd3 (kg/m3) C0 Ci C2 C3 C4 C5 C6 C7 C8 C9 CIQ
Coarse aggregate 2022 2049 2088 1975 1914 2029 2022 2025 2022 2022 2022
(1200) (1216) (1239) (1172) (1136) (1204) (1200) (1202) (1200) (1200) (1200)
Sand 1129 1127 1144 1083 1050 1112 1109 1110 1109 1109 1109
(670) (669) (679) (643) (623) (660) (658) (659) (658) (658) (658)
Cement 576 671 617 711 770 598 723 629 718 694 650
(342) (398) (366) (422) (457) (355) (429) (373) (426) (412) (386)
Silica Fume O 67.0 61.7 71.1 77.0 59.8 72.3 62.9 O 34.7 97.6
(O) (39,8) (36,6) (42,2) (45,7) (35,5) (42,9) (37,3) (O) (20,6) (57,9)
Superplast.* O 32.5 30.0 34.5 37.4 29.0 33.9 30.5 34.9 33.7 31.7
(O) (19,3) (17,8) (20,5) (22,2) (17,2) (20,1) (18,1) (20,7) (20,0) (18,8)
Added water 288 199 182 211 229 231 179 217 212 206 192
(171) (118) (108) (125) (136) (137) (106) (129) (126) (122) (114)
Density 2,36 2,43 2,45 2,42 2,42 2,41 2,43 2,41 2,43 2,43 2,43
Entrap, air % 1,9 0,6 1,2 0,7 0,4 0,7 0,9 0,5 1,2 0,8 0,6
Slump in (mm) 2.4 7.9 7.1 8.7 9.8 8.7 8.7 7.9
(60) (200) (180) (220) (250) (220) (220) (200)
Aggregate proportions 0.705 0.714 0.731 0.687 0.675 0.712 0.711 0.715 0.708 0.712 0.714
w/c** 0.50 0.33 0.33 0.33 0.33 0.42 0.28 0.38 0.33 0.33 0.33
s/c 0.0 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.0 0.05 0.15
Compressive strengths in ksi (MPa)
fci 1.65 3.68 3.88 3.31 4.54 2.99 4.96 2.07 3.68 3.70 4.18
(11,4) (25,4) (26,8) (22,8) (31,3) (20,6) (34,2) (14,3) (25,4) (25,5) (28,8)
Ic3 3.68 7.47 7.21 6.97 7.12 5.16 7.79 5.48 5.54 6.21 6.18
(25,4) (51,5) (49,7) (48,1) (49,1) (35,6) (53,7) (37,8) (38,2) (42,8) (42,6)
fc7 4.6 10.3 10.0 10.1 10.2 8.21 10.9 8.38) 8.29 9.35 9.71
(32,0) (70,7) (69,1) (69,5) (70,3) (56,6) (75,6) (57,8) (57,2) (64,5) (67,0)
fc28 6.31 13.4 13.7 13.5 14.4 10.8 14.1 11.5 9.74 10.8 13.7
(43,5) (92,1) (94,3) (93,3) (99,4) (74,6) (97,3) (79,5) (67,2) (74,6) (94,3)
Young's modulus in ksi x 103 (GPa)
Ei1 3.93 4.47 4.60 4.16 5.15 4.23 5.60 3.35 4.38 4.52 4.92
(27,1) (30,8) (31,7) (28,7) (35,5) (29,2) (38,6) (23,1) (30,2) (31,2) (33,9)
Ei3 4.95 6.25 6.26 5.97 6.13 5.96 6.82 5.90 6.05 6.15 6.38
(34,1) (43,1) (43,2) (41,2) (42,3) (41,1) (47,0) (40,7) (41,7) (42,4) (44,0)
Ei7 5.28 6.73 7.09 6.77 6.77 6.47 7.40 6.34 6.67 6.93 6.93
(36,4) (46,4) (48,9) (46,7) (46,7) (44,6) (51,0) (43,7) (46,0) (47,8) (47,8)
Ei28 5.99 7.31 7.51 7.70 7.05 6.80 7.74 6.57 6.67 7.22 7.50
(41,3) (50,4) (51,8) (53,1) (48,6) (46,9) (53,4) (45,3) (46,0) (49,8) (51,7)
Autogenous shrinkage between 72
and 5000 hours (1(H) 41 89 76 111 108 67 82 91 44 64 94
* Melamine resin at 30.9% dry extract
* * w/c = total water/cement alone
w/c = 0.28
w/c = 0.33
w/c = 0.38
fc (MPa)
w/c = 0.42
Normal-
strength
concrete
Age (days)
s/ c = 0.00
s/c = 0.05
s/c = 0.10
fc (MPa)
s/c = 0.15
Normal-
strength
concrete
Age (days)
vp = 0.296
vp = 0.286
fc (MPa)
vp = 0.313
vp = 0.325
Normal-
strength
concrete
Age (days)
Fig. 3.5 Compressive strength development for the ten concrete mixes given in Table 3.2.
Dash lines represent average values of the control mix, a 5000 psi (35 MPa) normal strength
concrete made with the same components
w/c = 0.38
w/c = 0.42
Normal-
strength
concrete
Age (days)
s/ c = 0.00
s/c = 0.05
s/c = 0.10
Ec [GPa]
s/c = 0.15
Normal-
strength
concrete
Age (days)
vp = 0.296
vp = 0.286
Ec [GPa]
vp = 0.313
vp = 0.325
Normal-
strength
concrete
Age (days)
Fig. 3.6 Young's modulus development for the ten concrete mixes given in Table 3.2. Dash
lines represent average values of the control mix, a 5000 psi (35 MPa) normal strength
concrete made with the same components
3.3 Creep
When a viscoelastic material is subjected to a stress, from a time ^0 > its
strain, measured parallel to the axis of the stress, changes over time. This
time dependent increase in strain of hardened concrete subjected to
sustained stress is termed creep. It is usually determined by subtracting,
from the total measured strain in a loaded specimen, the sum of the initial
instantaneous strain (usually considered elastic) due to sustained stress, the
shrinkage, and any thermal strain in an identical load-free specimen,
subjected to the same history of relative humidity and temperature
conditions. The principal variables that affect creep are summarized in
Table 3.1.
Creep is closely related to shrinkage and both phenomenon are related
to the hydrated cement paste. As a rule, a concrete that is resistant to
shrinkage also has a low creep potential. The principal parameter influenc-
ing creep is the load intensity as a function of time; however, creep is also
influenced by the composition of the concrete, the environmental condi-
tions and the size of the specimen.
A distinction must first be made between basic creep, which occurs in the
absence of hygrometric exchanges with the ambient medium, and drying
creep, that appears when the material dries. The former corresponds to the
creep of a very thick actual structural element, whereas the latter,
generally measured in the laboratory on small specimens - transverse
dimensions less than 0.5 ft (15 cm) - reflects the maximum creep exhibited
by thin structures. The basic creep and the total creep are the limits,
between which lies the creep of a part having any shape, drying at between
50 and 100% relative humidity.
There is an abundant literature on creep of normal strength concretes
and Neville23 has given an excellent summary. Information on creep of
concretes of higher strengths is limited.2'24
creep is intrinsic to the hydrates (the creep of the aggregates being zero or
negligible with respect to that of the matrix; the same for the anhydrous
cement). Its physical origin is poorly understood, and in any case no
consensus has been reached concerning it. But it must be controlled to a
great extent by the total volume of hydrates in the concrete microstructure.
The amplitude of the basic creep is also directly influenced by the presence
of free water in the microstructure of the material: a dry concrete exhibits
no creep.25
Therefore, two features of HSC microstructure are expected to modify
the creep behavior (compared with the one of NSC): the volume of
hydrates and the free water content. In Table 3.3, one can see the volume
of hydrates for cement pastes with water/binder ratio ranging between 0.2
to 0.4, and silica/cement ratio ranging from O to 16% (data taken from
Sellevold and Justnes26). Obtaining a higher strength concrete entails
imposing a low water/binder ratio; this gives a reduced volume of the
hydrate and the free water content. Moreover, self-desiccation appears.
These are factors which reduce creep.
£Cr(t, to) ~
Kcr.«.[(t-t0r/v+«-t0r]
J-, \J'^)
&i28
where / is the time (in hours), ^0 is the age of concrete at loading, a is the
applied stress, En^ is the elastic modulus, ecr is the creep strain, Kcr is the
creep coefficient, a and (3 are material parameters depending on ^ 0 .
The results of the creep tests, with specific creep values at the end of the
tests are given in Table 3.4. Creep curves for concrete Cl are shown in
Fig. 3.7. For each mix, it can be noted that the amplitude of specific creep
decreases with the age of the material. The ratio of specific creep of
concrete loaded at 1 day to the one of concrete loaded at 28 days is
approximately equal to 2. Another general feature is the lowering of the
kinetics: the concrete creeps faster when it is loaded at early age. As a
consequence, for some mixes, the specimens loaded at 28 days continued
to creep quite quickly at the end of the tests, so that the extrapolation (Kcr
value) is not very reliable. The results also indicate that creep is sensitive to
the paste content of concrete (Fig. 3.8) and it decreases when the water/
cement ratio decreases (Fig. 3.9). A low dosage of silica fume (5% by
weight of cement) leads to a decrease of the specific creep, but higher
dosage (up to 15%) increases the deformation (Fig. 3.10).
In summary, any change of the mix-design parameters involving an
increase of strength also leads to a decrease of creep, except when the silica
fume amount is more than 10%. On the other hand, the kinetics of
deformations do not show any obvious tendency when the values of
mix-design parameters are changed.
Time (hours)
Fig. 3.7 Basic creep of concrete Cl (mix composition, see Table 3.2)
there are other available models.29'30'31'32 The model of Bazant and Chern
based on log double-power law appears to be the quite effective analyti-
cally.
At LCPC (Paris), Auperin, de Larrard et al.33 have conducted in the
past a preliminary program of experiments on a silica fume concrete having
a strength of 11.6 ksi (80 MPa) at 28 days, with a view to characterizing its
creep vs. the age of loading. Based on the trend of the experimental data.
The following expression was proposed:
e cr (f, to) = *cr№(28)]/(f-f0) (3.3)
where Kcr is the creep coefficient, Et is the modulus and f(t —10) a kinetic
Specific creep (microstrain/MPa)
Water/cement ratio
Fig. 3.9 Influence of the w/c ratio on the basic creep of high strength concrete (specific creep
after two years)
function of time t, which is expressed in days. For the modulus, the creep
coefficient and the kinetic function, the following expressions were pro-
posed.
£,(/o) = 1.132exp[-0.42/(r0-0.4)1/2]£;28 (3.4)
1/2
Kcr(to) = 0.363 exp[9.3/(/0-0.34)] (3.5)
= exp[-1.7/(/-/Q + O.Q27)]1/3-exp(-4)
l-exp(-4)
Investigation of different loading paths (for HSC concrete 760 Ibs/cyd,
484.4 kg/m3, of cement, 8% silica fume and w/c of 0.38) showed that
superposition was also valid for higher strength concretes (Fig. 3.11).
Comparison of data of this high strength concrete with other data from the
literature revealed that its creep kinetics was exceptionally fast. The creep
amplitude for early loading (at one day) was also unusually high.
Specific creep (microstrain/MPa)
Silica/cement ratio
Fig. 3.10 Influence of the silica/cement on the basic creep of high strength concrete (specific
creep after two years)
4.30 MPa at 1.23 d. then 20.17 MPa at 7.19 d.
Fig. 3.11 Creep of high strength concrete subjected to two stress steps, and application of
the superposition method. fC 28 = 12 ksi (8SaMPa), cement: 760 Ib/yd (450 kg/m), silica
fume: 8%, w/c = 0.38; deformation in microstrains, time in days (linear scale)
4.1 Introduction
Many concrete structural members are subjected to repeated fluctuating
loads the magnitude of which is well below the maximum load under
monotonic loading. This type of loading is typically known as fatigue
loading. Contrary to static loading where sustained loads remain constant
with time, fatigue loading varies with time in an arbitrary manner. Fatigue
is a special case of dynamic loading in which inertia forces do not influence
the stresses. Examples of structures that are subjected to fatigue loading
include bridges, offshore structures and machine foundations.
Fatigue is one property of concrete that is not well understood, specially
in terms of the mechanism of failure, because of the difficult and tedious
experiments required for conducting research investigations. Detailed
presentations of research findings concerning the fatigue behavior of
concrete are included in the references.2'3'21'38'40 Use of high performance
concrete in structures subjected to fatigue loading requires knowledge
about its behavior under such loading. As mentioned previously (Chapter
2) the performance characteristics of concrete generally improve with the
strength attribute. Unfortunately, the available data on the fatigue be-
havior of high-strength concrete is very limited.4'5 This chapter presents an
overview of the results available in the published literature on concrete
fatigue in general, with emphasis on the behavior of high-strength concrete
(compressive strength >6000psi (42MPa)).
Fatigue is the process of cumulative damage that is caused by repeated
fluctuating loads. Fatigue loading types are generally distinctly divided
between high-cycle low amplitude and low-cycle high amplitude. Hsu20 has
summarized the range of cyclic loading into a spectrum of cycles. He
classified the fatigue loads into three ranges, Table 4.1. The low-cycle
fatigue loading occurs with less than 1000 cycles. The high-cycle fatigue
loading is defined in the range of 103 to 107 cycles. This range of fatigue
Table 4.1 Fatigue load spectrum20
Super-high-
Low-cycle fatigue High-cycle fatigue
cycle fatigue
transmit structures
Sea structures
Highway and
Mass rapid
railway bridges,
Airport highway
Structures subjected
pavements pavements,
to earthquake
and bridges concrete
railroad ties
Number of cycles
/\ COW
(a) £ R = -1
CO
'mean
(b) I
co R =O
CO
(C) I
CTD
O^R^1
Time
Fig. 4.1 Constant amplitude fatigue loading, (a) R = -1; (b) R = O; (c) 0</?< 1
A/ = / ma x-/min (4-1)
^. /max /min / j ~\
/mean ~ I V*-^
/. /max /min / ^ ^\
/amp — ^ V*-^/
R=f™L (4.4)
/max
Time
Fig. 4.2 Variable-amplitude random sequence fatigue loading
2)£
1=1 Ni = H> (4-6)
The factor w has been expressed as a function of the ratio (SminASc),
Fig. 4.3.
Concern with fatigue damage of concrete was recognized early in this
century. Van Ornum50'51 observed that cementitious composites possessed
the properties of progressive failure, which become total under the
repetition of load well below the ultimate strength of the material. He also
noticed that the stress-strain curve of concrete varies with the number of
repetitions, changing from concave towards the strain axis (with a hys-
teresis loop on unloading) to a straight line, which shifts at a decreasing
rate (plastic permanent deformation) and finally to concave toward the
stress axis, Fig. 4.4. The degree of this latter concavity is an indication of
how near the concrete is to failure.
The fatigue strength can be represented by means of S-N curves (known
also as the f-N curve or the Wohler's curve of the fatigue curve). In these
curves (Fig. 4.5.) S is a characteristic stress of the loading cycle, usually
indicating a stress range or a function of the maximum and minimum stress
and TV is the number of cycles to failure. .S is expressed with a linear scale
while W is presented using log scale. Using this format, usually the data can
be approximated to a straight regression line. At any point on the curve,
the stress value is the 'fatigue strength' (i.e., the value of stress range that
Stress level S
RMS
C sm Sc
°min
Time
Iog10 M(= Iog10fi))
min
lr»n
9io / ^— \
Io
\ Sc '
Fig. 4.3 Empirical relationship between fatigue loading parameters and the factor w
(= Miner sum at failure)19
will cause failure at a given number of stress cycles and at a given stress
ratio) and the number of cycles is the 'fatigue life' (i.e., the number of
stress cycles that will cause failure at a given stress range and stress ratio).
To include the effect of the minimum stress / min , and the stress range
/max~/min> the fatigue strength can also be represented by means of a
N = 10,390.
Concrete stress (MPa)
N-160
Concrete stress (psi)
N=1
N - 28,690
N - 22,480
Strain x 106
Fig. 4.4 Effect of repeated load on concrete strain3
Smax - Stress level - Maximum
Smin = 0.035
stress/static strength
Smax
Smin Run out
Time
Maximum
stress
Per cent of static ultimate strength
Range of
stress
Minimum
stress
Fig. 4.6 Modified Goodman diagram for concrete subjected to repeated axial loading16
proposed for either normal or high strength concrete. However, with the
present stage of knowledge the following two observations can be stated.42
1 The fatigue of concrete is associated with initiation and propagation of
internal microcracks at the cement paste-aggregate interface and/or
within the cement paste itself.
2 Cracks due to fatigue failure are more extensive than cracks initiated by
static compressive failure.
\ogN=ll\l-(fmax_~f™nWc'] (4.8)
L 1 /min'/c J
~ = 1.2-0.2/?-0.133(l-0.779/?)logN-0.05301(l-0.445/?)logr
*c (4.9b)
These equations were found to be applicable for the following conditions:
(a) normal weight concrete with/ c ' up to 8000 psi (56 MPa)
(b) for stress range O < R < 1
(c) for load frequency range between O to 150 cycles/sec
(d) for number of cycles from 1 to 20 million cycles
(e) for compression and flexure fatigue.
Chimamphant10 conducted series of experiments on uniaxial cyclic
compression of high-strength concrete. Concrete strength varied from
7500 psi (52MPa) to 12,000 psi (84MPa). Maximum stress level varied
from 0.4/c' to 0.9/c' while the minimum stress level was kept constant at
0.1/c'. Two different rates of loading were used, namely: 6 cycles/sec and
12 cycles/sec. He observed no significant difference in the fatigue behavior
of high strength concrete when compared to normal strength concrete. He
reported that up to 1 million cycles the S-W curve of high-strength concrete
was linear. He also observed no measurable effect on the fatigue strength
Notations
Fmax = Maximum stress of cyclic loading
fmin = Minimum stress of cyclic loading
fmax fsus = Sustained strength (or discontinuity strength) at 10 years
I^ fst = Static strength at period of 1 sec/cycle
R — ~~ f'min'/f"max
_fat_ N = Number of cycles
f« T = Period in seconds per cycle
F
ASTM R =O
loading R=1
rate
G Ci
R=1
R =1 R-O
R-O
log N (cycle)
log T (sec/cycle)
Fig. 4.7 Graphical representation off-N-T-R relationship20
Smin = O
Log N
Fig. 4.8 The S-N diagram for failure of high strength concrete in compression36
Friction
Bearing
Radial pressure
Bond strength
Fig. 4.9 Bond strength components for bar embedded in concrete50
+ 195.0/aux-h21.06(^7V)°-66
4.14 Summary
The use of high strength concrete in modern structures and the widespread
adoption of ultimate strength design procedures renewed the interest in
studying the fatigue and bond properties of high strength concrete.
Most of the fatigue investigations have been performed on concretes
of normal strength. These investigations have established the effect of the
range of stress, load history, rate of loading, stress gradient, curing and
material properties on the fatigue behavior. The limited fatigue studies
conducted on high strength concrete covered the effects of these variables
on the fatigue properties. Their results indicated no signifcant difference
between the fatigue behavior of normal and high strength concretes when
the stress levels are expressed relatively to the static strength of concrete.
The bond at the hydrated cement paste-aggregate interfaces is greatly
improved in high strength concrete because of the addition of pozzolanic
materials. This could result in the coarse aggregate characteristics becom-
ing the limiting factor of the high strength that can be achieved. The
average reinforcing steel-concrete matrix bond is increased for high
strength concrete when compared to normal strength concrete. More
research is needed on the morphology and micro-structure of the steel-
cement paste transition zone in order to characterize the effect of including
pozzolanic materials in high strength concrete on the bond properties.
References
1 Aas-Jakobson, K. (1970) Fatigue of concrete beams and columns, Bulletin No.
70-1. NTH Institute of Betonkonstruksjoner, Trondheim.
2 ACI Committee 215 (1974) Fatigue of concrete. American Concrete Institute,
SP-41.
3 ACI Committee 215 (1974) Consideration for design of concrete structures
subject to fatigue loading. Journal of the American Concrete Institute, Proceed-
ings, 71, 3, 97-121. Revised in 1986.
4 ACI Committee 363 (1984) State-of-the-art report on high-strength concrete.
ACI Journal, Title No. 81-34, 364-411.
5 ACI Committee 363 (1987) Research needs for high-strength concrete. ACI
Journal, Title No. 84-M49, 559-61.
6 Bartos, P. (ed.) (1982) Bond in concrete. Applied Science Publishers.
7 Bennett, E.W. and Muir, S.E.st.J. (1967) Some fatigue tests on high-strength
concrete in axial compression. Magazine of Concrete Research, London, 19,
No. 59, 113-17.
8 Brettmann, B., Darwin, D. and Donahey, R. (1986) Bond of reinforcement to
superplasticized concrete, Proceedings. ACI Journal, 98-107.
9 Chapman, R.A. and Shah, S.P. (1987) Early age bond strength in reinforced
concrete. ACl Journal, 84, No. 6, 501-10.
10 Chimamphant, S. (1989) Bond and Fatigue Characteristics of High-Strength
Cement-Based Composites, Ph.D. Dissertation, New Jersey Institute of
Technology, Newark, New Jersey.
11 Clark, P. (1949) Bond of concrete reinforcing bars. ACI Journal, Proceedings,
46, 3, 161-84.
12 Clemmer, H.E. (1922) Fatigue of concrete. Proceedings, ASTM, 22, Part II,
408-19.
13 Cornelissen, H. A.W. and Timmers, G. (1981) Fatigue of plain concrete in
uniaxial tension and alternating tension- compression, Report No. 5-81-7,
Stevin Laboratory, University of Technology, Delft.
14 Ezeldin, A.S. and Balaguru, P.N. (1989) Bond behavior of normal and
high-strength fiber reinforced concrete. ACI Materials Journal, Proceedings,
86, 5, September-October, 515-24.
15 Gj0rv, O.E., Monteiro, PJ. and Mehta, P.K. (1990) Effect of condensed silica
fume on the steel-concrete bond. ACI Materials Journal, Proceedings, 87, 6,
Nov-Dec, 573-80.
16 Graf, O. and Brenner, E. (1934/1936) Experiments for investigating
the resistance of concrete under often repeated loads (Versuche Zur Ermittlung
der Widerstandsfahigkeit Von Beton gegen oftmals Wiederholte Druck-
belastung), Bulletins No. 76 and No. 83, Deutscher Ausschuss fur Eisen-
beton.
17 Gray, W.H., McLaughlin, J.F. and Antrim, J.D. (1961) Fatigue properties of
lightweight aggregate concrete. ACl Journal, Title No. 58-6, August, 149-61.
18 Hilsdorf, H.K. and Kesler, C.E. (1966) Fatigue strength of concrete under
varying flexural stresses. Journal of American Concrete Institute, Proceedings,
63, 10, Oct, 1069-76.
19 Holmen, J.O. (1979) Fatigue of concrete by constant and variable amplitude
loading, Bulletin No. 79-1, Division of Concrete Structures, Norwegian
Institute of Technology, University of Trondheim.
20 Hsu, T.C. (1981) Fatigue of plain concrete. ACl Journal, Proceedings, Title
No. 78-27, July-August, 292-305.
21 (1982) International association for bridge and structural engineering, Proceed-
ings of Colloquium, Lausanne, IABSE Reports, 37.
22 Jinawath, P. (1974) Cumulative Fatigue Damage of Plain Concrete in Com-
pression, Ph.D. Thesis, University of Leeds.
23 Kakuta, Y. et al. (1982) New concepts for concrete fatigue design procedures
in Japan. Proceedings, IABSE, Lausanne, 37, 51-8.
24 Kemp, E. (1986) Bond on reinforced concrete: behavior and design criteria.
ACl Journal, Proceedings, 82, 1, Jan-Feb, 49-57.
25 Kesler, C.E. (1953) Effects of speed of testing on flexural fatigue strength of
plain concrete. Proceedings, Highway Research Board, 32, 251-8.
26 Klaiber, F.W. and Lee, D.Y. (1982) The effects of air content, water-cement
ratio, and aggregate type on the flexural fatigue strength of plain concrete.
American Concrete Institute, Special Publications, SP-75, 111-32.
27 Kolias, S. and Williams, R.LT. (1978) Cement-bound road materials: strength
and elastic properties measured in the laboratory, TRRL Report No. 344.
Transport and Research Laboratory, Crowthorne, Berkshire.
28 Leeuwen, J.V. and Siemes, J.M. (1979) Miner's rule with respect to plain
concrete. Heron, 24, 1, 34pp.
29 McCaIl, J.T. (1958) Probability of fatigue failure of plain concrete. ACI
Journal, Aug, 233-44.
30 Mindess, S. (1989) Interfaces in concrete, in Materials science of concrete /,
edited by Skalny, J.P. The American Ceramic Society, 163-180.
31 Mindess, S. (1988) Bonding in cementitious composites - how important is it,
in Bonding in cementitious composites, edited by Mindess, S. and Shah, S.P.
Materials Research Society, 114, 3-10.
32 Miner, M.A. (1945) Cumulative damage in fatigue. Transactions, American
Society of Mechanical Engineers, 67, A159-A164.
33 Murdock, J.W. and Kesler, C.E. (1958) Effect of range of stress on fatigue
strength of plain concrete beams. ACl Journal, August, 221-31.
34 OpIe, Jr, F.S. and Hulsbos, C.L. (1966) Probable fatigue life of plain concrete
with stress gradient, Research Report. ACl Journal, Title No. 63-2, January,
59-81.
35 Palmgren, A. (1924) 'Die Lebensdauer von Kugellagern, VDI.' Zeitschrift
Verein Deutscher Ingenieur, 68, 339-41.
36 Petkovic, G., Lenschow, R., Stemland, H. and Rosseland, S. (1991) Fatigue of
high-strength concrete. American Concrete Institute, Special Publication, SP
121-25, 505-25.
37 Raithby, K.D. and Galloway, J.W. (1974) Effects on moisture condition, age,
and rate of loading on fatigue of plain concrete. ACI Publications, SP-41,
15-34.
38 RILEM Committee 36-RDL (1984) Long term random dynamic loading of
concrete structures. RILEMs Materials and Structures, 17, 97, Jan, 1-27.
39 Saito, M. and Imai, S. (1983) Direct tensile fatigue of concrete by the use of
friction grips. ACl Journal, Title No. 80-42, Sept-Oct, 431-8.
40 Shah, S.P. (1982) Fatigue of concrete structures. American Concrete Institute,
SP-75.
41 Sparks, P.R. and Menzies, J.B. (1973) The effect of rate of loading upon the
static and fatigue strengths of plain concrete in compression. Magazine of
Concrete Research, 25, 83, June, 73-80.
42 Su, E.C.M. and Hsu, T.T.C. (1986) Biaxial compression fatigue of concrete,
Research Report UHCE 86-17. University of Houston, December.
43 Takhar, S.S., Jordaan, LJ. and Gamble, B.R. (1974) Fatigue of concrete under
lateral confining pressure. ACI Publications, SP-41, 59-69.
44 Tepfers, R. (1979) Tensile fatigue strength of plain concrete. ACl Journal,
Title No. 76-39, August, 919-33.
45 Tepfers, R. (1986) Fatigue of plain concrete subjected to stress reversals. ACI
Publications, SP-75, 195-215.
46 Tepfers, R., Gorlin, J. and Samuelsson, T. (1973) Concrete subjected to
pulsating load and pulsating deformation of different pulse wave-form.
Nordisk Betong, No. 4, 27-36.
47 Tepfers, R. and Kutti, T. (1979) Fatigue strength of plain, ordinary, and
lightweight concrete. ACl Journal, Title No. 76-29, May, 635-53.
48 Traina, L.A. and Jeragh, A.A. (1982) Fatigue of plain concrete subjected to
biaxial-cyclical loading. American Concrete Institute, Special Publications,
SP-75, 217-234.
49 Treece, R. A. and Jirsa, J.O. (1989) Bond strength of epoxy-coated reinforcing
bars. Proceedings, ACI Materials Journal, 86, 2, March-April, 167-74.
50 Van Ornum, J.L. (1903) Fatigue of cement products. Transactions, ASCE, 51,
443.
51 Van Ornum, J.L. (1907) Fatigue of concrete. Transactions, ASCE, 58,
294-320.
52 Williams, H. A. (1943) Fatigue tests of lightweight aggregate concrete beams.
ACI Journal, Proceedings, 39, April, 441-8.
^ Durability
O E Gj0rv
5.1 Introduction
Recent developments in concrete technology have made it possible to
produce concrete mixtures with strength properties that are beyond the
strengths that are currently used by the structural design practice. For
high-quality natural mineral aggregates, compressive strengths of up to
33,000 psi (230 MPa) can now be produced under laboratory conditions.1 If
the mineral aggregate is replaced by high-quality ceramic aggregate,
compressive strengths of up to 65,000 psi (460 MPa) can be achieved under
controlled laboratory conditions.2 Even with lightweight aggregate, com-
pressive strengths of more than 14,000 psi (100 MPa), with a density of less
than 3200 lb/yd3 (1900 kg/m3), can be obtained.3 However, very often it is
not the improved strength which is the primary objective but rather the
improved durability and overall performance. Therefore, the term 'high-
performance concrete', is inclusive of the term 'high-strength concrete'. In
this chapter, a brief summary of the most recent developments on
durability of high-strength concrete is presented.
5.2 Permeability
One of the main characteristics of high-strength concrete compared to that
of the normal-strength concrete is the more uniform and homogeneous
microstructure. When the portland cement is combined with the ultrafine
particles of silica fume in low w/c ratios, the microstructure of such systems
consists mainly of poorly crystalline hydrates forming a more dense matrix
of low porosity. With increasing silica fume content, larger content of the
calcium hydroxide is transformed into calcium silcicate hydrates (Fig. 5.1),
while the remaining calcium hydroxide tends to form smaller crystals
compared to that in pure portland cement pastes. From Table 5.1 it can be
seen that with the addition of higher percentage of silica fume, the
Calcium hydroxide
content by TGA
Cured:
52 days
110 days
(9/9 ign. cem.)
65 days
w/c = 0.60
and thus leads to a poorer bond between the aggregate and the cement
paste, is typically characterized by the following key elements:
1 The transition zone is richer in calcium hydroxide and ettringite than
the bulk phase, and the calcium hydroxide is oriented. A rim of massive
calcium hydroxide can often be observed around the aggregates.
2 The porosity of the transition zone is greater than that of the bulk
phase, and a gradient in porosity can be observed with a declining trend
as the distance from the aggregate surface increases.
For normal-strength concrete, the special microstructure of the transi-
tion zone is apparently related to the formation of water-filled spaces
around the aggregates in the fresh concrete, due to internal bleeding and to
a 'wall effect', which interferes with inefficient packing of the cement
particles around the aggregates. Calcium hydroxide and ettringite are
known to preferably grow in large pores, which accounts for the greater
contents of these phases in the transition zone.8'9 Also, this zone has larger
w/c ratio relative to the bulk10 and is therefore characterized by higher
porosity.
When silica fume is introduced into the system, and in particular in
high-strength concrete, considerable changes in the microstructure of the
transition zone take place. Regourd et al.n'12 and Aitcin et #/.13'14
observed that high-strength concrete with silica fume was not as crystal-
lized and porous as normal-strength concrete, and all of the space in the
vicinity of the aggregate was occupied with amorphous and dense calcium
silicate hydrates. Also, direct contact was formed between the aggregate
and the calcium silicate hydrates rather than with calcium hydroxide as in
normal-strength concrete. Scrivener et al.15 quantified the interfacial
microstructure and demonstrated that in high-strength concrete with silica
fume, the porosity of the transition zone was practically eliminated
(Fig. 5.2), and practically no gradient in porosity was observed, in contrast
to normal-strength concrete.
As demonstrated in Table 5.3 the addition of silica fume may have a
substantial effect on the permeability of concrete. By adding 20% of silica
fume to 169 lb/yd3 (100 kg/m3) of cement (OPC), the same permeability as
that of 421 lb/yd3 (250 kg/m3) cement is obtained. Addition of 10% silica
fume to 421 lb/yd3 (250 kg/m3) cement gives a permeability as low as
1.8.10~14m/s. For meeting the higher durability performance require-
ments for offshore concrete platforms in the North Sea, the permeability is
limited to 10~2m/s.17
Bulk
concrete Bulk
Transition zone matrix paste
180 days
Control
Porosity detected by BSE (% vol.)
Silica fume
1day
Control
Silica fume
For cement pastes with very low w/c ratio in the range of 0.20 to 0.30,
experiments18 have shown that a 10% replacement of the cement with
silica fume did only reduce the total porosity to a small extent. However, a
refinement of the pore size distribution took place in such a way that the
content of larger pores was reduced for decreasing w/c ratio. The effect of
w/c ratio on chloride diffusivity was substantially higher at high than at low
w/c ratios, but a 10% replacement with silica fume reduced the diffusivity
so much that the effect of w/c ratio became less significant.
The above qualitative and quantitative observations indicate that in
high-strength concrete the bulk matrix becomes very dense and typically,
this dense matrix extends up to the aggregate surface, in such a way that
the inhomogeneity of the transition zone is largely eliminated. It is now
Silica content
(as percentage cement replacement)
Silica content
(as percentage cement replacement)
400kg cement/m33
250kg cement/m3
100kg cement/m
Resistivity (kohm cm)
Environment
Marine
Urban
Industrial
Loss of section (%)
Permeability
Fig. 5.8 Relationship between resistance to chemical deterioration and permeability 16
reducing the calcium hydroxide, pozzolans, such as silica fume, also form
calcium silicate hydrates which are able to incorporate aluminum ions, thus
reducing the amount of aluminum available for ettringite formation.
The beneficial effect of silica fume in high sulfate-containing environ-
ments has been reported in several investigations.44'45 In these inves-
tigations, the performance achieved has been equal or better than that
obtained by use of sulfate resistant cements. Mehta46 exposed a number of
high-strength concretes to solutions of both 5% sodium sulfate and 1%
concentrations of sulfuric and hydrochloric acid for a period of up to 182
days. Although the portland cement contained 7% C3A, the results
showed that w/c ratios of 0.33 to 0.35 gave too low a permeability to cause
any deterioration. In more aggressive environments pure portland cements
have shown some deterioration, whereas addition of silica fume has given
practically unaffected performance.45 Even for normal-strength concrete
the presence of silica fume will improve the long-term performance of
concrete in very aggressive environments. In 1952 a large number of
concrete specimens were exposed in a field station in the underground of
Oslo city, which consists of very aggressive alum shale. In spite of a sulfate
content of up to 5 g/1 of SO3 and a pH of 2.8, a 15% replacement of
ordinary portland cement with silica fume gave the same good perform-
ance after 26 years of exposure as that by sulfate resisting cements (ASTM
Type V).47
The presence of pozzolans such as silica fume can also be used to control
the expansion caused by alkali-aggregate reaction.48'49 Porewater analyses
of cement paste with silica fume have demonstrated the ability of silica
fume to reduce the alkali concentration in the pore water rapidly, thus
making it unavailable for the slower reaction with reactive silica in the
aggregate.19'20 Also, for high-strength concrete the effect of self-
desiccation may reduce the moisture content to a level where no alkali-
aggregate reaction can take place.
Asgeirsson and Gudmundsson50 used silica fume with Icelandic high-
alkali cement and reactive sand in mortar bars to demonstrate the ability of
silica fume to reduce the expansion. Later on, field experience with
Icelandic cement blended with silica fume in 200 houses constructed during
the period of 1979-86 has confirmed these observations.51
e in %0
(a)
Series: OPC
100% = 32.9 N/mm2
a/ault(20°C)in%
8 in %o
(b)
Fig. 5.9 Effect of elevated temperature on the stress-strain relationship for (a) high-strength
concrete with ordinary Portlane cement and fly ash, and (b) ordinary Portlane cement52
Temperature in 0C
Fig. 5.10 Effect of elevated temperature on compressive strength52
S1 aU|t (2O0C0 = 106.6 N/mm22
Lt ault (2O C)- 84.5 N/mm
Trault(20°C)0 = 91.8N/mm2 2
OPC aU|t (2O C) - 32.9 N/mm
auit(T)/°uit(200C) in %
Temperature in 0C
Fig. 5.11 Effect of elevated temperature on modulus of elasticity52
Fig. 5.12 Accelerated load facility for testing the abrasion resistance of highway pavements
exposed to studded tires64
WeI Dry Wet Type of aggregate
Syeniteporphyr
Hornfels
Quartzdiorite
Partial abrasion (mm/104 rev.)
Jaspar
three years or less. Encouraged by the above test results several high-
strength concrete pavements have been completed over recent years.65
The wear of non-pavement surfaces such as concrete floors, sidewalks,
stairs, etc. is caused primarily by foot traffic, light vehicular traffic, and the
skidding, scraping, or sliding of objects on the surface. In some types of
industrial operation, the use of steel wheeled vehicles, forks, buckets of lift
trucks and loaders inflicts very severe damage to the concrete surfaces they
operate on. The movement of abrasive granular material in and out of
concrete storage facilities such as silos and bins also creates significant
abrasion problems. In the Scandinavian countries the utilization of high-
strength concrete for such problems has already started.
Hydraulic abrasion, or abrasion erosion is recognizable by the smooth,
worn concrete surface in addition to cavitation erosion, where the surface
is full of small holes and pits. Spillway aprons, stilling basins, sluiceways
and tunnel linings are particularly susceptible to abrasion erosion.
Most concretes used in hydraulic structures in the past could not be
classified as high-strength concrete. The concrete provided the mass, and
strength was a secondary consideration only. Concerns about thermal
cracking at early ages in the large concrete sections used in these structures
have also led to low cement contents to minimize heat of hydration, with
subsequent low concrete strengths. Where higher quality concrete is used,
the resistance to high water velocities has been satisfactory for many years,
but even these concretes do not fully resist the abrasive action, grinding or
repeating impacts of the debris. Observations of abrasion-erosion of
concrete surfaces in the stilling basins of several major USA dams have
varied from 2 in. (50mm) to 118 in. (3000mm). At Dworshak Dam,
2000 yd3 (1530 m 3 ) of concrete and bedrock were eroded from the stilling
basin. In many of these instances, the abrasion-erosion is accelerated
because of the impact forces of large rocks and boulders caught in
turbulent flows. These forces weaken the concrete surface and make it
more susceptible to removal.
In 1983 a major repair on the stilling basin of Kinzua Dam in Pennsylva-
nia, USA, was carried out by using high-strength concrete.66 The structure
which has been in operation since 1967, had already had an extensive
repair carried out in 1973-74. In 1983 approximately 1960 yd3 (1500 m 3 ) of
10 in. (250mm) slump concrete was placed, the 28 day compressive
strength of which was 12,857psi (9OMPa). Diver inspection of the
concrete after one year of service including a period with very large volume
of debris in the stilling basin, showed that the concrete was performing as
intended.
For marine or offshore concrete structures exposed to 'ice abrasion', the
actual mechanism that results in loss of surface is more complex than the
simple act of ice rubbing on the concrete surface. Research work and field
observations suggest that the concrete deterioration at or near the water-
line is due to a combination of environmental causes plus the impact of
loading of the concrete surface by repeated impacts from ice floes.
Pieces of ice, driven by wind and current, can possess significant kinetic
energy, much of which is dissipated into the concrete when the ice collides
with the concrete structure. Some of the energy is lost in the crushing of the
ice. The frequency of the loading is dependent on the circumstances which
occur at any given time at a particular structure and can vary from an
occasional impact to repeated impacts every few seconds. A large ice floe
in open waters will, upon initial contact with a structure, both load the
structure and begin to crush itself at the point of contact with the structure.
As the driving forces of the floe continue to move it forward against the
structure, the resistance of the structure continues to increase to a point
where the floe experiences a local failure in the ice, usually in flexure, some
distance from the initial point of contact with the structure. This momen-
tarily releases the load on the structure. The original ice, now damaged by
crushing and cracking, is shunted away by the moving flow and new,
undamaged ice in the floe then collides with the structure. The characteris-
tics of the ice and the floe, combined with the dynamic response of the
structure, will establish a 'ratcheting' effect on the concrete surface,
repeatedly loading and unloading it. With time, this repetitive loading
behavior can destroy the effectiveness of the aggregate bond near the
surface of the concrete and both cause and propagate existing microcracks
in the concrete.
Regardless of type of abrasion, both laboratory and field experience
indicate that compressive strength is the single most important factor in
determining the abrasion resistance of the concrete. Also, the abrasion
resistance can be significantly improved by the use of hard and dense
aggregate both in the upper and lower part of the grading curve. By
replacing the 0.079 to 0.157 in. (2 to 4mm) fraction of the natural sand
with crushed high quality material, the compressive strength in the
Norwegian highway pavement investigations65 decreased from 23,471 to
21,900 psi (164.3 to 153.3MPa), while the service life of the pavement
increased by 50%. Generally, the abrasion of concrete is higher in wet than
in dry condition, but the experience indicates that also this effect is reduced
by increased strength level. From Fig. 5.13 it can be seen that at 7143 psi
(5OMPa), the wet abrasion loss was approximately 100% higher than the
dry abrasion loss, while at 14,286 psi (100 MPa) the wet abrasion loss was
only 50% higher. At 214,429 psi (15OMPa), only a small difference
between the wet and dry abrasion loss was observed. For such a dense
concrete, it appears that the effect of moisture becomes more negligible.
References
1 Division of Building Materials, The Norwegian Institute of Technology, NTH,
Trondheim, Norway (unpublished data).
2 Elkem A/S Materials, Kristiansand, Norway (unpublished data).
3 Zhang, M.H. and Gj0rv, O.E. (1990) Development of high-strength light-
weight concrete. High-Strength Concrete, ACI SP-121, 667-81.
4 Sellevold, E.J., Eager, D.H., Klitgaard, J. and Knudsen, T. (1982) Silica
fume-cement pastes: hydration and pore structure, Report BML 82.610.
Division of Building Materials, the Norwegian Institute of Technology, NTH,
19-50.
5 Regourd, M., Mortureux, B. and Gautir, E. (1981) Hydraulic reactivity of
various pozzolans. Fifth International Symposium on Concrete Technology,
Monterey, Mexico.
6 Diamond, S. (1986) The microstructure of cement paste in concrete. Proceed-
ings, 8th Int. Congress on Chemistry of Cement, Vol. I, Rio de Janeiro, 122-47.
7 Maso, J.C. (1980) The bond between aggregates and hydrated cement paste.
Proceedings, 7th Int. Congress on Chemistry of Cement, Vol. I, Paris, VII 1/3
to VII 1/15.
8 Monteiro, P.J.M. and Mehta, P.K. (1985) Ettringite formation on the
aggregate-cement paste interface. Cement and Concrete Research, 15, 2,
378-80.
9 Olliver, J.P. and Grandet, J. (1982) Sequence of formation of the aureole of
transition. Proceedings, RILEM Coloq. Liaisons Pates de Ciment Materiaux
Associes, Tolouse, A14 to A22.
10 Hoshino, M. (1988) Difference of the w/c ratio, porosity and microscopical
aspect between the upper boundary paste and the lower boundary paste of the
aggregate in concrete. Materials and Structures, RILEM, 21, 175, 336-40.
11 Regourd, M. 'Microstructure of high strength cement paste systems,' Very
High-Strength Cement-Based Materials. Materials Research Society, Proceed-
ings, 42, 3-17.
12 Regourd, M., Mortureux, B., Aitcin, P.C. and Pinsonneault, P. (1983)
Microstructure of field concretes containing silica fume. Proceedings, 4th Int.
Symp. on Cement Microscopy, Nevada, USA, 249-60.
13 Sarkar, S.L. and Aitcin, P.C. (1987) Comparative study of the microstructure
of very high strength concretes. Cement, Concrete and Aggregates, 9, 2, 57-64.
14 Aitcin, P.C. (1989) From gigapascals to nanometers. Proceedings, Engr.
Foundation Conf. on Advances in Cement Manufacture and Use. The En-
gineering Foundation, 105-30.
15 Scrivener, K.L., Bentur, A. and Pratt, P.L. (1988) Quantitative characteriza-
tion of the transition zone in high strength concretes. Advances in Cement
Research, 1,2, 230-7.
16 Gj0rv, O.E. (1983) Chemical processes related to concrete. Proceedings,
CEB-RILEM International Workshop on Durability of Concrete Structures,
Copenhagen,341-4.
17 (1976) Regulations for the structural design of fixed structures on the Norwegian
continental shelf. Norwegian Petroleum Directorate, Stavanger.
18 Zhang, M.H. and Gj0rv, O.E. (1991) Effect of silica fume on pore structure
and diffusivity of low porosity cement pastes. Cement and Concrete Research,
Vol. 21,800-8.
19 Diamond, S. (1983) Effects of microsilica (silica fume) on pore-solution
chemistry of cement pastes. Journal of the American Ceramic Society, 66, 5,
C82-C84.
20 Page, C.L. and Vennesland, 0. (1983) Pore solution composition and chloride
binding capacity of silica-fume cement pastes. Materials and Structures,
RILEM, 16,91, 19-25.
21 Nakamura, N., Sakai, M., Koibuchi, K. and lijima, Y. (1986) Properties of
high-strength concrete incorporating very finely ground granulated blast
furnace slag. ACISP-91, 1361-80.
22 Byfors, J., Hansson, C.M. and Tritthart, J. (1986) Pore solution expression as
a method to determine influence of mineral additives on chloride binding.
Cement and Concrete Research, 16, 760-70.
23 Tuutti, K. (1982) Corrosion of steel in concrete, Report fo. 4-82. Swedish
Cement and Concrete Research Institute, Stockholm.
24 (1990) High-strength concrete, state-of-the-art-report, FIP/CEB.
25 CEB Bulletin d'Information No. 182 (1989) Durable concrete structures - CEB
Design Guide, Second edition.
26 Gj0rv, O.E., Vennesland, 0. and El-Busaidy, A.H.S. (1977) Electrical
resistivity of concrete in the oceans. Offshore Technology Conference, Pro-
ceedings, OTC 2803, Houston, Texas, 581-8.
27 Vennesland, 0. and Gj0rv, O.E. (1983) Silica concrete - protection against
corrosion of embedded steel. ACISP-79, Vol. II, 719-29.
28 Danish Great Belt Link (unpublished data).
29 Gewertz, M.W., Tremper, B., Beaton, J.L. and Stratfull, R.F. (1958) Causes
and repair of deterioration to a California bridge due to corrosion of
reinforcing steel in a marine environment. Highway Research Board, Bulletin
182.
30 Browne, R.D. (1980) Mechanisms of corrosion of steel in concrete in relation
to design, inspection, and repair of offshore and coastal structures. ACISP-65,
169-204.
31 Houston, A. and Furguson, P.M. (1972) Corrosion of reinforcing steel
embedded in structural concrete. Centre for Highway Research, The University
of Texas at Austin, Research Report No. 112-1-F, 148 pp.
32 Schiessl, P. (1975) 'Admissible crack width in reinforced concrete structures',
Contribution II 3-17, Inter-Association Colloquium on the Behaviour in
Service of Concrete Structures. Preliminary Reports, Vol. II, Liege.
33 Vennesland, 0. and Gj0rv, O.E. (1981) Effect of cracks in submerged
concrete sea structures on steel corrosion. Materials Performance, 20, 49-51.
34 Gj0rv, O.E. and Bathen, E. (1987) Quality control of the air-void system in
hardened concrete. Nordic Concrete Research, 95-110.
35 Gj0rv, O.E., Okkenhaug, K., Bathen, E. and Husevag, R. (1988) Frost
resistance and air-void characteristics in hardened concrete. Nordic Concrete
Research, 89-104.
36 Siebel, E. (1989) Air-void characteristics and freezing and thawing resistance
of superplasticized air-entrained concrete with high workability. ACI SP 119,
297-319.
37 Okada, E., Hisaka, M., Kazama, Y. and Hattori, K. (1981) Freeze-thaw
resistance of superplasticized concretes. Developments in the Use of Superplas-
ticizers, ACISP-68, 269-82.
38 Foy, C., Pigeon, M. and Bauthia, N. (1988) Freeze-thaw durability and deicer
salt scaling resistance of a 0.25 water-cement ratio concrete. Cement and
Concrete Research, 18, 604-14.
39 Gagne, R., Pigeon, M. and Aitcin, P.C. (1990) Durabilite au gel betons de
hautes performances mecaniques. Materials and Structures, RILEM, 23,
103-9.
40 Malhotra, V.M., Painter, K. and Bilodeau, A. (1987) Mechanical properties
and freezing and thawing resistance of high-strength concrete incorporating
silica fume. Proceedings, CANMET- ACI International Workshop on Conde-
nsed Silica Fume in Concrete, Montreal, Canada, 25 p.
41 Hammer, T.A. and Sellevold, EJ. (1990) Frost resistance of high strength
concrete. ACISP-121, 457-87.
42 Petersson, P.-E. (1984) Inverkan av salthaltiga milj0er pa betongens frostbes-
tandighet, Technical Report SP-RAPP 1984: 34 ISSN 0280-2503. National
Testing Institute, Boras, Sweden.
43 Biczok, I. (1972) Concrete corrosion, concrete protection. Akademiai Kiado,
Budapest.
44 Mather, K. (1982) Current research in sulfate resistance at the Waterways
Experiment Station. George Verbeck Symposium on Sulfate Resistance of
Concrete, ACISP-77, 63-74.
45 Cohen, M.D. and Bentur, A. (1988) Durability of portland cement - silica
fume pastes in magnesium and sodium sulfate solutions. ACI Materials
Journal, 85, 148-57.
46 Mehta, P.K. (1985) Studies on chemical resistance of low water-cement ratio
concretes. Cement and Concrete Research, 15, 6, 969-78.
47 Gj0rv, O.E. (1983) Durability of concrete containing condensed silica fume.
ACISP-79, Vol. II, 695-708.
48 Davis, G. and Oberholster, R.E. (1987) Use of the NBRI accelerated test to
evaluate the effectiveness of mineral admixtures in preventing the alkali-silica
reaction. Cement and Concrete Research, 16, 2, 97-107.
49 Hooton, R. (1987) Some aspects of durability with condensed silica fume in
concretes. Proceedings, CANMET -ACI International Workshop on Silica
Fume in Concrete, Montreal, Canada.
50 Asgeirsson, H. and Gudmundsson, G. (1979) Pozzolanic activity of silica dust.
Cement and Concrete Research, 9, 249-52.
51 Sveinbj0rnsson, S. (1987) Alkali-aggregate demages in concrete with microsilica
-field study. Icelandic Building Research Institute, Reykjavik.
52 Diederichs, U., Jumppanen, U.M., Penttala, V. (1988) Material properties of
high strength concrete at elevated temperatures. IABSE 13th Congress,
Helsinki, June.
53 Pedersen, S. (1981) Beregningsmetoder for varmepdvirkede betonkon-
struktioner. Institute of Building Design, Technical University of Denmark,
Lyngby.
54 Shirley, S.T., Burg, R.G. and Fiorato, A.E. (1988) Fire endurance of high
strength concrete slabs. ACI Materials Journal, 85, 2, 102-8.
55 American Society for Testing and Materials, ASTM, Philadelphia, El 19.
56 Jensen, J.J., Danielsen, U., Hansen, E.Aa. and Seglem, S. (1987) Offshore
structures exposed to hydrocarbon fire. Proceedings, First International Con-
ference on Concrete for Hazard Protection, Edinburgh, Sept, 113-25.
57 Williamson, R.B. and Rashed, A.I. (1987) A comparison of ASTM EI19 fire
endurance exposure of two EMSAC concretes with similar conventional con-
cretes. Fire Research Laboratory, University of California, Berkeley, July.
58 Hertz, K. (1984) Heat-induced explosion of dense concretes. Institute of
Building Design, Technical University of Denmark, Lyngby, Report no. 166.
59 Danielsen, U. (1989) Marine concrete structures exposed to hydrocarbon fires,
Nordic Seminar on Fire Resistance of Concrete. SINTEF Report STF65
A89036, Trondheim, 56-76.
60 (1984) Investigation of surface burning characteristics of fibermesh I fiber
reinforcement in regular and lightweight concrete. Southwest Research Labora-
tory, San Antonio, USA.
61 (1985) Small scale fire tests of fiber-enhanced concrete. Under-Writers Labor-
atories Inc., USA.
62 Chandra, S., Berntsen, L. and Anderberg, Y. (1980) Some effects of polymer
addition on the fire resistance of concrete. Cement and Concrete Research, 10,
367-75.
63 (1989) Norwegian Standard NS 3473, Concrete structures, Design rules. Oslo.
64 Gj0rv, O.E., Baerland, T. and R0nning, H.R. (1990) Abrasion resistance of
high-strength concrete pavements. Concrete International, 12, 1, 45-8.
65 Helland, S. (1990) High-strength concrete used in highway pavements. ACI
SP-121, 757-66.
66 Holland, T.C., Krysa, A., Luther, M.D. and Lin, T.C. (1986) Use of
silica-fume concrete to repair abrasion-erosion .damage in the Kinzua dam
stilling basin. ACISP-91, 841-63.
67 Godfrey, K. (1987) Concrete strength record jumps 36%. Civil Engineering,
84-6 and 88.
O Fracture mechanics
R Gettu and S P Shah
6.1 Introduction
Fracture mechanics is the study of crack propagation and the consequent
structural response. Tremendous research interest in the 1980s led to
fracture models that have been tailored to represent the quasi-brittle
behavior of concrete. The validation of these approaches has opened two
important avenues of application - materials engineering and structural
analysis.
The mechanical behavior of concrete that is designed to have high
strength is different in many aspects from that of normal concrete. These
differences have yielded some characteristics that are not beneficial, such
as brittleness. Fracture models can be used to understand the microstructu-
ral mechanics that control brittleness and crack resistance (toughness), and
to provide reliable means of quantifying them. New high-performance
concretes can then be engineered to possess higher toughness and lower
brittleness. Increased resistance to cracking may also lead to better
durability, long-term reliability and seismic resistance.
The application of fracture mechanics to structural analysis and design is
motivated by the fact that the failure of concrete structures is primarily due
to cracking, and several types of failures could occur catastrophically,
especially in high-strength concrete structures. Certain aspects of such
failures cannot be predicted satisfactorily by empirical relations obtained
from tests, but can be explained rationally through fracture mechanics. In
general, structural analysis based on fracture principles can lead to better
estimates of crack widths and deformations under service loads, the safety
factors under ultimate loads, and post-failure response during collapse.
The present chapter reviews the classical theory of fracture, its exten-
sions for the modeling of concrete behavior, and the microstructural
features that necessitate such models. Material characterization based on
fracture mechanics and its implications are also discussed. Some examples
of the applications are presented to demonstrate the scope of the fracture
approach. More pertinent details can be found in recent publications such
as the two RILEM reports,1'2 the ACI report,3 two detailed
bibliographies4'5 and several conference and workshop proceedings.^13
(a) (b)
Stress-flow
lines
y
(c)
ay
Crack-tip
x
Fig. 6.1 Stress-flow lines for a panel in tension: (a) without crack; (b) with crack; and (c)
stress concentration ahead of the crack
Model Mode Il Modelll
Fig. 6.2 Basic fracture modes
For a sharp notch or a crack, the stress (CT^) in the loading direction and
along the crack plane (y = O) is of the form:
jr
vy= r^— + C1 + C2VjC + C 3 JC+. . . (6.1)
VZTTJC
where C1, C 2 , C3, . . . depend on geometry and loading. Note that the
near-tip (jt^O) stresses depend only on the first term, and at the tip (x = O)
this term and the theoretical stress become infinite (see Fig. 6.Ic). The
parameter K1 in Equation (6.1) characterizes the 'intensity' of the stress-
field in the vicinity of the crack, and is called the stress intensity factor. It
depends linearly on the applied stress (or load), and is a function of the
geometry of the structure (or specimen) and the crack length:
K1 = (jFVm* (6.2)
where a = applied (nominal) stress as in Fig. 6.1b, a = crack length, and
F= function of geometry and crack length. The subscript of K represents
the mode of crack-tip deformation; the three basic modes (see Fig. 6.2) are
Mode I - tensile or opening, Mode II - in-plane shear or sliding, and
Mode III - anti-plane shear or tearing. When two or all three modes occur
simultaneously, the cracking is known as mixed-mode fracture. Only
tensile cracks are discussed in this chapter (unless mentioned otherwise)
since crack propagation in concrete is dominated by Model fracture.
Moreover, analytical treatments of the three modes are quite similar.
A more useful relation for K1, in terms of the load and structural
dimensions, can be obtained by writing Equation (6.2) as:
*, = ^/(«) (6-3)
a
W
d
a
B
(b)
Fig. 6.3 Fracture specimens: (a) single-edge-notched panel in tension; and (b) single-edge-
notched three points bend specimen
The advantage in Equation (6.3) is that it can also be used for several other
geometries, such as the beam in Fig. 6.3b with span/depth = s/d = 4.
Function/, however, is different for this beam:15
^fl.99-a(l-q)(2.15-3.93« + 2.7a 2 )]
/(00 = 6V (6 5)
^P (l + 2a)(l-a)- 1 '
From Equation (6.1) it is obvious that a stress-based criterion for crack
propagation is meaningless since the stress at the crack-tip is always
infinite, irrespective of the applied load. Griffith, 16 in his landmark paper,
suggested a criterion based on energy: Crack propagation occurs only if the
potential energy of the structure is thereby minimized. This thermodyna-
mic criterion forms the basis of fracture mechanics theory. Denoting the
potential energy per unit thickness as C/, the energy release rate is dU/da,
which is usually designated as G (after Griffith). If energy (per unit length
of crack-front) is consumed during fracture at the rate dW/da, denoted as
R (for resistance of the material against fracture), the fracture criterion is:
^ = ^ or G - * (6.6)
da da
There is no fracture when G<R, and the fracture is unstable (i.e., sudden
or catastrophic) when G>R. This criterion does not, however, imply that
crack propagation is independent of the stress-field at the crack-tip. Irwin 17
showed that the energy release rate and the crack-tip fields are directly
related by the equations suggested by Knott 18 :
G = K12IE' (6.7)
where E' = E for plane stress, E' = E/(l — v2) for plane strain,
E = modulus of elasticity, and v = Poisson's ratio. Using Equation (6.7),
the fracture criterion of Equation (6.6) can be written as:
K1=K1R9 K1n = VE7R (6.8)
In the classical theory of linear elastic fracture mechanisms (LEFM), R is
a constant (i.e., independent of the structural geometry and crack length),
usually denoted as G (for critical strain energy release rate). Then, G and
the associated critical stress intensity factor KIc(=\/E'Gc\ also known as
fracture toughness) are material properties. When LEFM governs, the
behavior is termed as Griffith or ideal-brittle fracture. More details of
LEFM are given in textbooks such as Knott18 and Broek.14
The LEFM parameter KIc (or Gc) can be determined from tests of
specimens such as the single edge-notched tension specimen (Fig. 6.3a)
and the three-point bend (3PB) specimen (Fig. 6.3b). Substitution of the
experimentally determined maximum load (fracture load) for P in Equa-
tion (6.3) yields Kf=KIc. Geometry function/(a) can be determined
using elastic analysis techniques including the finite element method.13'14
For several common geometries, f ( a ) can be found in fracture handbooks
such as Tada et al.15 and Murakami.19 Note that in the above-mentioned
specimens, failure occurs immediately, i.e., the crack length is the notch
length (^0) and a = a0 (= relative notch length = aid), at failure.
The main features of LEFM can be summarized as:
1 The fracture criterion involves only one material property, which is
related to the energy in the structure and the near-tip stress field.
2 The stresses near the crack-tip are proportional to the inverse of the
square root of distance from the tip and become infinite at the tip.
3 During fracture, the entire body is elastic and energy is dissipated only
at the crack-tip; i.e., fracture occurs at a point.
In reality, stresses do not become infinite, and some inelasticity always
exists at the crack-tip. However, LEFM can still be applied as long as the
inelastic region, called the fracture process zone, is of negligible size.
Materials where such conditions exist include glass, layered silicate,
diamond, and some high strength metals and ceramics (Chapter 4,
Knott 18 ). For other materials, the applicability of LEFM depends on the
size of the cracked body relative to the process zone. In general, LEFM
solutions can be used when the structural dimensions and the crack length
are much larger than the process zone size. The process zone in concrete
Crack
front
Fig. 6.4 Fracture process zone mechanisms: (a) microcracking; (b) crack bridging; (c) crack
deflection; and (d) crack bowing
(a) Load = 1324 Ib. (b) Load - 1461 Ib. (c) Load - 1601 Ib.
Fig. 6.5 Fracture process zone from holographic interferometry (1 Ib = 4.45 N)39
approach of Horii,42 each process zone mechanism is described individual-
ly. In general, however, the treatment of the process zone need not
distinguish between the different mechanisms; i.e., the toughening effect
of the bridging (wake) zone and the microcracked zone are equivalent.43
Therefore, all the shielding mechanisms can be lumped into a conceptual
fracture process zone that lies ahead of the traction-free zone of the crack
or in the wake of the 'actual' crack-front. This zone can also include the
toughening effects of other inelastic mechanisms such as crack-deflection
(tortuosity of the crack path; Fig. 6.4c) and crack-bowing (unevenness of
the crack front; Fig. 6Ad). Cracks in concrete follow paths of least
resistance, and subsequently, are considerably tortuous with rough crack-
faces. In the hardened cement paste, the crack usually passes around
unhydrated cement grains and along calcium hydroxide cleavage planes.21
The crack-face roughness appears to be less when the cement paste
contains silica fume. 44 In mortar, the crack follows the interfaces between
the sand grains and the hardened cement paste.30 Since the weakest phase
in normal concrete is the aggregate-mortar interface, cracks tend to avoid
the aggregates and propagate through the interfaces.45 In silica-fume
concrete the interfaces and the mortor are much stronger, and therefore,
cracks are less tortuous and sometimes pass through the gravel.27'29'46
Also, cracks propagate through coarse aggregates that are weaker than the
mortar, as in lightweight concrete.46'47 The tortuosity of the crack gives
rise to a higher R due to a larger surface area. Also, further shielding
occurs since the non-planar crack experiences a lower K1 than the
corresponding planar crack.47 Similarly, shielding exists when the crack
front is trapped by bridging particles and bows between the bridges until
they break or slip. This crack-bowing effect (Fig. 6.4d) can be an important
toughening mechanism for very-high strength concrete with strong inter-
faces where cracks propagate through the aggregates. The bowed crack-
front has a lesser K1 than a straight front.32'33
In conclusion, a sizable fracture process zone occurs ahead of a
propagating crack in cementitious materials. Energy is dissipated through-
out this zone causing crack-shielding or toughening. Its size and effective-
ness depend on the microstructure and inherent material heterogeneity;
therefore, toughening in mortar is greater than in hardened cement paste
and less than in concrete. The stress distribution (Fig. 6.6) in the process
V
LEFM stress
ay
Actual stress
Crack-tip
x
Fracture proces zone
Fig. 6.6 Stress distribution in the process zone
zone differs considerably from the LEFM distribution (Fig. 6.Ic). Within
the process zone, the stress and strain undergo 'softening' with a gradual
decrease to zero instead of an abrupt (brittle) drop. Models of fracture in
cementitious materials should account for these factors in order to predict
their behavior satisfactorily.
d = 102mm
Net tensile strength, au (MPa)
d = 200 mm
d = 300 mm
For hep, practically constant values of KIc are obtained (as in other works
such as Gj0rv et 0/. 55 ), but the values for mortar (with sand of maximum
size 0.3 mm) increase with notch length. It appears that the behavior of hep
is much closer to LEFM than mortar. Unless very large specimens are
tested, the observed values increase with notch length until a certain point
and then decrease.60 This trend can be seen in the tests of 3PB specimens
(50 x 100 x 40 mm) of high strength concrete (compressive strength
fc = 124 MPa) by Biolzi et al.61 Their data, in Fig. 6.8b, shows the variation
in LEFM-based KIc with increase in notch depth. It is obvious that
concrete, even of high strength, does not behave according to LEFM in
usual test specimens. It should be mentioned that the effect of notch length
on KIc varies considerably with specimen geometry and size. From tests
Mortar
K|C (N/mm372)
vN=KIcf(a0)^ (6.9)
Since KIc is constant in LEFM and/(a 0 ) is a shape-dependent constant,
Equation (6.9) implies that a^ = (constant)jcl/Vd in geometrically similar
structures. This is the size effect according to LEFM where the failure
stress decreases with increase in certain structural dimensions. Tests of
concrete specimens exhibit a more complex behavior, yielding £/c-values
that may be size-dependent. Therefore, the size effect on failure stress is
not always that of LEFM. From tests of notched beams, Walsh53 recog-
nized a significant trend in this size effect. For small specimens, the failure
stress is constant, as in failure criteria based on limit stress. For very large
specimens, the size effect is the strongest; (jN^l\rd as in LEFM.
Therefore, the structure should be greater than a certain size for LEFM to
apply. Such a transition in failure mode has also been observed in hep and
mortar. 57 The data of Tian et al.58 are shown in Figs. 6.9a and 6.9b for
compact tension specimens of concrete (maximum aggregate
size = 20mm). It is clear that the failure stress and KIc are significantly
size-dependent.
In summary, the fracture of cementitious materials that have a finer
microstructure (e.g., hep) is closer to LEFM than others (e.g., concrete,
fiber-reinforced mortar). The mode of structural failure generally lies
between two limits: the strength limit, and LEFM behavior. When the size
of the initial crack or the size of the uncracked ligament is of the same
order as the inherent material heterogeneity (i.e., much smaller than the
structure), failure is governed by limit criteria. This failure mode also
occurs when the characteristic structural dimensions are small (i.e., of the
same order as the size of the heterogeneities). When the structure is very
large, failure is governed by LEFM.
In order to model concrete failure, several methods have been de-
veloped for determining size-independent fracture properties using practic-
al-size specimens. Also, it is essential that analysis techniques are able to
simulate the observed geometry-effects on cracking and failure before they
can be applied to predict actual structural behavior.
Thickness b = 300mm
the crack, the fracture energies for several brittle materials were deter-
mined.
For LEFM-type behavior, the fracture energy is a constant equal to the
critical strain energy release rate, i.e. Gf= Gc. In practice however, G/is
not always constant. Tattersall and Tappin65 found that the fracture energy
increased with specimen size and decreased with a significant decrease in
stiffness of the testing machine. In the first application of the work-of-
fracture method to concrete, Moavenzadeh and Kuguel66 showed that the
determination of Gf was not straightforward. Since cracking in concrete is
tortuous, they argued that the crack area should be determined exactly
instead of simply using the ligament area. They obtained the crack area
from microscopy techniques, and thereby computed fracture energies that
were almost the same for hep, mortar and concrete. Since the measure-
ment of 'true' crack area is difficult, and to a certain extent subjective, later
researchers have usually circumvented the problem by taking the ligament
(or projected) area as the nominal area of the crack. Then, the fracture
energy is also a nominal value. Moavenzadeh and Kuguel66 used straight
notches for their concrete beams, as have later researcher who have been
able to obtain the entire load-displacement curve using stiff servo-
hydraulic testing machines.
Petersson67 proposed that the work-of-fracture from a notched beam
test would provide a material constant for Gf if the following conditions are
satisfied: (1) energy consumption outside the fracture process zone is
negligible, (2) energy consumption is independent of specimen geometry,
and (3) the fracture is always stable. Due to the effect of toughening or the
increase in R with crack extension (see Section 6.6.2), condition (2) can
never be satisfied exactly. However, as the sizes of ligament and specimen
increase, Gf would tend asymptotically towards a constant value. There-
fore, Gf approaches a material property when the fracture zone is
negligible compared to the specimen size.68 Accordingly, the work-of-
fracture method was recommended by RILEM69 with a lower limit on the
specimen size. For concrete with a maximum aggregate size of 16-32 mm,
the required beam depth is 200mm and the length is 1.2m. The notch
length should be half the depth of the beam. If the total energy consumed
(including the work done by the weight of the beam) is Wf, the fracture
energy is:69'70
Gf=Wf/A^ (6.10)
where A}ig = area of the ligament. In the case of a beam with a straight
notch, Ajig = b (d — a0). The corresponding value of fracture toughness can
be calculated from Equation (6.7). Condition (3) is satisfied by using a
testing machine with high stiffness and/or servo-control.
It has been shown that the specimen sizes recommended by RILEM do
not always provide size-independent values of G/. For concrete, Hillerborg
(1983) concluded on the basis of the fictitious crack model (see Section
6.6.1) that the work-of-fracture method would provide constant values
only at the LEFM limit when the beam was 2-6 m deep and has a notch
longer than 200-400 mm. This would be the point at which condition (2) is
practically satisfied. However, with increasing loads there is some crushing
at the loading points and possible energy dissipation outside the process
zone. Due to this, Petersson's condition (1) may not be satisfied in certain
cases. Nallathabi et al.11 conducted a detailed experimental study of
concrete with a maximum aggregate size of 20mm. The work-of-fracture
of beams with depths ranging from 150 mm to 300 mm was determined. It
was shown that Gf increased with an increase in beam depth, and
decreased with an increase in notch depth. All the values obtained were
much greater than Gc-values calculated from LEFM (Equation (6.3)). By
comparing several studies, Hillerborg72 concluded that in some cases the
work-of-fracture method yielded size-independent values, but in others
doubling the specimen size resulted in a 20% increase in Gf. A 20%
increase in G was also observed for high strength concrete beams by Gettu
et al.73 by increasing the beam depth from 200 mm to 400 mm. Such a size
effect on G/has been observed in compact tension specimens by Wittmann
et al.14 and was attributed to the increase in the width of the fracture zone
with increase in the ligament length. Other possible sources of size effects
and errors in the work-of-fracture method have been described in detail by
Planas and Elices,75 and Brameshuber and Hilsdorf.76
Though the method yields size-dependent values, it has widely been
used due to its simplicity. Availability of extensive data has led to empirical
'code-type' relations linking fracture energy to conventional design prop-
erties. One such equation is provided by the CEB-FIP Model Code:77'78
Gf = xFfcm°-7 (6.11)
where XF is a tabulated coefficient that depends on the aggregate size (e.g.,
for a maximum aggregate size = 16 mm, XF= 6), fcm = mean compressive
strength of the concrete in MPa, and Gf is obtained in N/m. The
commentary to the code cautions that Equation (6.11) gives values that
may be size dependent with deviations up to ±20%. Nevertheless, the
relation is useful when experimental data is lacking.
It should be emphasized that G/, by itself, is not a reliable measure of
toughness or ductility, and that using Gf as the sole fracture parameter
could lead to erroneous conclusions. If one were to conclude from the
observed increase in Gf with compressive strength (as suggested by
Equation (6.11)) that ductility increases with the strength, this would be
wrong. With the use of additional parameters (see Section 6.7), the higher
brittleness in high-strength concrete can be adequately characterized.
Wc w
a0 f
af
ft
Area Gf
W
Wc
where CTOD = opening at the intial crack-tip. The first term in Equation
(6.14) is due to the singularity and the second term gives the energy
dissipated by the cohesive stresses. The use of a singular model will lead to
realistic predictions of crack profiles and near-tip deformations. However,
in the analysis of global structural behavior, singular and non-singular
models give practically the same results.
Crack opening displacement (microns) Predicted crack profile
K1 - O, Linear CP
Predicted length correct
Profiles do not match
Load = 7.9 kN
Measured
Predicted
Elastic
zone
Nonsingular Singular
micro-crack- micro-crack
tip tip
Fracture
process
zone
Macro-crack Macro-crack
tip tip
if\m iw\2m
=l
7
\fj + \w
Hcj <6-15)
where m is a material parameter; for 24MPa concrete, m = 0.27 and for
83 MPa concrete, m = 0.2.
Several other experimental techniques have been proposed for deter-
mining the stress-opening relation. The method of Li,104 based on the
J-integral (Chapter 4 of Broek14), requires the load, displacement and
crack-tip opening of two fracture specimens to be monitored until failure.
Using this data, the/(w) curve including its parameters,/,, wc and G/, can
be obtained.104
The ability to measure deformations continuously over a wide field with
sufficient accuracy by means of optical interferometry has led to ex-
perimental-numerical approaches for determining the constitutive rela-
tions of cracking. In these investigations, deformations near the propagat-
ing crack in a test specimen were obtained using moire interferometry37'105
or sandwich-hologram interferometry.38'39 From these data, the para-
meters of the/(w) curve having a prescribed shape were obtained. In the
works of Castro-Montero et al.39 and Miller et al. ,106 holograms of an area
near the notches of center-cracked plates were made after each loading
step, with a laser light source. The interference fringes observed in a
sandwich of two consecutive holograms, with the same illumination
direction, represent one component of the relative displacement under-
gone during the load step. To obtain the correct displacement vectors three
holograms have to be made at each step with different illumination
directions. An image analysis system was used to analyze the fringes
objectively. The crack-opening and the strain field were thereby computed
at each load step. Using finite element analysis with the FCM, the
parameters of the f ( w ) relation were determined such that the calculated
crack-openings matched those computed from the holograms.
Other experimental-numerical approaches (e.g. Roelfstra and
Wittmann107) use the global load-deformation response of fracture speci-
mens to obtain the parameters of the f ( w ) relation. Wittmann et al.108
assumed a bilinear form for/(w) characterized by four constants/,,/ c ,/i
and W 1 , with a change of slope at (fa ,W 1 ). By fitting the load-displacement
curve from tests of compact tension specimens, they found that several
/(w) relations could produce the same global behavior. However, the
computed fracture energy Gf (as in Equation (6.12c)) was almost the
same. They proposed that, in order to get unique results, the tensile
strength ft should be determined independently and the ratio ft/fa should
be set at a certain value (in the range of 3-5). The resulting stress-opening
relations seem to be independent of specimen size and loading rate.
The global structural behavior that is predicted by fracture analysis can
be strongly influenced by the shape of the f ( w ) relation.107'109'110 From
available data, several code-type formulations have been proposed for the
stress-opening curve. The CEB-FIP Model Code77'78 recommends a
bilinear relation whose parameters are/, Gf, wc and Xf, where wc and XF
(see Equation (6.11)) have tabulated values that depend on the aggregate
size. For a maximum aggregate size of 16mm, wc = 0.15 mm. Values for
Gf can be obtained from Equation (6.11). The change in slope is at
/ = 0.15/,, and the mean tensile strength can be estimated from:
/ = 0.30/C2/3 (6.16)
where fc is the characteristic compressive strength in MPa. A similar
relation has also been formulated by Liaw et al.nl
Some recent works have focused on establishing rational relations
between the microstructure of the concrete and the stress-opening be-
havior of the crack. The relation of Li and Huang112 for/(VM) is formulated
in terms of the aggregate content, maximum aggregate size, fracture
toughness of the hep and the characteristics of the aggregate-hep interface.
Their model assumes that microcracking, frictional debonding of the
aggregates and crack-deflection are the main toughening mechanisms.
Another model proposed by Duda113 incorporates also the aggregate size
distribution through a probabilistic formulation. Note that these models
are only valid for concrete with weak interfaces, and with aggregates that
are stronger than the hep.
/?-curve models
Consider a concrete specimen with an initial crack or flaw. As the load
increases, a process zone at the tip will form and increase in size until the
traction-free crack propagates. An effective crack can be defined such that
its compliance is equal to that of the actual crack including its fracture
zone. Accordingly, this effective crack will extend as the process zone
grows. Since toughening increases with process zone sizes, its effect can
now be modeled very simply by assuming that LEFM relations can be
applied to the effective crack, and that the crack resistance varies with the
effective crack extension. Then, the fracture criterion is G = R as in
Equation (6.6) (or K1 = KIR as in Equation (6.8), where R is not constant
G1R
G3(P3) G2(P2) G1(Pi)
Gc R (toughening)
P
R (LEFM)
a0 Cf
c (crack extension)
Fig. 6.12 The R-curve model
^ DS ^ 8\y) c c
£'(<*o)/g(7) , \ /^10X
f
forc<c r R(c=Gc -, - = \~77-^-y +ao} (6.18a)
g'(VLQ) cf cf g(CLQ) \g'(y) /
foTc^Cf R(c) = Gc (6.18b)
where g(a) = {/(a)}2 (see Equation (6.3)), g'(^) ~ derivative of g(a)
with respect to a, Gc and c/ are the fracture energy and process zone size
defined by the size effect model (see below), and y is a dummy parameter.
Equation (6.18) was derived for an infinitely large specimen where the
process zone develops without restrictions. For finite structures, the
maximum load occurs at c<c/and R<GC. Bazant et a/.118 suggested later
that for finite size, Equation (6.18) is valid until the maximum load and
subsequently the R-CUTVQ is constant. With this modification, Gettu et «/.73
satisfactorily predicted load-deflection curves of high strength concrete,
using Equation (6.18).
Ouyang et al.ul proposed an R-CUYVQ that is the envelope of several
specimens that are geometrically similar and different in size, but with the
same initial notch length. For infinite size, their R-CUTVQ is given as:
*-4'-(7^)(—rv«>*
[ V 1 K-K +l/\ a-a0 J J <6-i9a>
, i+^I±]i+^i_(izi)T' (,I9b)
2 K J4 K \ K / J
where K and £ are functions of « 0 , E, KIc, COTDc and specimen
geometry.126 KIc and COTDC are the critical stress intensity factor and
critical CTOD, respectively, defined according to the two parameter
fracture model (see the next Section). For finite specimens, the plateau of
the 7?-curve begins at the maximum load. Using this model, the load-
deformation response of several fracture specimens have been
predicted.117
The R-CUTVQ can be easily used in nonlinear fracture analysis of
structures when the R(C) and G(a) functions are available. However, the
R-CUTVQ approach has some limitations: (1) R(C) is not a true material
property but also depends on structual geometry, and (2) G (a) functions
cannot be determined without knowing the crack pattern.
CTODc KI — Kfc
p2
'max CT
KI - Kfc
CTODc
effective crack length ac. The associated critical stress intensity factor is
KIcs, and the critical CTOD is CTODC. The fracture criteria of the model
are:
K1 = KIc and CTOD = CTODC (6.20)
These criteria have to be satisfied simultaneously for fracture to occur (see
Fig. 6.13a). The TPFM has also been extended to mixed mode fracture.127
A RILEM128 recommendation describes the procedure for obtaining the
parameters of the TPFM from the load-CMOD response of a 3PB
specimen, where CMOD (for crack-mouth opening displacement) is the
opening of the notch mouth. As shown in Fig. 6.13b, the unloading
compliance C6 is determined just after the peak load Pmax (within 95% of
Pma*), and used along with the initial compliance C/ in LEFM relations to
get ac. The parameters KIcs and CTODC are then computed for the
effective crack at load P max . It has been demonstrated that this method
yields parameters that are practically size-independent.110 Though several
specimen geometries can be used, the RILEM recommendation proposes
the 3PB specimen since the method has been extensively verified with this
geometry. Typical values of the TPFM parameters are given in Section 6.7.
An effective crack model has been proposed by Karihaloo and
Nallathambi 129 ' 130 that is similar in principle to the TPFM. In their model,
the critical length of the effective crack is determined such that the
deflection of the effective specimen is the same as that of the actual
specimen, under peak load Pmax. The KIc-values determined from this
method are comparable to those of the TPFM; typical values are about 31
MPaVmm for normal concrete with/ c = 27 MPa, and about 58 MPaVmm
for 78 MPa high strength concrete.131 See Section 6.7 for more details.
a = (6 21)
" vm?'
V l + p P = T-
d 0 -
where <JN = nominal failure stress (see Equation (6.9)), d = characteristic
structural dimension (for the 3PB specimen, d = depth), and B and d0 are
empirical parameters. Equation (6.21) has the form shown graphically in
Fig. 6.14, which implies that for small sizes, the failure is governed by limit
stress criteria (no size effect), and that for large sizes, the failure is
governed by LEFM ((TTV <* 1/Va).
Since Equation (6.21) relates the failure stress of small specimens to
LEFM behavior, Bazant proposed that the LEFM asymptote be used to
define fracture parameters unambiguously for an infinitely large size. Test
data could then be extrapolated to the limit of d-^> °°, where effects of the
specimen geometry are theoretically absent. This approach was used to
determine fracture parameters for concrete and mortar by Bazant and
Strength
criteria
Size effect
curve
oN/B
d/d 0 = 3
Fig. 6.14 The size effect model
Pfeiffer. 133 This also lead to a reformulation of Equation (6.21) in terms of
two fracture parameters:125
where KIc(=\/E'Gc) and Cf are the fracture toughness and the maximum
process zone length, for the LEFM limit of finite size. (See Fig. 6.12 and
Equation (6.18).) The functions g and g' are the same as in Equation
(6.18). In this model, the effective crack is similar to that of the TPFM.
The SEM has been used to model the fracture of several types of specimens
and structures.73'125
The procedure for obtaining the fracture parameters of the SEM has
been proposed as a RILEM134 recommendation. The peak loads of at least
three sizes of 3PB specimens are needed to calibrate Equations (6.21) and
(6.22), and consequently determine KIc, Gc and cf. Typical values for
these parameters are given in Section 6.7.
8.1 Introduction
It has been shown for quite some time that concrete becomes less
deformable or more brittle as its compressive strength increases. Figure 8.1
shows a high-strength concrete cylinder being tested in compression. The
failure is obviously explosive, indicating that the material is brittle. The
same fact is depicted in a different way by Fig. 8.21 which shows the
Stress, ksi
Strain rate = 10 (ae/sec
fc(ksi)
Fig. 8.2 Complete axial stress versus axial strain curves for normal weight concretes of
different strengths1
axial stress-strain curves and axial-lateral strains curves in compression of
normal weight concretes having different strength levels. Low-strength
concrete obviously can develop only a modest stress level, but it can sustain
that stress over a significant range of strains. High-strength concrete attains
a much higher stress level, but then cannot sustain it over any meaningful
range of strains. The load-carrying capacity drops precipitously beyond the
peak of the stress-strain relationship.
Figure 8.31 shows the stress-strain curves of lightweight concretes having
different compressive strengths. These curves were obtained by Ahmad in
an investigation on mechanical properties of high strength lightweight
Lightweight
Peak
Inflection
point
Stress, ksi
Data points
Strain
Fig. 8.3 Stress-strain curves of lightweight concretes of different strengths1
Normal weight
Lightweight
Stress, ksi
Normal weight
Lightweight
Data points
Strain
Fig. 8.4 Comparison of stress-strain curves of normal weight and lightweight concretes1
ductility index JJL (the ratio of maximum ultimate deflection to the deflec-
tion at the end of the initial linear portion of the load-deflection curve)
decreased drastically. It is evident from Table 8.1 that the ductility index
drops to quite low values for p/pb = 0.69, whereas a ratio of up to 0.75 is
allowed by the ACI Code.3 However, this observation probably requires
some qualification. Table 8.2 from Leslie et al.2 lists the average ductility
indices for increasing values of p. Also listed in the table are average values
of ultimate deflection and yield deflection. First, neither of these terms was
clearly defined in Leslie et al.2 Secondly, as can be seen, the ductility index
decreased with increasing p not so much because the ultimate deflection
decreased as because the yield deflection increased. Thus, without know-
ing precisely how the yield deflection was determined, it is difficult to tell if
this investigation was indicative of any lack of inelastic deformability on
the part of high-strength concrete beams containing moderately high
amounts of tension reinforcement (above 1.5%). It should be noted from
Table 8.1 that ^x values varied widely for the beams having the same tension
reinforcement ratio, and that there was no correlation between this
variation and the variation in concrete strength.
A comprehensive investigation on the deformability of high-strength
concrete beams was carried out by Pastor et al.4 Two series of tests, A and
B, were conducted.
Series A consisted of four beams of high-strength concrete (HSC), one
of medium strength concrete (MSC) and one of low-strength concrete
(LSC). The scope was limited to singly reinforced, unconfined rectangular
members subject to short-term 1/3 point loading. Beam Series A details are
given in Fig. 8.6 and Table 8.3. Concrete compressive strength, fc' (at the
time of testing), and the tensile steel ratio, p, were the experimental
variables.
The scope of Series B was limited to high-strength concrete rectangular
Table 8.1 Specimen details and ductility indices2
fc , PSi b, in. d, in.
Specimen(a) (MPa) (mm) (mm) A(b) P Pb P/Pb
^
7.5-1 9310 8.25 10.63 2#6 0.01 0.0045 0.30 5.9
(64.1) (210) (270)
8.0-1 10,660 8.25 10.63 2#6 0.01 0.046 0.26 8.0
(73.5) (210) (270)
9.0-1 10,620 8.25 10.63 2#6 0.01 0.045 0.27 4.3
(73.2) (210) (270)
7.5-1.5 9720 8.00 10.56 2#7 0.014 0.051 0.44 4.5
(67.0) (203) (268)
8.0-1.5 11,400 8.13 10.56 2#7 0.014 0.051 0.31 2.5
(78.6) (207) (268)
9.0-1.5 11,630 8.50 10.56 2#7 0.013 0.051 0.30 4.2
(80.2) (216) (268)
7.5-2 10,850 8.50 10.50 2#8 0.018 0.040 6.69 3.2
(748) (216) (267)
8.0-2 10,610 7.88 10.50 2#8 0.019 0.040 0.63 2.4
(73.1) (200) (267)
9.0-2 11,780 8.13 10.50 2#8 0.019 0.039 0.56 2.7
(81.2) (207) (267)
7.5-3 11,650 8.38 10.50 3#8 0.027 0.039 0.82 1.9
(80.3) (213) (267)
8.0-3 11,730 8.25 10.50 3#8 0.027 0.039 0.71 2.1
(80.9) (210) (267)
9.0-3 11,210 8.25 10.50 3#8 0.027 0.039 0.77 1.5
(77.3) (210) (267)
(a)
The first number indicates cement content in sacks/cu.yd. The second number
indicates the nominal percentage of longitudinal reinforcement.
(b)
The yield strengths for No. 6, No. 7 and No. 8 bars were 60.22 ksi (415 MPa),
55.83 ksi (385 MPa) and 66.88 ksi (461 MPa), respectively.
stirrup
fc = 9284 psi (64.1 MPa), while the amount of tensile steel was the same
for all beams and consisted of three No. 7 (22mm diameter) deformed
bars.
To facilitate the evaluation of the moment-curvature and load-deflection
data, experimental curves are represented as shown in Fig. 8.8 (Series A)
and Fig. 8.9 (Series B). Key load-deformation values, identified in each
Load point
B1.B3, B5 B2a,
B4,B6 Section B-B Hoop detail
(typ.)
Fig. 8.7 Details of Series B beams tested by Pastor et al.4
Table 8.4 Beam Series B details4
Compressive Age
strength at
Test region dimensions Tensile steel Compressive steel Tie steel(a)
attest, test,
Beam psi days /i, in. ^, in. d, in. d' 7 in. A /,,ksi P plpb AS, /,,PSi P' As", in. 2 Ps
in. 2 in. 2
B-I 8534 186 12.13 6.94 10.69 1.76 62 0.024 0.43 2#6 69 0.012 No. 2 at 12 in. 0.0023
(0.88) (0.05)
B-2a 9284 7 12.25 7.19 10.81 1.70 60 0.023 0.31 3#7 60 0.023 No. 2 at 12 in. 0.0023
(1.80) (0.05)
B-3 8578 189 12.25 6.88 10.81 1.88 3#7 62 0.024 0.46 2#6 69 0.012 No. 2 at 6 in. 0.0047
(1.80) (0.88) (0.05)
B-4 8478 186 12.19 6.69 10.75 1.88 62 0.025 0.36 3#7 62 0.025 No. 2 at 6 in. 0.0047
(1.80) (0.05)
B-5 8516 194 12.06 7.00 10.63 1.69 62 0.024 0.43 2#6 69 0.012 No. 2 at 3 in. 0.0093
(0.88) (0.05)
B-6 8468 190 12.06 6.88 10.63 1.75 62 0.025 0.36 3#7 62 0.025 No. 2 at 3 in. 0.0093
(1.80) (0.05)
(a)
Yield strength = 51 ksi
1 in. = 25.4 mm
1000 psi = 6.895 MPa
Pre-cracking
portion Pre-yielding stage Post-yielding stage
Post-cracking
Failure
portion
Deformation
Fig. 8.8 Idealization of beam Series A load-deformation curves4
Failure
Loss of
concrete cover
Load
Yielding of
tensile steel
Deformation
Fig. 8.9 Idealization of beam Series B load-deformation curves4
Table 8.5 Moment-curvature data and curvature ductility indices for Series A and B beams 4(a ^
Ductility
Moments M (ft - kips) Curvatures 4> x 1(T6 (in." 1 ) Slopes K = I O " 3 (kip-in. 2 ) index
Beam Mcr My M3 M4 Mu *cr *y *3 4>4 *,, KOI K12 K2, K45 M-C
Key
LSC (fc - 3.7 ksi)
MSC (fc - 6.5 ksi)
^c ^ 4>u/<l>y
p - As/bd
Fig. 8.10 Beam Series A curvature and displacement ductility indices versus tensile steel
ratio4
Key
LSC (fi - 3.7 ksi)
MSC (fc - 6.5 ksi)
HSC(8.5 ^fJs=
9.3 ksi)
1/c, mm 1
1/c, inr1
p = As/bd
Fig. 8.11 Inverse values of beam Series A neutral axis depths at failure versus tensile steel
ratio4
bars in the tensile zone with a concrete cover of approximately 0.4 in. or
10 mm. The amount of transverse reinforcement provided in the part of the
beam subjected to variable moment was sufficient to prevent shear
cracking.
To emphasize the different behavior of the two concretes with low,
medium and high percentages of reinforcement, reinforced concrete beams
were prepared with 0.87, 1.97 and 4.61% tension steel. 8.88% steel was
also used for the very high-strength concrete. Balanced steel percentages
were 3.85 and 13.49 for the ordinary and very high-strength concretes,
respectively.
Ductility was represented by the ratio of ultimate curvature to yield
curvature. The yield curvature was defined as corresponding to a tensile
steel strain of 0.2%, and the ultimate curvature to a tensile steel strain of
0.1% or a compressive concrete strain of 0.35%.
Figure 8.12 shows that the beams made with VHSC are more ductile
than those made with LSC when the reinforcement is about or over 1%.
Moreover, the ductility of the beams made with LSC quickly decreases as
the reinforcement ratio increases, whereas that of the beams made with
VHSC declines gently up to a reinforcement ratio of about 5%.
It is worth noting that the curvature ductility of all heavily reinforced
sections (p>p 6 ) theoretically approaches unity regardless of concrete
strength, f c '. This suggests that the increase in JJLC with/ c ' tends to decrease
with p. There seems to be, therefore, a limiting tensile steel ratio beyond
which JJLC values are practically the same regardless of f c r .
Results shown in the lower half of Fig. 8.10 pertaining to \Ld exhibit the
same general characteristics previously described for JUL C . Differences
S 1000
Ductility
S 400
Reinforcement percentage
Fig. 8.12 Curvature ductility versus reinforcement percentage for LSC and VHSC beams5
between \id values due to differences in fcr are, however, much smaller,
mainly because the expected reduction from \LC to JJL^ is greater for the
lightly to moderately reinforced HSC beams than for the LSC and MSC
members. This suggests that the hypothetical p limit beyond which \Ld is no
longer influenced by/ c ' is comparatively smaller than that for JUL C . In any
case, it seems reasonable to assume that JJL^ increases with//, although
much less significantly than JJLC .
The above conclusion emphasizes the difference between material,
sectional and member ductility. The ductility of concrete as a material
depends on the post-peak deformation of its stress-strain response, a
property influenced primarily by fc'. The ductility of a singly reinforced
concrete section, on the other hand, depends on the post-yield deforma-
tion of its momentum-curvature response. Similarly, the ductility of a
singly reinforced beam as a structural unit depends on the post-yield
deformation of its load-deflection response. The two latter characteristics
are associated with the position of the neutral axis at failure, a property
influenced not only by/ c ', but also by p and/^.
As shown in Fig. 8.11, neutral axis depths at failure for the LSC and
MSC beams are greater than the corresponding HSC values. It follows,
therefore, that the decrease in c with// more than compensates for the loss
of material ductility in the HSC range, and helps explain the observed
increase in sectional and member ductility with// for constant p and/^.
Consider Fig. 8.134 where JJLC and \Ld are plotted versus p/p^. It is
interesting to note that differences between LSC, MSC and HSC JJLC values
shown in Fig. 8.13 are greatly reduced compared to those observed in the
upper half of Fig. 8.10. This can be explained in terms of the inverse
proportionality that exists between p/p^ and//. In effect, everything else
being equal, pb will be higher for a beam with higher//. Therefore, a beam
of low // with the same p/p^ ratio as that of a higher strength member
necessarily has less area of steel. Less tensile ateel area implies a shallower
neutral axis depth at failure. Consequently, the difference in ultimate
Key
LSC (fj = 3.7 ksi)
MSC (fj = 6.5 ksi)
HSC (8.5 ss fc sc 9.3 ksi)
Uc = <t>u/<l>y
Note: pb according to
equation 5.1
Ud = AU/Ay
P/Pb
Fig. 8.13 Beam Series A curvature and displacement ductility indices versus tensile steel
ratio expressed as a fraction of balanced steel ratio4
P" W
Fig. 8.14 Influence of compression and transverse reinforcement on the displacement
ductility of doubly reinforced confined HSC beams4
Table 8.7 Properties of beams tested by Fajardo4
Age, /c', /y bxh, L, d, <T, A P, A,' P' P^ Confining
Beam days psi ksi in. ft in. in. in/ O/
/O in/ % % P/P6 hoops
B-7 121 12,800 65.0 7x12 12.0 10.50 1.44 2.35 3.2 1.20 1.6 7.7 0.42 #3 at 12 in.
(3#8) (2#7)
B-8 134 12,650 65.0 7x12 12.0 10.50 1.50 2.35 3.2 2.35 3.2 9.2 0.35 #3 at 12 in.
(3#8) (3#8)
B-9 128 12,800 65.0 7x12 12.0 10.50 1.44 2.35 3.2 1.20 1.6 7.7 0.42 #3 at 6 in.
(3#8) (2#7)
B-IO 133 12,000 65.0 7x12 12.0 10.50 1.50 2.35 3.2 2.35 3.2 8.9 0.36 #3 at 6 in.
(3#8) (3#8)
B-Il 132 12,650 65.0 7x12 12.0 10.50 1.25 2.35 3.2 0.40 0.50 6.6 0.48 #3 at 6 in.
(3#8) (2#4)
1 in. = 25.4mm
1000 psi = 6.895 MPa
p"fy'7fc'<Q.ll exhibited a load-deformation behavior with virtually no
post-spalling response. These beams failed almost immediately after losing
their cover concrete. Since significant inelastic dilatancy of the compress-
ion zone concrete occurs only after loss of the cover concrete, the presence
of longitudinal compression and lateral tie steel has little or no influence on
the ductility of these members. Consequently, the post-yielding load-
deformation behavior and the displacement ductility indices were essen-
tially the same as those of the corresponding singly reinforced unconfined
member (A-4).
On the other hand, the relatively steep slope of the line to the right
suggests that the post-spalling lateral expansion of the compressed con-
crete was restrained by the compression zone reinforcement. For beams
with p"////c'>0.11, therefore, ductility was influenced significantly by the
presence of longitudinal compressive and lateral tie steel. It should be
mentioned that testing of the two beams that gave the highest JUL^ values
(B-8 and B-IO) was halted when excessive deflections could not be
accommodated by the test setup.
The relative efficiency of the longitudinal compressive and the lateral tie
steel in increasing JJL^ is examined in Fig. 8.15. Figure 8.15(a) plots JJL^
versus psfy'7fc' for two values of p7p, while Fig. 8.15(b) plots JJL^ versus
V*Iy Ifc f°r three values of ps. In both cases straight lines are used to
represent the variation between data points.
Consider first Fig. 8.15(a). For a relatively low psfy'7fc' value of 0.0175,
increasing A5' by a factor of two augments \Ld by a factor of almost four.
Consider now Fig. 8.15(b). For a significant range of p'////«,' values,
increasing ps 2.5 times results in a relatively small increase in \Ld. When ps
is increased by a factor of five, the effect of JJL^ is, as expected, more
pronounced. Nevertheless, the increase in JJL^ is still smaller than the
increase obtained by doubling the area of longitudinal compressive steel.
Results in Fig. 8.15 suggest that lateral ties are not as efficient in
Pastor, Nilson and Slate (average f£ = 8.8 ksi)
Fajardo (average \'c = 12.6 ksi)
Au/Ay
Hd =
Foil gauges
(both sides) 3, #7 bars
Whittmore gauges
(both sides)
3 stirrups at 11"c/c
7 bars
cover
f; = 11,500psi(79MPa)
at 60 days
2 bar #3 stirrups bars
at11"c/c
cover
fi= 12,310 psi (85 MPa)
at 60 days
bar 4 stirrups at 3" c/c bars in two rows
Beam 5C
Beam 5B
Beam 5A
Deflection, in.
Fig. 8.17 Load versus midspan displacement of beams tested by Swartz et al.6
specimens in each set, two groups containing four No. 3 (10 mm diameter)
bars, two more groups having four No. 5 (16mm diameter) bars, and the
last two groups being reinforced with four No. 9 (29 mm diameter) bars at
the four corners. The difference between two groups of specimens with the
same concrete strength and longitudinal reinforcement was in the spacing
Compression
reinforcement
Tension
reinforcement
Dial gauge
D.C.D.T.
12200(84.1)
15000(103.4)
p/pb
Fig. 8.19 Flexural ductility, as defined by Equation (8.2), under monotonic loading7^9
is a more practical measure of ductility for beams tested than JJLO , has rather
substantial values.
Figures 8.19 and 8.20 generally show that the same amounts of longitu-
dinal and confinement reinforcement, the ductility indices rise sharply as
the concrete strength increases from 4 ksi (27.6 MPa) to 12 ksi (82.7 MPa)
(nominal values), but then decrease somewhat as/c' increases further from
12 ksi (82.7 MPa) to 15 ksi (103.4 MPa).
The confinement reinforcement spacing, within the range studied, did
P/Pb
Fig. 8.20 Flexural ductility, as defined by Equation (8.3), under monotonic loading7'9
The beam specimen
Load
increment
Fig. 8.22 Deformation sequence for reversed cyclic loading tests by Kamara et al.y-10
Deflection, in.
p/pb
Fig. 8.24 Flexural ductility, as defined in Equation (8.3), under reversed cyclic loading -
upward deflection9'10
ductility as is likely to be required of them in practical situations. It should
be remembered, however, that the specimens tested in the course of
Kamara's investigation were under zero axial load.
For the same amounts of longitudinal reinforcement and confinement
reinforcement, the ductility ratios were found to increase with increasing
concrete strength. For the same concrete strength, the ductility ratio
decreased with increasing amounts of longitudinal reinforcement. Within
the range studied in Kamara's work, the spacing between ties appeared to
have virtually no effect on the ductility of the tested specimens.
fipsi
Ductility index, jif = A f /A y
P/Pb
Fig. 8.25 Flexural ductility, as defined in Equation (8.3), under reversed cyclic loading -
downward deflection9'10
Lightweight concrete beams under monotonic loading
Ahmad and Barker11 reported limited experimental data on the flexural
behavior of high-strength lightweight concrete beams. Flexural tests were
conducted on six singly reinforced beams. Experimental variables were the
compressive strength of concrete (5200 <//<!!,000 psi or
35.9</ c '<75.9MPa) and the reinforcement ratio, p/ph (0.18 <p/
P^ < 0.54). No compression or lateral reinforcement was used in the beams.
A summary of the experimental program is presented in Table 8.8.
Deflection ductility was defined as the ratio of the deflection at ultimate
to the deflection at yielding of the tensile steel. Ultimate was defined as the
stage beyond which it was felt during testing that a beam would not be able
to sustain additional deformation at the same load intensity.
The deflection ductility index JJL^ decreased with an increase in tensile
steel content p (Fig. 8.26 from Ahmad and Barker 11 ). The results of Figs.
8.26 and 8.2711 show that for an approximately equal p/pb ratio, JUL^
decreases with an increase in fc'. As discussed earlier in connection with
Pastor's investigation,4 this is because the value of pb increases with greater
concrete strengths. Therefore, for a constant value of p / p b , a beam with a
higher strength concrete contains more steel than one with a lower strength
concrete, which in turn decreases JJL^. For the beams with compressive
strengths of 5000 and 8000 psi (34.5 and 55.9 MPa), a sharp reduction in \Ld
occurred with an increase in p/pb. However, for beams of 11,000 psi
(75.9MPa) concrete, the value of JUL^ appeared to be less sensitive to
changes in p/pb. The trend indicated that beyond a certain range of p/ph,
LR5-19
LR6-22
Ud
LR11-24
LR5-41
LR8-51
LR11-54
p
Fig. 8.26 Effect of the reinforcement ratio, p, on the displacement ductility of singly
reinforced lightweight concrete beams tested by Ahmad and Barker 11
Table 8.8 Test program11
Rein-
Age at forcing
b, d, testing, /c,(b) £,(0
Ec bar
fc, A,,2
Beam(a) in. (mm) in. (mm) in. (mm) days psi (MPa) psi (MPa) ksi (MPa) detail in. (mm2) P Pb p/pfc
LR5-19 6.0(152.4) 12.0(304.8) 10.25 (260) 2 5470(37.7) 5200(35.9) 3500 (2415) 2#4 0.4 (258) 0.0065 0.0344 0.189
LR5-41 6.0(152.4) 12.0 (304.8) 9.25 (235) 2 5690 (39.3) 5410 (37.3) 3520 (2429) 4#4 0.8 (5160) 0.0144 0.0354 0.407
LR8-22 6.0(152.4) 12.0 (304.8) 10.19(259) 5 8770 (60.5) 8330(57.5) 3760 (2594) 2#5 0.62 (400) 0.0101 0.0454 0.222
LR8-51 6.0(152.4) 12.0 (304.8) 9.13(231) 5 8550 (59.0) 8120 (56.0) 3750 (2588) 4#5 1.24 (800) 0.0226 0.0443 0.511
LRl 1-24 6.0(152.4) 12.0 (304.8) 10.13(257) 49 11560(79.8) 10980 (75.8) 4410 (3043) 2#6 0.88 (568) 0.0145 0.0599 0.242
LR11-54 6.0(152.4) 12.0(304.8) 9.00 (229) 49 11590(80.0) 11010(76.0) 4770 (3291) 4#6 1.76(1135) 0.0326 0.0600 0.543
(a)
Beam nomenclature: for Beam LR5-19, '5' indicates the approximate concrete compressive strength in ksi and '19' indicates reinforcement ratio p/pfo
(b)
Based on 4 x 8-in. cylinder strength
(c)
Using equivalent 6 x 12-in. cylinder strength, assumed to be 95 per cent of 4 x 8-in. cylinder strength
LR5-19
LR8-22
f£ = 5 ksi
Ud
fi = 8 ksi
LR11-24
LR5-41
fi = 11ksi
LR8-51
LR11-54
P/Pb
Fig. 8.27 Effect of the reinforcement ratio, p/p 6 , on the displacement ductility of singly
reinforced lightweight concrete beams tested by Ahmad and Barker11
0.19<p/pb<0.24
0.41 < p/pb < 0.54
LR5-19
LR8-22
Ud
LR5-41
LR11-24
LR8-51
LR11-54
fc
Fig. 8.28 Effect of the concrete strength, fc' fg on the displacement ductility of singly
reinforced lightweight concrete beams tested by Ahmad and Barker
Beam nomenclature
NC9-0.49
NC9-0.22 indicates reinforcement
ratio
fc is in ksi indicates approximate
compressive strength
concrete normal weight
LR5-19 Lightweight concrete beams (present study)
Normal weight concrete beams [10]
Ud
M, = ^ based on
LR11-24 the test data of present study
LR5-41 NC9-0.49
NC9-0.69 NC9-1.10
LR8-51 LR11-54
Pfc
Fig. 8.29 Comparison of ductilities of lightweight and normal weight high-strength singly
reinforced lightweight concrete beams tested by Ahmad and Barker 11
indicates that p has a similar influence on JUL^ for beams made of both types
of concrete. The \Ld values obtained in Ahmad's study11 were lower than
those reported by Pastor and Nilson4 for reinforced normal weight
high-strength concrete beams.
Ahmad and Baits12 developed limited experimental data on the flexural
behavior of doubly reinforced high-strength lightweight concrete beams
with web reinforcement. Flexural tests were conducted on six doubly
reinforced beams. Experimental variables were the compressive strength
of concrete (6700</ c '<ll,060 psi or 46.2</ c '<76.3 MPa) and the rein-
forcement ratio, p/pb (0.16<p/p6<0.47). All the beams had compression
and web reinforcement. The compression reinforcement was kept to
approximately half of the tension reinforcement, and web reinforcement
was provided by No. 2 (6 mm diameter) smooth bars placed as stirrups at a
spacing equal to half the depth of the section. A summary of the
experimental program is presented in Table 8.9.
Deflection ductility was defined as in Ahmad and Barker.11 The index JJL^
decreased with an increase in the tensile steel content p (Fig. 8.30 from
Ahmad and Batts 12 ). The results in Figs. 8.31 and 8.3212 show that for an
approximately equal p/pb ratio, JJL^ decreases with an increase in /c', for
reasons discussed earlier. For beams with concrete strengths of 8000 and
11,000 psi (56 and 77MPa), the ductility decreased relatively less with an
increase in p/pb than for beams with 5000 psi (35 MPa) concrete. Test
results (Fig. 8.33 from Ahmad and Batts12) showed that the ductility
decreases with increasing values of the product pfcf. However, for larger
values of the product p/ c ', the ductility essentially becomes constant.
A comparison of the results presented in Figs. 8.26-8.29 and those given
Table 8.9 Test program12
Compressive Tensile steel
Test region dimensions Web steel
strength at
test/c"a), fc' <b)
P/P6,
Beam psi psi /z, in. b, in. d, in. d'' , in. A 9 , in.2 p in. 2 As', in. 2 P' A5", in.2 P,
LJ-6-16 6700 6380 12 6 10.00 1.90 0.40 0.0067 0.16 0.22 0.0037 0.05 0.012
LJ-7-31 7720 7330 12 6 9.00 2.00 0.80 0.0148 0.31 0.40 0.0074 0.05 0.017
LJ-8-21 8080 7680 12 6 9.94 2.00 0.62 0.0104 0.21 0.40 0.0067 0.05 0.017
LJ-8-44 8360 7940 12 6 8.88 2.10 1.24 0.0233 0.44 0.62 0.0116 0.05 0.023
LJ-1 1-22 11740 11150 12 6 9.88 2.00 0.88 0.0148 0.22 0.40 0.0067 0.05 0.017
LJ-1 1-47 11060 10510 12 6 8.75 2.10 1.76 0.0335 0.47 0.88 0.0168 0.05 0.030
(a)
Based on 4 x 8-in. cylinder strength 1 in. = 25.4mm
(b)
Using equivalent 6 x 12-in. cylinder strength, assumed to be 95 per cent of 1000 psi = 6.895 MPa 4 x 8-in. cylinder strength
The tension and compression reinforcement consisted of ASTM A 615 Grade 60 deformed bar
The stirrups used were No. 2 smooth bars with yield strength of 60 ksi
LJ-6-16 ,d ^ (R = 0.88)
LJ-7-31
Ud
LJ-8-21
LJ-11-22 LJ-8-44
LJ-11-47
P
Fig. 8.30 Effect of reinforcement ratio, p, on the displacement ductility of doubly reinforced
(p' ~0.5p) lightweight concrete beams tested by Ahmad and Batts12
LJ-8-21 LJ-11-22
LJ-8-44
LJ-11-47
1
C
Fig. 8.31 Effect of concrete strength on the displacement ductility of doubly reinforced
(p' ~0.5p) lightweight concrete beams tested by Ahmad and Batts12
LJ-6-16
f^ = 6.7 ksi
LJ-7-31
^d
LJ-8-21
f£ = 8 ksi
LJ-8-44
LJ-11-22
fj = 11 ksi
LJ-11-47
P/Pb
Fig. 8.32 Effect of the reinforcement ratio, p/pb, on the displacement ductility of doubly
reinforced (p' ~0.5p) lightweight concrete beams tested by Ahmad and Batts12
2 018 R
LJ-6-16 ^ = P^7o^-
1
C ( = 0-94)
o LJ-7-31
^d
LJ-8-21 LJ-8-44
LJ-11-22 LJ-11-47
LJ-7-31, data not used in the regression
pfc
Fig. 8.33 Displacement ductility versus p/r' for doubly reinforced (p'« 0.5p) lightweight
concrete beams tested by Ahmad and Batts12
Conclusions
The following conclusions can be drawn with respect to the deformability
of high-strength concrete beams:
1 Although high-strength concrete is a less deformable material than
lower strength concrete, the curvature ductility, JJL C , of a singly
reinforced concrete section increases with/,' for the same value of the
reinforcement ratio, p = As/bd. This is because the neutral axis depth
c decreases with increasing concrete strength, and the decrease in c
with// more than compensates for the loss of material ductility in the
HSC range. For the same//, JULC decreases with increasing p, because
the neutral axis depth, c increases with increasing values of p.
2 The curvature ductility of all heavily reinforced sections (p>p^)
theoretically approaches unity regardless of concrete strength, fc'.
Thus, the increase in |JLC with fc' tends to decrease with p. There
Upward
fc psi (MPa)
Ductility index (ji0)
P/Pb
(a)
Downward
fc psi (MPa)
Ductility index (|i0)
p/pb
(b)
Fig. 8.34 Ductility index JULO versus p/pb for doubly reinforced (p = p') lightweight concrete
beams13
appears to be a limiting tensile steel ratio beyond which JJLC values are
practically the same regardless of fc'.
3 The curvature ductility, JJL C , decreases with increasing values of the
reinforcement ratio, p/p^. At the same p/p/,, the differences between
LSC, MSC and HSC curvature ductility (u,c) values are greatly
reduced compared to those observed for a constant value of p. This is
Upward
fi psi (MPa)
Ductility index fa)
P/Pb
(a)
ft psi (MPa)
(W) xepuj AjjiipriQ
p/pb
(b)
Fig. 8.35 Ductility index (jy versus p/pfe for doubly reinforced (p = p') lightweight concrete
beams13
9.1 Introduction
The successful application of high-strength concrete requires the complete
cooperation of the owner, architect, structural engineer, contractor,
concrete supplier and testing laboratory. In locations where applications of
high-strength concrete have been successfully accomplished, the construc-
tion team has worked together for their mutual benefit and for the benefit
of the engineering community. In the case of buildings, the owner must be
willing to allow the building to become a state-of-the-art structure. The
architect must be willing to design a structure that will utilize the benefits of
high-strength concrete. The structural engineer must have the knowledge
and ability to adapt structural design concepts based on lower strength
concretes for use with higher strength materials. At the same time, the
designer must work within the boundaries of the codes and specifications.
The contractor must be willing to work with different materials and must
accept the need for a higher degree of quality control. The concrete
supplier must be able to supply concrete of the specified strength. Finally,
the quality control testing laboratory must have the capability to prepare
the test specimens and to test them appropriately. This requires that all
members of the team work together from the inception of the project until
its completion. Where this approach has been adopted, the development
and applications of high-strength concrete have been successful; higher
strength materials have been utilized and taller buildings have been built.
As the development of high-strength concrete has continued, the
definition of high-strength concrete in North America has changed. In the
1950s, a compressive strength of 5000 psi (34MPa) was considered high
strength. In the 1960s, commercial usage of 6000 and 7500 psi (41 and
52MPa) concrete was achieved. In the early 1970s, 9000 psi (62MPa)
concrete was being used. In the 1980s, design strengths of 14,000 psi
(97 MPa) were used for commercial applications in buildings. Strengths as
high as 19,000 psi (131 MPa) have been used, although their commercial
application has been limited to one geographic location. The primary
applications of these higher strength concretes have been in the columns of
high-rise buildings. However, there is an increasing interest in the use of
higher strength concretes in long-span bridges and offshore structures. The
following sections of this chapter describe structural design considerations,
construction considerations and quality control aspects and some specific
applications of high-strength concrete. Although the specific applications
are predominantly in North America, there have been applications in other
countries.1"4
Flexural strength
The ACI Building Code6 allows the use of rectangular, trapezoidal,
parabolic or other stress distribution in design provided that the predicted
strength is in substantial agreement with the results of comprehensive tests.
However, in most situations, it is convenient to utilize an equivalent
rectangular compressive stress distribution. Based on published data, 7 it
appears that, for under-reinforced beams, the present ACI methods can be
used for concretes with compressive strengths up to 15,000 psi (103 MPa.)
Additional information is required for concretes with compressive
strengths in excess of 15,000 psi (103 MPa).
Shear strength
Recent tests8 have indicated the need to modify the shear strength design
provisions of the ACI Building Code.6 Test results have indicated that,
with the utilization of higher-strength concretes in reinforced concrete
beams, the specified minimum amount of web reinforcement must be
increased as the concrete compressive strength increases. This has been
found necessary for concretes with compressive strengths in excess of
10,000 psi (69MPa). This increase in the minimum amount of web
reinforcement is needed to control the extent of shear cracking in the
beams and to provide ductile behavior. The 1989 version of the ACI
Building Code6 contains provisions to achieve this.
Development length
Design provisions in the United States permit the use of shorter develop-
ment lengths or anchorage lengths for reinforcing bars as the concrete
compressive strength increases. However, due to a lack of experimental
data, the ACI Building Code6 was modified to limit the design provisions
for development length to concrete with a compressive strength of less than
10,000 psi (69MPa). Consequently, although a concrete with a compress-
ive strength in excess of 10,000 psi (69 MPa) may be used, the design must
be based on a concrete compressive strength of 10,000 psi (69 MPa). This
limitation removes one of the advantages of utilizing higher-strength
concretes. Currently, there does not seem to be any reason why the
limitation should apply. However, data are needed to substantiate removal
of the limitation.
Long-term deformations
The use of high-strength concretes in high-rise buildings requires that
special attention be paid to the long-term length changes that occur in
high-strength concrete members.9 Long-term deformations result from
creep and shinkage. In addition, instantaneous deformations occur
whenever load is added to the building.
As with lower-strength concretes, creep deformations in a member
depend on the creep properties of the concrete at age of loading, stress
level in the concrete, size of member and amount of reinforcement. The
creep per unit stress of high-strength concretes is less than the creep per
unit stress of lower-strength concretes. This means that, at the same stress
level, creep deformations will be less for higher-strength concretes.
Alternatively, for concretes loaded to the same ratio of stress to strength,
the creep deformations will be about the same irrespective of concrete
strength.
Shrinkage of most high-strength concrete is about the same as that of
lower-strength concretes and is more dependent on factors other than the
strength level. Some admixtures are said to reduce shrinkage. However,
the reduced shrinkage may be the result of lower water content in the mix
rather than the use of a specific admixture.
Instantaneous deformations also constitute a major source of shortening
in the lower story columns of high-rise buildings. Instantaneous deforma-
tions are primarily a function of the modulus of elasticity at the age of
loading and can be calculated from the following equation:
(9 l}
ACEC + ASES '
where Ac = area of concrete
As — area of steel reinforcement
Ec = modulus of elasticity of concrete at age of loading
Es = modulus of elasticity of steel reinforcement
P = applied axial load.
The ACI Building Code6 contains the following equation for calculation of
the modulus of elasticity:
l 5 112
Ec = Wc - 33(fc') (9.2)
where wc = unit weight of concrete in Ib/cu ft
fc = compressive strength of concrete as measured on 6x 12-in.
(152 x 305-mm) cylinders.
Equation (9.2) was developed by Pauw10 on the basis of concretes with
compressive strengths up to about 5500 psi (38MPa). As additional data
have become available on higher-strength concretes, various investigators
have made comparisons between the equation and the data for higher-
strength concretes. Based on published data, Martinez et al.n recom-
mended a modified equation for the calculation of modulus of elasticity.
This revised equation predicted a lower modulus of elasticity for higher-
strength concretes when compared with Equation (9.2). However, more
recent data published by Cook12 indicate that Equation (9.2) underesti-
mates the modulus of elasticity for the higher-strength concretes. It should
be noted that the higher-strength concretes reported by Martinez et al.
were obtained using a smaller aggregate size. However, Cook was able to
produce higher-strength concretes with larger-size aggregates.
It should be recognized that Equation (9.2) was based on a statistical
analysis of the available data. As such, there is considerable scatter in the
relationship between modulus of elasticity and concrete compressive
strength. Consequently, it is recommended that, when accurate calcula-
tions are required for high-strength concretes, the modulus of elasticity
should be measured as part of the concrete mix design preparation.
Alternatively, the engineer must assume that there is going to be some
deviation from the values predicted by Equation (9.2) or any other selected
equation.
The value of modulus of elasticity used in Equation (9.1) should be the
modulus of elasticity of the concrete in the column. However, the modulus
of elasticity is measured on smaller specimens. The question, therefore,
arises whether the modulus of elasticity as measured on 6 x 12-in.
(152 x 305-mm) concrete cylinders or other plain concrete specimens is
applicable to large structural members. Currently, there are no published
data on this topic related to high-strength concrete. However, Hester4 and
Cook12 have extracted cores from large concrete members and measured
their modulus of elasticity. For the strength levels used in his program,
Cook showed that the modulus of elasticity as measured on cores varied as
the measured compressive strength varied. If the core strength was low,
the modulus of elasticity was also low. Hester4 also observed the same
phenomena. However, in Hester's work, both strength and modulus of
elasticity of cores were considerably lower than corresponding values
measured on cylinders.
Despite the variations that exist in high-strength concrete and the lack of
information about some of the properties, good correlation has been
obtained between calculated and measured deformations on real
structures.13'14 These correlations have been conducted for structural
members with concrete compressive strengths up to 10,000 psi (69MPa).
There is currently a need to extend these types of correlations to members
with higher concrete compressive strengths.
In the design of 311 South Wacker Drive, special consideration was
given to vertical shortening of different structural elements. In the design
process, column loads were first calculated using a conventional analysis.
The column loads were then used in the calculation of vertical shortening
of column stacks. The calculated differential shortening was then utilized
in subsequent analysis of the forces in the structural frame. If the
differential movements were too large, concrete compressive strengths and
percentages of reinforcement for the columns were revised to minimize the
differential. This iterative analysis process was repeated several times. The
calculated differential movements were then utilized to specify formwork
chamber for the floor slabs. During construction differential movements
were monitored. Very close agreement with design values was obtained.
Heat of hydration
Heat development in cement-rich high-strength concretes can result in high
internal temperatures. Field measurements on 30-in. (760-mm) thick
columns have shown internal temperatures of 150 to 18O0F (66to 82 0 C).
Thus, consideration must be given to minimizing thermal gradients as these
can result in cracking. The most common solutions to this problem are to
keep mixing temperatures as low as possible, use the lowest amount of
cement needed to obtain the specified strength level, use mineral admix-
tures or low-heat-generation cement, and insulate forms as necessary to
maintain a more uniform temperature distribution until concrete strengths
are sufficient to resist thermally induced tensile stresses. Thermal blankets
and expanded polystyrene insulation have been used to insulate the
concrete, even in hot climates.
Testing procedures
As concrete strengths increase, the use of standard 6 x 12-in. (152x305-
mm) cylinders becomes problematic for many laboratories in North
America because test machine capabilities may be insufficient. However,
since building code provisions are referenced to 6 x 12-in. (152 x 305-mm)
cylinders, designers are naturally reluctant to accept test results from
smaller specimens. A potential solution to this dilemma is to develop (by
testing) correlation curves between larger and smaller specimens for the
job concrete mixes. The smaller cylinders can then be used for quality
control with direct correlation to equivalent larger specimen strengths.
This approach was successfully used on the construction of Two Union
Square in Seattle. It should also be noted that, in areas where the use of
high-strength concrete is becoming more prevalent, testing laboratories are
upgrading to higher-capacity machines.
Tests have shown that concrete specimens made in either cardboard,
plastic or tin molds attain lower strengths than specimens made in steel
molds. The Canadian Standards Association (CSA) requires that molds
other than steel are acceptable, if documentation is available and the
cylinders produced from non-steel molds have compressive strengths
equivalent to those obtained using steel molds. The most common practice
is to use plastic molds for lower-strength concretes. It is reasonable to
continue this practice for high-strength concretes. If measured strengths
come into question, the CSA correlation approach can be used to verify
that the mold material is not causing the problem. Re-use of plastic molds
should not be permitted for high-strength concretes.
For high-strength concrete, test strengths are particularly sensitive to
specimen end conditions. The American Society for Testing and Materials
(ASTM) C39,15 Standard Test Method for Compressive Strengths of
Cylindrical Concrete Specimens, provides guidance on perpendicularity
and planeness, and ASTM C 617,16 Standard Practice for Capping Cylin-
drical Concrete Specimens, covers capping procedures. High-strength
capping compounds should be used with a uniform thickness of Jin.
(3mm). As an alternative to capping, cylinders can be lapped to meet
ASTM end-condition requirements, but this procedure is generally more
costly. If lapping is used, it should be done a day or two after casting
because lapping becomes more difficult as the concrete gains strength.
Compression testing machines should meet the requirements of ASTM
C39,15 or other applicable standards. Careful attention should be paid to
platen smoothness as small deviations can cause increased errors in
measured strengths. It is recommended that machines be calibrated every
six months rather than annually when used for high-strength concrete
testing.
In-place strengths
Several national codes address the use of concrete core data when
strengths of concrete quality control specimens do not exceed the specified
strengths. For example, the ACI Building Code6 allows concrete to be
accepted if the average strength measured on the cores exceeds 85% of the
specified strength and no single strength is less than 75% of the specified
strength. The factors of 75 and 85% are based on lower-strength concretes
and were determined from comparisons of core strengths with cylinder
strengths. Data on concretes with compressive strengths up to 17,000psi
(83MPa), have indicated that the factors are still valid. However, in at
least two projects using concrete with specified strengths in excess of
12,000 psi (83 MPa), the concrete core strengths have been lower than 85%
of the cylinder strengths. This has raised the question of the validity of the
factors for high-strength concrete. Designers and contractors are cautioned
to address this issue before construction begins.
8% steel
4% steel
1% steel
(52MPa) concrete at the time Lake Point Tower was built, the use of
reinforcement with a yield stress of 75 ksi (520 MPa) was also required.
Construction of the first story of the building began in June 1966 and the
70th story was completed in December 1967. Weather permitting, the
building was built at the rate of one floor every three working days. The
Fig. 9.2 Lake Point Tower
Table 9.2 Properties of interior columns and core walls at Lake Point Tower
Concrete Grade of Column Core wall
design strength reinforcement diameter thickness
Story psi ksi in. in.
68-59 3,500 60 30 0,12
58-44 5,000 60 30 12
43-35 5,000 60 36 12,14
3^30 6,000 60 36 16
29-17 6,000 60 40 16, 18, 20
16-12 7,500 60 40 22
11-1 7,500 75 40 24,30
Metric equivalents:
1000 psi = 1 ksi = 6.895 MPa
1 in. = 25.4mm
Fig. 9.3 Water Tower Place
9.6 Bridges
Without doubt, the largest use of high-strength concrete has been in the
columns of buildings. However, high-strength concrete is receiving more
and more attention for use in bridge structures. Three examples of the
usage in North America are illustrated in Fig. 9.7. The tensile strength of
high-strength concrete increases with compressive strength. This is bene-
ficial in the design of prestressed concrete members such as bridge girders
where the tensile strength may control the design. The reduced creep of
high-strength concrete is also beneficial in reducing prestress losses in
bridge girders. Consequently, utilization of high-strength concrete results
in economies in prestressed concrete girders. Table 9.4 lists a selection of
bridges that have utilized concrete strengths of 6000 psi (42MPa) or
greater.7
1. Lightweight concrete
Joigny Bridge
The Joigny Bridge is an experimental bridge designed to demonstrate the
possibility of producing high-strength concrete bridges using existing
batching plants and local aggregates.37 The bridge is a three-span structure
with span lengths of 111, 151 and 111 ft (34, 46 and 34 m). In cross section,
the bridge is a double-T with a deck width of 49 ft (15 m). The bridge was
designed based on a concrete with a characteristic compressive strength of
Fig. 9.8 Glomar Beaufort Sea 1
8700 psi (60 MPa) after 28 days and utilizes external longitudinal prestres-
sing. This bridge was part of the French program to introduce high
performance concretes into bridge construction.
Miscellaneous
Miscellaneous applications of high-strength concrete have been summa-
rized by ACI Committee 3637 and CEB/FIP.1 For example:
Precast panels for dam - 9000 psi (62 MPa)
Prestressed concrete poles - 10,000 psi (69 MPa)
Grandstand roofs - 7500 and 8850 psi (52 and 61 MPa)
Marine foundations - 8000 psi (55 MPa)
Underwater bridge - 9400 psi (65 MPa)
Grandstand elements - 8700 psi (60 MPa)
Avalanche shelters - 10,900 psi (75 MPa)
Piles - 10,900 psi (75 MPa)
Acknowledgements
The author acknowledges information provided by the designers and
contractors for the various structures described in this chapter. Portions of
this chapter were written by S.H. Gebler and D.A. Whiting of Construc-
tion Technology Laboratories. Photographs were provided by the Portland
Cement Association.
References
1 CEB/FIB Working Group on HSC (1990) High strength concrete - state of the
art. The Institution of Structural Engineers, London.
2 Burnett, LD. (1989) High-strength concrete in Melbourne, Australia. Con-
crete International Design and Construction. American Concrete Institute, 11,
No. 4, April, 17-25.
3 Holand, L, Helland, S., Jakobsen, B. and Lenschow, R. (1987) Utilization of
high strength concrete, Proceedings, Symposium in Stavangar, Norway, June
15-18. Tapir, Trondheim, 688 pp.
4 Hester, W.T. (1990) High-strength concrete, second international symposium,
Publication SP-121. American Concrete Institute, Detroit, 786 pp.
5 Russell, H.G. (1990) Use of high-strength concretes. Building Research and
Practice, 18, No. 3, May/June, 146-152.
6 ACI Committee 318 (1989) Building Code Requirements for Reinforced
Concrete. American Concrete Institute, Detroit, 353 pp.
7 ACI Committee 363 (1984) State-of-the-art report on high strength concrete.
ACI Journal, 81, No. 4, July/August, 364^411.
8 Roller, JJ. and Russell, H.G. (1990) Shear strength of high-strength concrete
beams with web reinforcement. ACI Structural Journal, 87, No. 2, March-
April, 191-8.
9 Russell, H.G. (1985) High-rise concrete buildings: shrinkage, creep and
temperature effects, Analysis and design of high-rise buildings, Publication
SP-97. American Concrete Institute, Detroit, 125-37.
10 Pauw, A. (1960) Static modulus of elasticity of concrete as affected by density.
Journal of the American Institute, Proceedings 32, No. 6, December, 679-787.
11 Martinez, S., Nelson, A.H. and Slate, P.O. (1982) Spirally-reinforced high
strength concrete columns, Research Report No. 82-10. Department of
Structural Engineering, Cornell University.
12 Cook, J.E. (1989) Research and application of 10,000psi (//) high-strength
concrete, Concrete International Design & Construction. American Concrete
Institute, 11, No. 10, October, 67-75.
13 Pfeifer, D.W., Magura, D.D., Russell, H.G. and Corley, W.G. (1971)
Time-dependent deformations in a 70-story structure, Designing for effects of
creep, shrinkage and temperature in concrete structures, Publication SP27-7.
American Concrete Institute, Detroit, 159-85.
14 Russell, H.G. and Corley, W.G. (1977) Time-dependent behavior of columns
in Water Tower Place, Douglas McHenry international symposium on concrete
and concrete structures, Publication SP-55-14. American Concrete Institute,
Detroit. Also printed as PCA Research and Development Bulletin
RD052.01B, Portland Cement Association, Skokie, Illinois, 10 pp.
15 Standard test method for compressive strengths of cylindrical concrete speci-
mens, ASTM C 39. American Society for Testing and Materials, Philadelphia.
16 Standard practice for capping cylindrical concrete specimens, ASTM C 617.
American Society for Testing and Materials, Philadelphia.
17 Schmidt, W. and Hoffman, E.S. (1975) 9000-psi concrete - why?, why not?
Civil Engineering, ASCE, 45, No. 5, May, 52-5.
18 Giraldi, A. (1989) High-strength concrete in Washington, D.C. Concrete
International Design and Construction. American Concrete Institute, 11,
No. 3, March, 52-5.
19 Russell, H.G. and Larson, S.C. (1989) Thirteen years of deformations in
Water Tower Place. ACI Structural Journal, 86, No. 2, March-April, 182-91.
20 Colaco, J.P. (1985) 75-Story Texas Commerce Plaza, Houston - the use of
high-strength concrete. High-Strength Concrete, Publication SP-87. American
Concrete Institute, Detroit, 1-8.
21 Pickard, S.S. (1981) Ruptured composite tube design for Houston's Texas
Commerce Tower. Concrete International Design & Construction. American
Concrete Institute, 3, No. 7, July, 13-19.
22 (1983) Tower touches few bases. Engineering News Record, June 16, 24-25.
23 LeMessurier, WJ. (1982) Toward the ultimate in composite frames, Building
Design and Construction. Cahners Publication Company, 23, No. 11, Novem-
ber, 14-21.
24 Case history report - the world's tallest concrete skyscrapers, Bulletin No. 40.
Concrete Reinforcing Steel Institute, Schaumburg, IL, 8 pp.
25 (1989) Tall concrete buildings come of age, Engineering News Record,
November 30, 25-27.
26 Page, K.M. (1990) Pumping high-strength on world's tallest concrete building,
Concrete International Design & Construction. American Concrete Institute,
12, No. 7, January, 26-8.
27 Moreno, J. (1990) 225 W. Wacker Drive, Concrete International Design &
Construction. American Concrete Institute, 12, No. 1, January, 35-9.
28 (1989) Put that in your pipe and cure it. ENR, February 16, 44-53.
29 Thornton, C.H. (1990) The world's tallest building - Chicago's Miglin-Beitler
Tower. Engineered Concrete Structures, 3, No. 3, Portland Cement Associa-
tion, December, 1-2.
30 Thornton, C.A., Hungspruke, O. and DeScena, R.P. (1991) Looking down at
the Sears Tower. Modern Steel Construction. American Institute of Steel
Construction, August, 27-30.
31 Jobse, HJ. (1981) Applications of high-strength concrete for highway bridges,
Executive Summary, U.S. Department of Transportation, Federal Highway
Administration, Washington, DC, Report No. FHWA/RD 81/096, October,
27pp.
32 Jobse, HJ. and Moustafa, S.E. (1984) Applications of high strength concrete
for highway bridges. Journal of the Prestressed Concrete Institute, 29, No. 3,
May-June, 44—73.
33 Zia, P., Schemmel, JJ. and Tallman, T.E. (1989) Structural applications of
high strength concrete, Report No. FHWA/NC/89-006, Center for Transporta-
tion Engineering Studies, North Carolina State University, Raleigh, June, 330
pp.
34 (1984) Hybrid girder in cable-stay debut. Engineering News Record, Novem-
ber 15, 32-6.
35 Taylor, P.R. and Torrejon, J.E. (1987) Annacis Bridge - design and construc-
tion of the cable-stayed span. Quarterly Journal of the Federation Interna-
tionale de Ia Precontrainte, 4, 18-23.
36 (1986) Stayed girder reaches a record with simplicity. Engineering News
Record, May 22.
37 Pliskin, L. and Malier, Y. (1990) The French R&D Project, 'New develop-
ments for concrete, the high strength concrete Bridge of Joigny', Preprint No.
89-0586, Transportation Research Board, 69th Annual Meeting, Washington,
D.C., January.
38 Ronneberg, H. and Sandvik, M. (1990) High strength concrete for North Sea
platforms, Concrete International Design & Construction. American Concrete
Institute, 12, No. 1, January, 29-34.
39 Fiorato, A.E., Person, A. and Pfeifer, D.W. (1984) The first large scale use of
high-strength lightweight concrete in the Arctic environment, Second Sympo-
sium on Arctic Offshore Drilling Platforms, Houston, Texas, April.
^j f\ High strength
A vf lightweight aggregate
concrete
T A Holm and T W Bremner
10.1 Introduction
It may be argued that the first practical use of high strength concrete took
place in World War I when the American Emergency Fleet Corporation
built lightweight concrete ships with specified compressive strengths of
5000 psi when commercial normal weight concrete strengths of that time
were 2000 psi. It was fully recognized by these forward looking engineers
that high self-weight was the major impediment in the use of structural
concrete.
Concrete density can be reduced in several ways, i.e., lightweight
aggregates, cellular foams, high air contents, no fines mixes, etc., but only
high quality structural grade lightweight aggregates can develop high
strength lightweight aggregate concretes. As such, the letter 4 A' in the
abbreviation LAC for lightweight aggregate concrete will be dropped and
similarly HSLC indicates high strength lightweight aggregate concrete. In a
similar fashion, LWA, NWA, NWC and HSNWC represent lightweight
aggregate, normal weight aggregate, normal weight concrete and high
strength normal weight concrete respectively.
Structural efficiency
The entire hull structure of the USS Selma was constructed with HSLC in a
shipyard in Mobile, Alabama and launched in 1919. The strength/density
(S/D) ratio (structural efficiency) of 50 used in the USS Selma (4-5000
psi/100 pcf) was extraordinary for that time.1 Improvements in structural
efficiency of concrete since that time are shown schematically in Fig. 10.1,
revealing upward trends in the 1950s with introduction of prestressed
specified compressive strength (psi)
Proposed
offshore
"Selma" SLWC
• WW1 structures
SLWC
ships
°°
"High quality
Ratl
concrete"
Commercial concrete
Fig. 10.1 The structural efficiency of concrete. The ratio of specified compressive strength
density (psi/pcf) through the recent history of construction (from Holm and Bremner 1 )
11020mmSolite(Flyash)
Strength ceiling
maximum attainable
Strength - Compressive or tensile
concrete strength
as limited by the
HSLW coarse LWA
Fine "Paste"
aggregate Cement, water
& air voids
Coarse
aggregate
inclusion
magnitude that extensive microcracking may take place before any super-
imposed stress is applied.
Natural aggregates have an extremely wide range of elastic moduli
resulting from large differences of mineralogy, porosity, flaws, lamina-
tions, grain size and bonding. It is not uncommon for a fine-grained
diabase rock to have an elastic modulus greater than 13 x 106 psi (90 GPa)
while poorly bonded, highly porous natural aggregates have been known to
have values lower than 3 XlO 6 psi (2OGPa). Aggregate description by
name of rock is insufficiently precise, as demonstrated in one rock
mechanics text which reported a range of elastic modulus of 3 to 10 x 106
psi (20 to 69 GPa) for one rock type.9
Figure 10.6 illustrates compressive strength and stiffness characteristics
reported for several rock types and compares these wide ranges with the
modulus of elasticity of concrete as suggested by equation Ec ^33(O1 5Vj^7
(in psi) of ACI 318-89 Code. The ratio of the coarse aggregate modulus to
that of the concrete composite can be shown to be as much as 3, signaling a
further difference between the two interacting phases (mortar and coarse
aggregate) of as much as 5 to 1. That the strength-making potential of the
stone or gravel is normally not fully developed is evident from visual
examination of fracture surfaces of concrete cylinders after compression
testing. The nature of the fracture surface of concretes is strongly
influenced by the degree of heterogeneity between the two phases and the
extent to which they are securely bonded together. Shah10 reported on the
profound influence exerted by the contact zone in compressive strength
tests on concretes in which aggregate surface area was modified by
coatings. The degree of heterogeneity and the behavior of the contact zone
between the two phases are the principal reasons for the departure of some
concretes from estimates of strength based upon the water-to-cement (w/c)
ratio. As has been suggested,8 undue preoccupation with the matrix w/c
ratio may lead to faulty estimates of compressive strength and even greater
misunderstanding of concrete's behavior from durability, permeability and
tensile type loading conditions.
Obviously the characteristics of the NWA will have a major effect on
elastic compatibility. The interaction between the absolute volume percen-
tage of coarse aggregate (±35%) and the mortar phase (±65%) will result
in a concrete with a modulus intermediate between the two fractions. At
usual commercial strength levels the elastic mismatch within structural LC
is considerably reduced due to the limited range of elastic properties of
usual LWA particles.
Stone aggregate
Towards elastic
Elastic mismatch matching of stone,
between stone, C.A.and mortar 1 LTWT coarse agg. particle
coarse aggregate 2 A/E sanded Itwt concrete
Elastic
mismatch Itwt.
coarse agg. and
Elastic compatability A/E mortar
of Itwt. coarse agg.
and mortar phases
Cementitious materials
Cement
Portland cements used for HSC should conform to the requirements of
ASTM C150. Granulated iron blast furnace slags used as a replacement for
Portland cement should conform to ASTM C989. Compressive data on
cements, supplementary cementitious materials, admixtures, etc. are re-
ported on in ACI363.13
Supplementary cementitious materials
Production of ultra-HSLC generally requires the use of supplementary
cementitious materials. High quality fly ash meeting the requirements of
ASTM C618 will reduce permeability, improve placing qualities, lower
heat of hydration, and improve long term strength characteristics. Microsi-
lica will improve compressive strength at all ages and also provide
significantly improved resistance to chloride penetration.
Supplementary cementitious materials function very effectively in
HSLC, because pozzolanic activity requires the combination of the calcium
hydroxide liberated during cement hydration with finely divided silica in
the presence of moisture. As shown in Fig. 10.2, mixes incorporating
mineral admixtures achieved higher strength ceilings than the control
concretes. Favorable hydrating environments will be provided for a longer
time due to the internal curing provided by the LWA absorbed moisture,
thus promoting increased activity of the pozzolanic materials.
Admixtures
When used in HSLC, admixtures offer reduced water demand, enhance
durability, and improved workability in a manner comparable to that of
HSNWC. Water reducers, retarders and high range water reducers should
conform to ASTM C494 and be dosed according to manufacturers'
recommendations.
Air entrainment
LC mixtures, normally contain entrained air. Entrained air serves to
increase the cohesiveness of the fresh concrete mix, and to make concrete
resistant to the effects of freezing and thawing when in a wet environment.
When freezing and thawing is not a consideration, then small amounts of
entrained air (3 to 5%) are adequate. Entrained air volumes should meet
the requirements of ACI201 according to the severity of the exposure
conditions. While air entrainment will diminish strength making character-
istics of the cementitious matrix, it will also lower water and sand volumes
necessary to achieve satisfactory workability, with the net effect being only
a modest reduction in the strength of HSLC.
Coarse aggregate
HSLCs normally require only coarse LWA. As reported earlier, most, but
not all, HSLC mixes require a reduction of the LWA top size, particularly
in the 7000 to 10,000 psi (48 to 7OMPa) range. Certain LWAs, however,
because of the strength of the vitreous material enveloping the pores, have
routinely used the 3/4 to #4 (20 to 5 mm) gradation in production of high
strength precast concrete for more than four decades. Most LWA manu-
facturing plants will limit coarse aggregate to two sizes to minimize
production and stockpiling problems, but these plants will entertain other
gradations if project volumes warrant.
Gradations of 3/4 to #4 (20 to 5 mm) or 1/2 to #4 (13 to 5 mm) will
normally be appropriate for usual size HSLC members while 3/8 to #8 (10
to 5 mm) gradations may be necessary in highly reinforced members to
allow adequate placement conditions.
Fine aggregate
HSLC normally incorporates normal weight sand as the fine aggregate
fraction. Quality criteria developed for sands used in HSNWC (e.g., FM of
about 3.0 for optimum workability and strength, etc) are identical to those
used in manufacturing HSLC.
Tensile strength
Tensile strength of HSLC is limited by the fact that approximately 50% of
the aggregate volume is pore space. The ACI code (ACI 318-89) requires
LWA producers to supply tensile splitting test data on concrete incorporat-
ing their aggregate, allowing structural engineers to modify code equations
for shear, torsion and cracking.
High tensile splitting developed on mature specimens of HSLC have
shown clearly visible high moisture contents on the split surface, demon-
strating that well compacted mixes with high binder content, and particu-
larly those incorporating mineral admixtures (microsilica, fly ash), are
essentially impermeable and will release moisture very slowly. High
strength specimens drying in laboratory air for over several months were
still visibly moist over 90% of the split diameter.5 The reductions in
splitting strength observed in tests on air dried commerical strength LC
that are caused by differential dry moisture gradients in the concrete prior
to reaching hygro-equilibrium are significantly delayed and diminished in
high binder content HSLC.
Elastic properties
Modulus of elasticity
Concrete is a composite material composed of a continuous matrix
enveloping particulate inclusions. Stiffness of the composite is related to
the stiffness of its constituents in a rather complex way, and it is surprising
that the recommended ACI formula has been so effective. One factor
affecting stiffness is the variation of aggregate modulus of elasticity within
a particular density range. At the same specific gravity, LaRue found the
modulus of elasticity of natural aggregates could vary by a factor of as
much as three.42 Concrete strengths also tended towards a maximum
where the aggregate modules matched the modulus of the concrete made
from them.
Although the ACI 318 formula, Ec = 33 ^1 -5Vf? (Ec = 0.04Sw1 V/V)
has provided satisfactory results in estimating the elastic modulus of NWC
and LC in the usual commercial strength range from 3000 to 5000 psi (20 to
35 MPa), it has not been adequately calibrated to predict the modulus of
high strength concretes. Practical modification of the formula was first
provided by ACI 213-7743 to more reasonably estimate the modulus of
HSLC:
Ec = Cul-5Vf?
C = 31 for 5000 psi (C = 0.40 for 35 MPa)
C = 29 for 6000 psi (C = 0.38 for 41 MPa)
When designs are controlled by elastic properties (e.g. deflections, buck-
ling, etc), the specific value of Ec should be measured on the proposed
concrete mixture in accordance with the procedure of ASTM C 469,
Standard Test Method for Static Modulus of Elasticity and Poisson's Ratio
of Concrete in Compression.
In general, structural grade rotary kiln produced LWAs have a compara-
ble chemical composition and are manufactured under a similar tempera-
ture regime. They achieve low density by formation of a vesicular structure
in which the vesicles are essentially spherical non-interconnected pores
enveloped in a vitreous matrix. It would be expected that, with such
similarities, the variability in stiffness of the aggregate would be principally
due to the density as determined by the pore volume system. Aggregate
density, in turn, is reflected in a reduced concrete unit weight which is
accounted for in the first term in the above equation.
As with NWC, increasing matrix stiffness is directly related to matrix
strength which, in turn, affects concrete strength. When large percentages
of cementitious materials are used, the LC strength ceiling may be reached
causing the above equation to overestimate the stiffness of the concrete.
Poisson's ratio
Testing programs investigating the elastic properties of HSLC have
reported an average Poisson's ratio of 0.20, with only slight variations due
to age, strength level, or aggregates used.44'45
Dimensional stability
Shrinkage
Figure 10.9 demonstrates the shape and ultimate shrinkage strains from
Steel
Appied
beam
load
Concrete
cylinder
Fracture
107 20 mm LWC
100 20 mm LWC
108 15mm stone
Age (days)
Fig. 10.9 Shrinkage of high strength lightweight and stone concretes
molds. Brass wafers were epoxied at one day to the bar surface at a 10 in.
(250mm) gage distance with mechanical measurement by a Whitemore
gage. Reference readings were established 7 days after fabrication after
which specimens were allowed to dry in laboratory air, 7O 0 F (210C),
(50% ± 5 RH), with no further curing. Shrinkage and weight readings were
taken weekly for three months then monthly with results shown to one
year. Ten year shrinkage strains were only slightly higher than one year
results and will be reported in another publication.47
Creep
Rogers reported that the one year creep strains measured on several high
strength North Carolina and Virginia LCs were similar to those measured
on companion NWC.48 Greater creep strains measured on HSLC contain-
ing both fine and coarse LWA, when compared to reference HSNWC in
reports by Reichard44 and Shideler,45 could be anticipated because of the
larger matrix volume required because of the angular particle shape of the
LWA fines.
While the Prestressed Concrete Institute provides recommendations for
increasing stress losses due to creep when using HSLC, it may be advisable
to obtain accurate design coefficients for long span HSLC structures by
conducting pre-bid laboratory tests in accordance with the procedures of
ASTM C512, Standard Test Method for Creep of Concrete in Compression.
As reported in ACI 213,43 and shown in Fig. 10.10, specific creep values
decrease significantly with increasing strength of LC. Additionally, at
higher strength levels, the creep strain envelope developed from a wide
range of LCs tested converged towards the performance of the reference
NWC.
Thermal properties
Accurate physical property input data is essential when considering the
thermal response of restrained members in exposed structures. Obvious
Depth of carbonation (mm)
Japan
structures Concrete
NY MD ships
bridge bridge
Time in years
Fig. 10.10 Measured depth of carbonation (mm) of exposed lightweight concrete structures
cases in point include exposed exterior columns of multi-story cast-in-place
concrete frames as well as massive offshore concrete structures constructed
in temperate zones and then towed to harsh Arctic marine environments.
As thermal behavior of concrete is a function of the thermal contributions
of the separate components, the low expansions, conductivities and
diffusivities of HSLC are predictable in the light of thermal measurement
data reported on regular strength LC.43 A summary of data relating to the
coefficient of linear thermal expansion, thermal conductivity, and diffusiv-
ity data extracted from the results obtained in the joint industry investiga-
tion digested by Hoff 17 is shown below.
• Coefficient of linear thermal expansion
At high moisture contents typical of marine exposed concrete, the
coefficient of linear thermal expansion for HSLC ranged from 4 to 7
microstrain/°F (7 to 13 microstrain/°C).17
• Thermal conductivity
Over a temperature range of -50 to 7O 0 F (-46 to 210C), thermal
conductivity values of the HSLC varied from 5.6 to 7.6 Btu.in/hr.sf. F
(0.8 to 0.9 W/m. K) According to density and mix composition
factors.17
• Specific heat
The specific heat of HSLC tested averaged 0.23 Btu/lb. F (1.0 J/
kg K).17
• Thermal diffusivity
The thermal diffusivity of the HSLC tested ranged from 0.020 to 0.022
sq ft/hr (18.3 to 20.8 sq cm/hr) for the concrete densities tested.17
Durability
Numerous laboratory durability testing programs evaluating freeze/thaw
resistance have confirmed field observations of the long-term, proven
performance of highly exposed HSLC. The United States Federal Highway
Administration closely examined 12 study cases of more than 400 North
American LC bridge decks and concluded:
'Some have questioned the durability, wear resistance, and long-term
freeze/thaw qualities of lightweight concrete. No evidence was found
that these properties differ from those of normal weight concrete. In
fact, there is evidence that these properties could be better for
lightweight concrete, especially if the normal weight concrete is of
poor quality. This leads to the suggestion that the designer might
consider specifying lightweight concrete if natural aggregates are not
of high quality. Although lightweight aggregates vary depending on
the raw material source, they are usually of a more consistent quality
than some natural aggregates. Specified material tests will provide the
necessary quality characteristics.'4
These findings supported similar conclusions reported earlier in the LC
Bridge Deck Survey of the Expanded Shale Clay and Slate Institute.24
Reports of the inspections of the durability of the HSLC used in World
War I and World War II concrete ships also attested to excellent long-term
performance.2'3'6'7'49
Freeze/thaw resistance
Most accelerated freeze/thaw testing programs conducted on structural LC
have incorporated HSLC on the high end of the compressive strength
range of the specimens tested. Investigations in North America25'50 and in
Europe39'51 researching the influence of entrained air volume, cement
content, aggregate moisture content, specimen drying times, and testing
environments have arrived at essentially the same conclusion: air entrained
LCs properly proportioned with high quality binders provide satisfactory
results when tested under usual laboratory freeze/thaw testing programs.
Core samples taken from the hulls of 70 year old LC ships as well as the
30 to 40 year old LC bridge decks have demonstrated concretes with high
internal integrity and low levels of microcracking. This proven record of
high resistance to weathering and corrosion is due to physical and chemical
mechanisms; they include superior resistance to microcracking developed
by significantly higher aggregate/matrix contact zone adhesion as well as
internal stress reduction due to the elastic matching of coarse aggregate
and matrix phases. High ultimate strain capacity is also provided by
concrete with a high strength/modulus ratio. In addition, because of elastic
compatibility, the stress/strength ratio at which the disruptive disintegra-
tion of concrete begins is higher for LC than for equal strength NWC. A
well dispersed void system provided by lightweight fine aggregates will
assist the entrained air pore system, and may also serve an absorption
function by reducing disruptive mechanisms in the matrix phase. Addi-
tionally, long-term pozzolanic action is provided by the silica rich ex-
panded aggregate combining with calcium hydroxide liberated during
cement hydration. This will reduce permeability and minimize leaching of
soluble compounds.
It is widely recognized that while ASTM C666, Standard Test Method for
Resistance of Concrete to Freezing and Thawing, provides a useful compa-
rative testing procedure, there remains inadequate correlation between
accelerated laboratory test results and the observed behavior of mature
concretes exposed to natural freezing and thawing. Inadequate laboratory/
field correlation observed when testing NWC is compounded when inter-
preting results from laboratory tests on structural LC prepared with high
aggregate moisture contents. A proposed modification to ASTM C666
recommends a 14 day air drying period prior to the first freezing cycle, to
improve correlation between laboratory test data and observed field
performance.2
Durability characteristics of any concrete, both NWC and LC, are
decisively influenced by the protective qualities of the paste fraction. It is
imperative that the concrete matrix provide high quality, low permeability
characteristics in order to protect steel reinforcing from corrosion, which is
clearly the dominant form of structural deterioration observed in current
construction. The protective quality of the matrix in concretes prop-
ortioned primarily for thermal resistance that incorporate high air contents
and low cement quantities will be significantly reduced. Very low density,
non-structural LC will not provide resistance to the intrusion of chlorides,
carbonation, etc comparable to the long-term satisfactory performance
demonstrated with high quality, structural grade LC.2
For a number of years field exposure testing programs have been
conducted by the Canadian Department of Minerals, Energy and Technol-
ogy (CANMET) on various types of concrete exposed to a cold marine
environment at the Treat Island Severe Weather Exposure Station main-
tained by the U.S. Army Corps of Engineers at Eastport, Maine.52'53
Concrete specimens placed on a mid-tide wharf experience alternating
conditions of sea water immersion followed by cold air exposure at low
tide. In typical winters the specimens experience about 100 cycles of
freezing and thawing. In 1978, a series of prisms were cast using commer-
cial NWAs with various cement types and including supplementary ce-
mentitious materials. W/c ratios of 0.40, 0.50, and 0.60 were used to
produce 28 day compressive strengths of 4350, 3770, and 3480 psi (30, 26
and 24MPa) respectively. In 1980, these mixes were essentially repeated
with the exception being that the Ii in. (40mm) gravel aggregate was
replaced with a 1 in. (25 mm) expanded shale LWA. Fine aggregates used
in both 1978 and 1980 were commercially available natural sands. Cement
contents for the semi-LC mixtures were 800, 600, and 400 pcy (480, 360,
and 240 kg/m 3 ), which produced compressive strengths of 5220, 4350 and
2755 psi (36, 30 and 19 MPa) respectively. All specimens continue to be
evaluated annually for ultrasonic pulse velocity, and resonant frequence as
well as visually ratings. Ultrasonic pulse velocities are measured centrally
along the long axis of the prisms. Negligible differences exist between the
structural LC (8 years) and NWC (10 years) after exposure to twice daily
sea water submission and approximately 1000 cycles of freezing and
thawing. 2
Permeability
Conventional strength concrete employs a matrix with a w/c ratio signi-
ficantly higher than that required for the chemical reaction associated with
hydration of portland cement. This excess of uncombined water is free to
either move due to an applied hydraulic gradient or, if allowed to
evaporate, provide conduits through which gas can either diffuse into the
concrete or flow in response to a pressure differential. Concentration
gradients of chemicals in liquid form can also diffuse into saturated
concrete or can be absorbed by an initially dry concrete.
The above defines a material which is porous, in that it contains both
pores and conduits that communicate with a free surface and, as a result, is
permeable to liquids and gases. The normal definition of the adjective
'porous' usually contains any, or all, of the following: possessing pores, or
containing vessels and conduits, or being permeable to liquids and gases.
In the case of concrete of usual strength, all of the above are applicable.
Both HSNWC and HSLC are usually made with water contents that only
slightly exceed that required to hydrate the cement. This means that while
the hydrated cement paste still contains pores and conduits, these channels
are not fully continuous nor do all of them communicate with the surface.
High strength concrete normally contains both silica fume and a superplas-
ticizer, which leads to a densification of the cement paste matrix after a
relatively short period of moist curing. Slag cement, fly ash, or both, are
frequently incorporated into the mix which further densifies the cement
paste matrix.
Aggregates inserted into the cementitious matrix are surrounded by the
paste and isolated from one another by this essentially impermeable
matrix. Microcracks, however, may form in the concrete as a result of the
volume changes associated with hydration of the cement paste matrix.
Microcracking can also result from stresses that arise in concrete as a result
of differing aggregate/matrix thermal expansion coefficients when the
concrete is heated or cooled. Aggregates and matrix fractions expand and
contract at different rates as the concrete gains and loses water, further
increasing prospects for microcracking. Microcracking will also result from
the lack of elastic compatibility that exists between the aggregate inclusion
and the cement paste matrix when the composite concrete is subjected to
an applied stress. It is these crack networks that, more than any other
factor, render concrete permeable to gases and liquids. The high stiffness
of the matrix fraction of high strength concrete results in a closer elastic
match and a lower propensity to form stress-induced microcracks.
For all concrete types, the ratio of aggregate to matrix stiffness starts out
at infinity as hydration begins, therefore any volume change due to the
hydration process, moisture change, or thermal changes may lead to
microcracking in the concrete prior to the superposition of any design loads
on the concrete. In most instances where liquid permeability is of concern,
or where presence of moisture is associated with deterioration due to
corrosion, autogenous healing also is operative. Laboratory testing is,
however, not usually conducted on mature concrete and this beneficial
effect is not observed. Field and early age laboratory tests on high strength
concrete do indicate permeability coefficients substantially lower by sever-
al orders of magnitude, however, when compared with commercial con-
crete. In fact, satisfactory techniques to satisfactorily measure the per-
meability of high strength concrete in most cases have not been developed.
The role the aggregate plays in the liquid and gas permeability of
concrete is minor for both LC and NWC. With both types of aggregate,
small cracks formed during the final crushing to size of the aggregate
particles may have some small effect. However, few of the cracks would be
oriented in the direction of flow, and the matrix essentially seals the cracks
at the surface of the aggregates. Scanning electron microscopy studies have
indicated that structural grade LWAs have a vesicular structure with
essentially no interconnection of pores so that flow of liquid or gas is
insignificant. Porous aggregate particles will not be permeable because of
the lack of continuous channels through the particles. Permeability inves-
tigations conducted on LC and NWCs exposed to the same testing criteria
have been reported by Khokrin,51 Nishi,54 Keeton,55 Bamforth, 56 and
Bremner.57 It is of interest that in every case, despite wide variations in
concrete strengths, testing media (water, gas and oil) and testing techni-
ques (specimen size, medium pressure and equipment) structural LC had
equal or lower permeability than its heavier counterparts. This result has
been attributed to the reduction of microcracks in elasticly compatible LC
and the enhanced bond and superior contact zone present in structural LC.
Corrosion resistance
Corrosion resistance of HSLC is at least comparable to the performance of
HSNWC. Investigations of mature bridges and marine structures report
that internal integrity and minimal microcracking have effectively limited
rapid intrusion of aggressive forces into the concrete.2'3'6 Internal integrity
effectively limits disruptive effects to diffusion mechanisms which are
orders of magnitude slower in their deteriorating actions.
Carbonate resistance
Penetration of the carbonation front into concrete is primarily determined
by the vapor diffusion characteristics of the mortar matrix. With high
quality matrixes typical for all HSCs, this issue will not be a concern. Field
measurements of carbonation depths in LC marine and bridge structures
have demonstrated that the rate of carbonation is extremely low, with
results shown in Figs. 10.10 and 10.11. Adequate protection against
Depth of carbonation (mm)
Never
exposed
Structural concrete
Non-structural concrete
Normal wt. concrete
Time in years
Fig. 10.11 Measured depth of carbonation (mm) of laboratory specimens of lightweight
concrete
carbonation depassivating reinforcing steel in the service life of high
quality structures is provided by covers recommended by ACI318.
Abrasion resistance
Investigations into the abrasion resistance of LC using tests conducted in
accordance with ASTM C779, Standard Test Method for Abrasion Resist-
ance of Horizontal Concrete Surfaces, reported that the
'wear trends were similar for both normal weight and lightweight
concretes and, because of the small differences in the wear amounts
between both types of concretes, it appears that lightweight concrete
could be used satisfactorily in abrasive situations where normal weight
concrete might normally be specified.'17
Bare LC bridge decks exposed to more than 100 million vehicles crossing
over a more than 20 year period have demonstrated wear patterns
comparable to NWC.
Unit weight
The fresh unit weight of HSLC is a function of mix proportioning, air
contents, water demand, and the specific gravity and moisture content of
the LWA. Decrease in density of exposed concrete is due to moisture loss
which, in turn, is a function of ambient conditions and surface area/volume
ratio of the member. Design professionals should specify a maximum fresh
unit weight for LC, as limits of acceptability should be controlled at time of
placement.
Dead loads used for design should be based upon equilibrium density
which, because of the very low permeability of HSLC, may be assumed to
be reached after 180 days for moderate sized members. Extensive tests
conducted during North American durability studies demonstrated that
despite wide initial variations of aggregate moisture content, equilibrium
unit weight for commercial strength LC was found to be 3.1 pcf (50 kg/m 3 )
above the oven dry unit weight. European recommendations for in-service
density are similar.12 For HSLC however, reduction in density from fresh
to equilibrium will be significantly less due to the low permeability and high
cementitious content which, because of internal curing periods, will serve
to hydrate a larger fraction of the mix water. Because of the high
cementitious content of HSLC, equilibrium density of non-submerged
concrete will be close to the fresh density. Hoff reported that HSLC
mixtures containing supplementary cementitious materials exposed to
hydrostatic pressures equal to 200 ft. (61 m) of water demonstrated an
increase of density of less than 4 pcf (64 kg/m 3 ). 17
Ductility
The ductility of concrete structural frames should be analyzed as a
composite system - that is, as reinforced concrete. Ahmad's studies
indicate that the ACI rectangular stress block is adequate for strength
predictions of HSLC beams and that the recommendation of 0.003 as the
maximum usable concrete strain is an acceptable lower bound for HSLC
members with strengths not exceeding 11,000 psi (76.5 MPa) and p/ph
values less than 0.54.58 Moreno found that while LC exhibited a distinctive
descending portion of the stress-strain curve, it was possible to obtain a flat
descending curve with reinforced LC members that were provided with a
sufficient amount of confining reinforcement slightly greater than that with
NWC.59 This report also included studies that showed that it was economi-
cally feasible to obtain desired ductility when increasing the amounts of
steel confinement.
Rabbat et at., came to similar conclusions when analyzing the seismic
behavior of LC and NWC columns.60 This report focused on how properly
detailed reinforced concrete columns could provide ductility and maintain
strength when subjected to inelastic deformations from moment reversals.
These investigations concluded that properly detailed columns made with
LC performed as well under moment reversals as NW columns.
Pumping
LWA may absorb part of the mixing water when exposed to high pumping
pressures. To avoid loss of line workability, it is essential to raise the level
of absorption of the LWA prior to pumping. Presoaking is best accom-
plished at the aggregate production plant where a uniform moisture
content is achieved by applying water by spray bars directly to the
aggregate moving on belts. This moisture content can be maintained and
supplemented at the concrete plant by stockpile hose and sprinkler systems
for at least one, but preferably, three days.
Presoaking will significantly reduce LWA rate of absorption, minimizing
water transfer from the matrix fraction which can cause slump loss during
pumping. Higher moisture contents developed during presoaking will
result in an increased specific gravity which, in turn, develops higher fresh
concrete density. High water content due to presoaking will eventually
diffuse out of the concrete, developing a longer period of internal curing
and a larger fresh to equilibrium density differential than that usually
associated with LC using aggregates of a lower moisture content. Aggre-
gate suppliers should be consulted for mix design recommendations
necessary for consistent pumpability. Mix designs and the physical prop-
erties measured on samples of HSLC pumped 830 ft. (250 m) to the 60th
floor of the NationsBank project in Charlotte, NC (Fig. 10.12) are shown
in Table 10.1.
Fig. 10.12
Table 10.1 Mix design and physical properties for concretes pumped 830 ft.
(268 m) on Nationsbank Building, Charlotte, NC, 1991
Mix# 1 2* 3
Mix proportions:
Cement type III (lbs.) 550 650 750
Flyash(lbs.) 140 140 140
Solite3/4to#4(lbs.) 900 900 900
Sand (lbs.) 1370 1287 1203
Water (gals.) 35.5 36.5 37.2
WRA (oz.) 27.6 31.6 35.6
Superplasticizer 55.2 81.4 80.1
Fresh concrete properties:
Initial slump (inches) 2i 2 21
Slump after Superplasticizer 5i 74 61
% air 2.5 2.5 2.3
Unit weight (pcf.) 117.8 118.0 118.0
Compressive strength (psi):
4 days 4290 5110 5710
7 days 4870 5790 6440
28 days (avg.) 6270 6810 7450
Splitting tensile strength (psi): 520 540 565
* Mix selected and used on project
Buildings
The 450 ft. (140 m) multi-purpose Federal Post Office and Office Building
constructed in 1967 with five post office floors and 27 office tower floors
was the first major New York City building application of post-tensioned
floor slabs. Concrete tensioning strengths of 3500 psi (24MPa) were
routinely achieved at 3 days for the 30 X 30 ft ( 9 x 9 m) floor slabs with a
design target strength of 6000 psi (41 MPa) at 28 days. Approximately
30,000 cubic yd (23,00Om3) of structural LC were incorporated into the
floors, and the cast-in-place architectural envelope serves a structural as
Fig. 10.14
HSNWC HSNWC
High strength
SU\Cin Regular strength
floor slab SLAC in floor slab
Fig. 10.15 Alternate construction schemes for transfer of HSNWC column loads through
floor slabs
from the column and intermingled with the regular strength floor slab
concrete. The technique used in the North Pier project avoids delicate
timing considerations that are necessary to avoid cold joints.
Bridges
Of the more than 400 LC bridge decks constructed throughout North
America, most have been produced with concretes at higher than usual
commercial level. The Sebastian Inlet Bridge, which utilizes extra long
HSLC drip-in spans during its construction in 1965, is included in one
LWA supplier's listing of almost 100 completed bridges. Transportation
engineers generally specify higher concrete strengths on bridge decks,
primarily to insure high quality mortar fractions (high strength combined
with high air content) that will minimize maintenance costs. One state
authority has completed more than 20 bridges utilizing HSLC using a
target strength of 5200 psi (36MPa), 6-9% air content, and an air dry
density of 115 pcf (1850 kg/m 3 ). Recent studies have identified tens of
thousands of bridges in the United States which are functionally obsolete
with low load capacity, unsound concrete, or insufficient number of traffic
lanes. To remedy limited lane capacity, Washington, DC engineers have
replaced a 4-lane bridge with 5 new lanes providing a 50% increase in
one-way, rush hour traffic, without replacing the existing structure, piers,
or foundations.
Marine structures
Because many offshore concrete structures will be constructed in shipyards
located in lower latitudes and then floated and towed to the project site,
there is a special need to reduce weight and improve the structural
efficiency of the cast-in-place structure. Because shallow water conditions
mandate lower draft structures the submerged density ratio of
HSNWC 2.50-1.00 _
HSLC 2.00-1.00 ~ L 5 °
which is greater than the air density ratio
2.50
loo = L25
becomes increasingly important.
These requirements have already been satisfied by several projects
placed in the Arctic, e.g., HSLC was used in 1981 in the TARSUIT
CAISSON Retained Island project constructed in Vancouver, Canada and
transported to the Beaufort Sea on the Alaskan North Slope.61 This
project was followed in 1984 with the HSLC used in the construction of the
Concrete Island Drilling System built in Japan and also towed to the
Table 10.2 Physical properties, strength and chloride ion permeability of
structural lightweight concrete microsilica1
Lightweight concrete mix K O F C N+ N -(D
(With (Without
Microsilica) Microsilica)
Fresh Wet 116.5 120.5 117.9 116.5 118.2 116.8
concrete density (1870) (1930) (1870) (1870) (1890) (1870)
properties psi (kg/m3)
Slump 7.5 8.5 8.5 8.8 8 6
in (mm) (190) (215) (215) (225) (205) (150)
Air
content % 5.5 5.2 5.8 5.2 5.8 5.2
2 6360 4850 4180 5920 5800 2890
(43.9) (33.4) (28.8) (40.8) (40.0) (19.9)
Compressive 7 8270 5740 5310 7470 6990 3960
strength (57.0) (39.6) (36.6) (51.5) (48.2) (27.3)
at age (days)
28 9648 6840 6050 8600. 7460 5050
(66.5) (47.2) (41.7) (59.3) (51.0) (34.8)
90 9855 6550 6225 8990 7525 5270
(68.0) (45.2) (42.6) (62.0) (51.9) (36.3)
W/B ratio (2) .36 .33 .35 .31 .32 .43
Chare passed (coulombs)(3) 260 450 450 220 370 4800
(1) Mix N- identical to N+ with exception of no microsilica/superplasticizer.
(2) Includes water from slurry.
(3) Rapid determination of the chloride permeability of concrete. AASHTO T277. Microsilica Elkem,
Pittsburgh, PA.
Beaufort Sea.62 This project called for compressive strengths of 6500 psi
(45MPa) with density of 115pcf (1840 kg/m 3 ). In addition to reducing
draft during construction and towing, use of HSLC in offshore gravity
based structures can be justified by the improved floating stability as well as
the opportunity to carry more topside loads.
11.1 Introduction
In Japan high strength concrete was first achieved as early as the 1930s. For
example, Yoshida reported in 1930 that high strength concrete with 28-day
compressive strength of 102MPa was obtained. 1 This result was obtained
by a combination of compression and vibration processes without chemical
or mineral admixtures. This isolated development was not followed by
systematic development in the production and use of high strength
concrete till the mid 1960s.
In 1968, high-strength reinforced mortar piles were developed by a
process of autoclave curing with the use of silica powder in cement. The
compressive strength of concrete used was ll,300psi (78MPa). 2 In 1970,
high strength concrete with a compressive strength of 12,700 psi (88 MPa)
was developed and used for prestressed concrete piles.3 This high strength
concrete was produced by using superplasticizer and autoclave curing. In
the 1970s high strength concrete was utilized in the construction of railway
bridges,4 and since then its use in the construction industry has continued
to increase. The growth of the use of high strength concrete is primarily
due to the development of naphthalene sulfonate condensed superplasti-
cizer, which was developed in Japan in 1964.5
The definition of high strength concrete as defined by the Japan Society
of Civil Engineering (JSCE) is different from that of Architectural Institute
of Japan (AIJ). In 1980 JSCE published Proposed Recommendations for
the Design and Construction of High Strength Concrete,6 which defined
high strength concrete as a concrete with a design compressive strength
8500 to 11,500psi (59 to 79MPa). On the other hand, AIJ defines high
strength concrete as a concrete with strength of 3900 to 5100 psi (27 to
35 MPa) for normal weight concrete, and with strength of 3500 to 3900 psi
(24 to 27 MPa) for a lightweight concrete.7
For civil engineering structures in Japan, high strengtrh strength con-
crete is used mainly for bridges, high-rise buildings, and piles.4 Studies in
the use of high strength concrete in super high-rise reinforced concrete
buildings are being promoted under the leadership of the Japanese
National Project by Ministry of Construction (MOC). Studies related to
the development and production of high strength concrete of strengths up
to 17,100psi (118MPa) are under way and techniques for making these
concretes practical for high-rise construction are being investigated.8
In South East Asia, studies into the utilization of high strength concrete
have just started. However, in some of the countries, especially in
Singapore, practical applications of high strength concrete have already
commenced in high-rise buildings.
Size of aggregate
Bauxite Fe powder
Slag sand
Silica sand
11.3 Applications
The applications of high strength concrete in Japan are summarized in
Table 11.1. The details of these examples including the methods of
strength development employed are described below.
Bridges
High strength concrete has been used for producing prestressed concrete
(PC) girders for the purpose of reducing the dead load and for achieving
longer span.8 Japanese Industrial Standard (JIS) requires the compressive
strength of concrete used for PC girders to be above 7100 psi (49 MPa) as
shown in Table 11.1. Most of the high strength concrete used for railway
bridges is usually a superplasticized concrete with a low water cement
ratio. Examples include the Dai-ni-Ayaragigawa Bridge, Ootanabe
Bridge, Iwahana Bridge, Kazuki Bridge and Akkagawa Bridge. The
strengths of concretes used in the highway bridges are in the range of
Table 11.1 Application of high strength concrete in Japan
Concrete constructions and Design
concrete products strength (MPa) Techniques for high strength Notes
Railway bridges 59-79 Superplasticizer (+high early cement) Lightened bridge weight
prestressed concrete
Highway bridges 59-69 Superplasticizer (+high early cement)
Prestressed concrete beams Superplasticizer (+ high early cement)
for slab bridges >49 JISA5313
for beam bridges >49 JISA5316
for light load slab bridges >49 JISA5319
Diaphram walls -49 Superplasticizer + low heat cement Massive concrete
Oil drilling rigs 58-65 Superplasticizer + silica fume Light weight concrete
Abrasion resistance concrete 39-79 Superplasticizer + silica fume or Repair of dam (floor)
anhydrous gypsum based additive
High rise RC buildings 35-47 Superplasticizer or new type National project
Superplasticizer (+ultra-fine particles) 60 -112 MPa
Reinforced spun concrete piles >39 (Superplasticizer) JISA5310 centrifugal force
Pretensioned spun concrete piles >49 Superplasticizer JISA5335 centrifugal force
Posttentioned spun concrete piles >49 Superplasticizer JISA5336 centrifugal force
Pretensioned spun high strength >79 Superplasticizer + autoclave curing JISA5337 centrifugal force
concrete piles or anhydrous gypsum based additive
Steel concrete composite piles >79 Superplasticizer + autoclave curing centrifugal force
or anhydrous gypsum based additive
+ expansive additives
Prestressed spun concrete poles >49 (or 79) Superplasticizer (+autoclave curing) JISA5309 centrifugal force
Centrifugal reinforced 49-69 Superplasticizer + anhydrous gypsum Centrifugal force
concrete pipes based additive or low slump concrete Jacked pipe
Railway sleepers 39-49 Superplasticizer (+high early cement)
Concrete segments 49 Superplasticizer
Machine beds 79-108 Superplasticizer + silica fume (DSP) Damping capability
Fig. 11.3 Iwahana Railway Bridge (Sanyo shinkan-Sen)
8400 psi to 9800 psi (59 to 69MPa). Examples include the Nitto, Kami-
noshima, Jodoji and Seto Highway Bridges. In addition to the above
examples, concretes with design strengths in the range of 5600 to 6400 psi
(39 to 44 MPa) has been utilized in many bridges in the form of prestressed
or cast-in-place concrete construction.
The mix proportions of the concrete used for these bridges are listed in
Table 11.2. In the Dai-ni-Ayaragigawa Railway Bridge, high strength
concrete was employed in order to reduce the dead load. The weight of the
main girder with 5600 psi (39 MPa) normal strength concrete would have
been 170 tons. This was reduced to 150 tons by the use of 8500 psi (59 MPa)
strength concrete. The strength of the concrete with job-site curing
resulted in 9000 to 10,400 psi (62 to 72MPa), with an average of 9400 psi
(65 MPa). The production of the PC girders was done in a yard near the
construction site, and the concrete used was produced in ready-mix
concrete plant about 2 km from the yard.24
In the construction of Iwahana Bridge (Fig. 11.3), precast concrete
members were prifnarily used but for the joints cast-in-place concrete was
used. Concrete mix information such as cement type, design strength, w/c
ratio, dosage of superplasticizer etc. is given in Table 11.2. The w/c ratio
for cast-in-place concrete was about 0.29. The average strength attained
was 10,700 psi (74 MPa).25
In the construction of the Akkagawa Bridge, concrete with a design
strength of 11,400 psi (79 MPa) was employed, and it was produced with
autoclave curing. The average concrete strength and the standard devia-
tion of the strength after curing, were 14,100 psi (97 MPa) and 580 psi
(4 MPa) respectively.26
The merit of using high strength concrete for piers was examined.27 It
was reported that by increasing the design strength of concrete from 3900
to 8600 psi (27 to 59 MPa), the cross section of a pier for monorail could be
reduced by 36%. It was also estimated that by increasing the design
strength from 3900 psi (27MPa) to 5700 and 11,400 psi (39 and 79MPa),
the cross section of a highway bridge pier could be reduced by 13% and
56% respectively.27
Table 11.2 Mix proportions of high strength concrete for railway bridges
Design Average of
Name of Type of strength w/c s/a Dosage of strength Con-
bridges structure (MPa) Type of cement (%) (%) SP(CX%) (MPa) struction
Da-ni- 1 Girder bridge 59 Normal cement 30 40 0.75 65 1973
Ayaragigawa span: 49 m
Iwahana Span: 45 m 79 High early cement 23 38,5 1.5 83 1974
Akkagawa Span: 45. 9 m 79 Normal cement 30 39.5 1.5 93* 1976
(Slump: 12 + 2.5 cm, G max : 20 mm, * autoclave curing)
Table 11.3 Mix proportions of the main tower for Aomori Oohashi Road Bridge (Type of structure: PC cable-stay bridge, design strength:
59 MPa)
The main towers of Aomori Bay Bridge were constructed in 1989 and
high strength concrete with a design strength of 8600 psi (59 MPa) was
used.28 Figure 11.4 shows the bridge and Table 11.3 lists the mix propor-
tions of the concrete. In this prestressed concrete cable-stayed highway
bridge, high strength concrete was employed to achieve a slender structure
with a reduced dead load and attractive aesthetic appearance. The
concrete incorporated a new type of superplasticizer. Due to the presence
of carboxylate polymer, this new superplasticizer reduced the slump loss
which is caused by conventional sulfonated naphthalene formaldehyde
condensate or sulfonated melamine condensate superplasticizer. Among
these new types of superplasticizer, the one with air-entraining agent is
currently being used.
Diaphragm walls
Usually, for diaphragm walls concrete of 3000 to 3500 psi (21 to 24 MPa) is
used. However, it is desirable to use higher strength concretes for deep
underground large scale walls, which are subjected to higher water and
earth pressures. This is very important for Japan, since there is a strong
need to develop underground space because of the limited land mass. The
Table 11.4 Mix proportions of high strength concrete using light weight aggregate
for oil drilling rigs
Unit weight (kg/m 3 ) Compressive
w/c Silica Fine Coarse Slump Air strength
(%) W C fume aggregate aggregate (mm) (%) (MPa, for 28d)
30.8 160 520 52 603 493 185 7.7 58
28.5 143 500 75 604 456 170 5.2 65
(with superplasticizer and air entraining agent)
Concrete piles
Concrete strengths for this application range between 5600 psi (39 MPa) to
11,500 psi (79MPa). The types of concrete pile used in Japan include
reinforced spun concrete piles, pretensioned spun concrete piles, post-
tensioned spun concrete piles and pretensioned spun high strength con-
crete piles (Table 11.1). The production and use of Pretensioned High-
strength Concrete (PHSC) piles is increasing very fast and presently
exceeds 50% of the total market of piles. This is because: (1) A
pretensioned high strength concrete pile can withstand higher vertical and
horizontal loads compared with other kinds of pile which reduces the
number of piles required, resulting in savings, (2) It has higher impact
resistance, (3) The production period is shorter. Figure 11.6 shows a
prestressed high strength concrete pile being driven. The typical mix
proportions for high strength concrete used for concrete piles is given in
Table 11.9. The high strength concrete piles are generally made using the
autoclave curing method, and sometimes incorporate silica fine powders.6
An alternate method for producing high strength concrete piles in Japan,
is by the use of ordinary steam curing with an anhydrous gypsum based
Table 11.9 Mix proportions of high strength concrete for concrete piles
Unit weight (kg/m3 )
Design Target Type of ^max Slump W/C s/a Dosage of SP
Products strength (MPa) strength (MPa) cement (mm) (cm) (%) (%) C W (CX%)
PHC piles* 78 88 Normal 20 y**# 32 40 450 142 1.25
PC piles** 49 57 Normal 20 4 36 42 420 152 0.85
(* : Steam curing + autoclave curing, * * : Steam curing, * * * : By using of pump - 12 cm)
No addition
10% addition
Compressive strength (MPa)
days months
Age
Fig. 11.7 Effect of addition of anhydrous gypsum based additive on compressive strength of
concrete
Alite CH
CS
TSH (Ett.)
CSH2
Time (hr)
Fig. 11.8 Hydration and unhydration products of cement with anhydrous gypsum based
additive
hand, when the composite piles are produced with ordinary steam curing, a
combination of anhydrous gypsum based additive and expansive additive is
used.37
In addition to the piles, high strength concrete is being used for the
production of concrete poles. The mix proportions and the manufacturing
method are basically similar to those used for the production of the piles.
Table 11.10 Mix proportions of high strength concrete for steel concrete
composite piles
Unit weight (kg/m3)
G max Slump W/Binder s/a Expansive Dosage of SP
(mm) (cm) (%) (%) W C additive (Binder X %)
20 13 30.7 42 132 404 26 1.5
(Autoclave curing, main component of expansive additive - CaO)
10,000psi (69MPa) strengths have been produced with centrifugal com-
paction and have been used in a number of applications. For these
concretes, sometimes only superplasticizer, sometimes a combination of
superplasticizer and anhydrous gypsum based additive is used. For the
placement of concretes with very low water cement ratio, vibratory
centrifugal compaction method is employed.
Prestressed concrete sleepers are used for railway tracks. In addition to
the prestressed concrete mentioned in Table 11.1, JIS specifications for
various kinds of prestressed concrete products have been established. The
specifications require that the design strength of the concrete, and the
strength at the time of prestressing must exceed 7100 psi (49 MPa) and
5700 psi (39MPa) respectively. Because of this requirement, a combina-
tion of high-early-strength cement and superplasticizer is usually employed
in the manufacturing process.
Concrete shield segments are being used as lining in shield tunneling.
High-strength concrete is being used for these segments to make them
lighter hence reducing the transportation and erection costs and to make
them stronger to better withstand the high earth pressures. Figure 11.9
shows concrete segments ready for use in tunnel construction by the shield
tunneling method.
One of the unique applications of high strength concrete is the use of
15,600 psi (108 MPa) concrete with DSP material for the beds of machine
tools. Because of the high degree of rigidity and vibration damping
properties of higher strength concretes, it is very advantageous for this
application.39 This application is an example of high performance concrete
use in which attributes other than strength govern its choice. Figure 11.10
(by Nippei Toyama Corporation) shows a machine tool, the bed for which
is made by using concrete with a strength of 15,600 psi (108 MPa).
A
AASHTO T277 (chloride permeability) 370
Abrasion resistance, LC 362
Abrasion/erosion resistance 153
Japanese applications 384 386
ACI 213-77 353
ACI 318-89
compressive strength 42 221 314 317
368
concrete contribution 225 229
core strengths 321
deflection 224
development length 315
hoop reinforcement 296
modulus of elasticity 50 353
shear strength 314
slender column design 232
tensile strength 353
ACI Code 1983 225
ACI Standard 211.1 13
ACI-ASCE 351 290
ACI-ASCE 352 290 293 296 298
Admixtures 350
effect on shrinkage 67
Age, effect on fatigue 128
Aggregates 11
coarse 12 350
This page has been reformatted by Knovel to provide easier navigation. 399
400
Index terms Links
Aggregates (Continued)
quality 29
coarse/fine ratios 16
effect on shrinkage 66
effect of unit weight 288
elastic modulus 346
fine 13 351
Air-entrainment 348 350 363
frost resistance 148
Alfe’s shrinkage model 90
Ambient conditions, effect on shrinkage 66
American Concrete Institute Committee 363 (1987) 31 40
mix design procedures 2
State of the Art Report (1984) 46 48 69
see also ACI
American Society for Testing and Materials. See ASTM
Anchorage
steel 227
limits on length 315
Annacis Bridge 336
Architectural Institute of Japan (AIJ) 375 388
ASTM C31 (test specimens) 19
ASTM C39 (compressive strength) 20 23 321
ASTM C78 (flexural loading test) 46
ASTM C150 (Portland cement) 3
ASTM C172 (sampling) 366
ASTM C173 (air content) 363 366
ASTM C469 (static modulus) 50 353
ASTM C470 (molds) 19
ASTM C496 (splitting tension test) 46
ASTM C512 (creep test) 356
B
Barenblatt model 176
Beam-column joints/connections 232
deformability 290
Beams
deflection 222
ductility 249
flexure 219
prestressed 229
shear 224 227
Biaxial loading, effect of 129
Blaine finenesses 3
Blast furnace slag 7 349
limitations on use 7
Bond properties
interfacial 130
reinforcing-steel 131 227
Bridge structures 334
Japan 379 381 383
lightweight concrete 369
rehabilitation of 370
Brittleness 189 196 198
D
Darcy's process 71
Deflection ductility 263 272
F
f-N curve. See S-N relationships
Fatigue damage 115 118
Fatigue life 116 119 128
Fatigue loading 115
failure 120
Fatigue strength 118 121 124 219
Federal Post Office and Office Building, New York City 367
Federation Internationale de la Precontrainte (FTP), Manual
of Lightweight Aggregate Concrete 347
Fibers, use of 196 197
Pick's process 72
Fictitious crack model (FCM) 176
Field control. See Construction procedures
FIP/CEB State-of-the-Art Report 1
Fire resistance 150
Flexural fatigue 124
Flexural strength 48 228
Flexural tension 59
G
Global load-deformation response 180
Glomar Beaufort Sea 1 337
Goodman diagram, modified 120 121
Griffith (ideal-brittle) fracture 165
I
Ice abrasion 155
In-place strengths 321
Interfacial fracture 195
Interferometry, use in stress-strain studies 58 180
Interfirst Plaza, Dallas 329
J
Japan, applications in 375 379
Japan Society of Civil Engineering (JSCE) 375
Joigny Bridge 107 336
L
Laboratory control 365
Laboratory testing programs 351
Lake Point Tower, Chicago 323
Lateral steel, effect of 215
LC (lightweight concrete) 341
Le Chatelier's contraction 70
Leaching corrosion 149
LEFM (Linear Elastic Fracture Mechanics) 162
Lightweight aggregate, pore system 343
Lignosulfonate-based superplasticizers 8
Loading history
effect on fracture 194
effect of 281
M
Marine structures. See Offshore structures
Materials engineering, use of fracture mechanics 196
Materials selection 2
Melamine sulfonate superplasticizers 8
Melment (superplasticizer) 8
Member size, effect on shrinkage 66
Microcracking 140 166 360
Microstructure 80 196
Miglin-Beitler Building, Chicago 332
Mix design/proportions 13
Mixed-mode fracture 194
Modulus of elasticity 50 316 346
high strength lightweight concrete 353
Modulus of rupture. See Flexural strength
Moisture condition, effect on fatigue 128
Molds, specimen 19 41 321
N
900 N. Michigan Avenue, Chicago 323
Naphthalene sulfonate superplasticizers 9
North Pier Tower, Chicago 368
Notch sensitivity 169
Nuclear containment structure, Civaux 102
O
Offshore structures 337
oil drilling rigs 383
use of high strength lightweight concrete 369
Oxygen availability, effect on corrosion resistance 147
P
Palmgren-Miner (P-M) hypothesis 117
Parking decks, rehabilitation 370
Particulate composite, elastic compatability 345
Pavement life 153 155
Permeability 139 149 359
high strength lightweight concrete 359
PHSC (Pretensioned High Strength Concrete) 388
Piles, Japan 388 390
Plastic settlement cracks 320
Plastic shrinkage cracks 319
Poisson's ratio 59
high strength lightweight concrete 354
Polymer impregnated concrete (PIC) 376
Portland cement. See Cement, Portland
Precast structures, HSLC 366
Prestressed concrete beams/girders 229 334
Prestressed structures, high strength lightweight concrete 366
Proportioning rules 363
Puddling 318
Pullout failure (reinforcing bar) 132
Pumping, concrete 364
R
R-curve models (fracture resistance) 181 190
Rehabilitation procedures 370
Reinforcement
column ties 217
column transverse 285
confinement 281
volumetric ratio 281
high-strength 283
spiral 217 286
steel-concrete bond 131
web 225 227
Research programs 351
Rest periods, effect of 127
Retarders 11
Reversed loading, effect of 126
Revibration 320
Richmond-Adelaide Centre, Toronto 323
RILEM draft recommendation 185 187
S
S-N relationships (stress/fatigue) 118
Salt scaling 148
Sedimentation 70
Seismic region construction 309
Self-desiccation shrinkage 70 71 75 80
Serviceability monitoring 107
Setting process 70
T
Temperature
during setting 8 72
effect on crack resistance 194
See also Thermal properties
Tensile strength
direct 45
high strength lightweight concrete 352
indirect 46
Tension, cyclic 124
Tension test, uniaxial 179
Test ages 18
Test specimens see Specimens
Testing machines 22
V
Vertical load transmission 317
W
Water Tower Place, Chicago 326
Water/binder (w/b) ratio. See Water/cementitious (w/c) ratio
Water/cementious (w/c) ratio 13 32
effect on shrinkage 66
Web reinforcement 225 227
Web-shear cracking 228
225 West Wacker Drive, Chicago 331
Wnuk model 176
Wohler's curve. See S-N relationships
Work-of-fracture 174