Fracture Mechanics of Concrete and Concrete Structures -

Recent Advances in Fracture Mechanics of Concrete - B. H. Oh, et al.(eds)

ⓒ 2010 Korea Concrete Institute, Seoul, ISBN 978-89-5708-180-8

Fracture mechanics of early-age concrete

V.T.N. Dao, P.H. Morris & P.F. Dux
School of Civil Engineering, The University of Queensland, Australia

ABSTRACT: In this paper, the fracture mechanics of concrete structures is first reviewed. This clearly dem-
onstrates that, despite the successful application of fracture mechanics to study the fracture behavior of ma-
ture concrete, its application to the cracking of very early-age concrete (i.e. within several hours after mixing)
is still in its infancy, with very limited literature available. This is believed principally to be the result of diffi-
culties that arise in the experimental determination of the properties of extremely fragile concrete at very
early ages. In a recent research at the University of Queensland, a test apparatus and experimental procedures
have been developed that enable the complete tensile stress-displacement behavior of concrete specimens at
very early ages to be captured reliably. Based on the data obtained, the paper shows that current models for
the stress-separation relationship of mature concrete may not apply for concrete at very early ages. Revised
mathematical models for the stress-separation relationship of early-age concrete are proposed.

1 REVIEW OF FRACTURE MECHANICS OF
CONCRETE STRUCTURES
The risk of failure due to the growth of cracks can be
treated using the science known as fracture mechan-
ics, which arose initially from the work of Griffith
(1920) on the fracture of brittle materials such as
glass. Its most significant applications, however,
have been in controlling the brittle failure and fa-
tigue failure of metallic structures such as pressure
vessels, airplanes, and ships. Considerable develop-
ment has taken place in the last several decades to
account for the ductility typical of metals.
Portland cement concrete is a relatively brittle
material and as a result, its mechanical behavior is
critically influenced by crack propagation. Many at-
tempts have been made to apply fracture mechanics
concepts to cement-based composites, such as mor-
tar and mature-age concrete. The first application of
fracture mechanics to concrete appears to have been
made by (Neville 1959a, b, c), while Kaplan (1961)
appears to have published the first experimental
study of the application of fracture mechanics to
concrete. Excellent reviews of the application of
fracture mechanics to mature cement and concrete
are available in the literature, and it is thus not cov-
ered in detail here.
There are fundamental differences between the
fracture behavior of concrete, a quasi-brittle mate-
rial, and that of brittle and ductile-brittle materials.
While the nonlinear zone is practically absent in
brittle materials, most of the nonlinear zone in front
of a crack tip in a ductile-brittle material (such as a
metal) involves hardening plasticity or perfect yield-
ing (Fig. 1). In ductile-brittle materials, the fracture
process zone (FPZ), which is the zone in which the
material undergoes softening damage, is quite small.
In quasi-brittle materials (such as mature concrete,
rock, and ceramics), however, plastic flow is almost
nonexistent and the FPZ fills almost entirely the
nonlinear zone (Fig. 1). It has been suggested that
the width and length of the FPZ in mature concrete
are of the order of three and twelve times the maxi-
mum aggregate size, respectively (Bazant & Oh
1983, Otsuka & Date 2000). The development of a
sizable FPZ necessitates the application of nonlinear
fracture mechanics (Bazant 2002), which differs sig-
nificantly from the fracture mechanics traditionally
applied to ductile-brittle materials. Linear elastic
fracture mechanics is applicable only to relatively
large-scale structures in which the effect of the non-
linear FPZ can be neglected.
Figure 1. Fracture process zone in metal and concrete (Bazant
2002).
Various models have been developed to describe
the FPZ in front of a crack in concrete, including no-
tably the fictitious crack model (FCM) proposed by
Hillerborg et al. (1976) and the crack band model
proposed by Bazant & Oh (1983) (Fig. 2). The for-
 D (h, T )∇the
J = −models
mer h process zone as a geometrically dis- (1) explicitly with
decreases accounts for the
increasing evolution
tensile of hydration
strength.
continuous crack with characteristics after cracking reaction
However, and the SF increase
content.inThis sorptionenergy
the fracture isotherm
of
thatThe
are described by a stress-crack
proportionality coefficient opening
D(h,T) isrelation-
called reads with its strength should not be confused
concrete
ship. The latter imagines the process zone
moisture permeability and it is a nonlinear function to exist with an increase in its ductility. The fracture energy
within a certainhumidity
of the relative finite bandwidth in whichT the
h and temperature mi-
(Bažant alone cannot distinguish⎡ ductility from brittleness, ⎤
crocracks are uniformly distributed and describes
& Najjar 1972). The moisture mass balance requires the nor does the reverse
we (h αbec αevaluated apply.
⎢ Additional parameters
⎥
s ) = G (α ctoα smake
) − + 1

behavior after cracking by a stress-strain

that the variation in time of the water mass per unitrelation- must , , ,
⎢ this
1

(g α
distinction.
∞ )h
⎥ It has
c − α clength,
1

ship.
volume Figure 2 shows(water
of concrete the stress distribution
content w) be equalandtosof-
the been suggested that the⎢⎣ characteristic
e
10
⎥
⎦
1 lch,
(4)
tening stress-separation
divergence of the moisture(orflux
equivalent
J strain) curve which is defined by ∞
for each model. The two models are essentially ⎡ g α c − α c )h
10(
⎤

equivalent (Karihaloo 1995, Bazant 2003), and are K (α c α s )⎢e 1

− ⎥
EG , 1

− ∂ = ∇ • Jreferred to collectively as cohesive mod- (2) lch = 2F ⎢ ⎥

(2)
1

sometimes
w
⎣ ⎦
∂
els or fracture process models. In this paper, the
t ft
FCM Theiswater
applied to study
content thebefracture
w can parameters
expressed as the sumof where the first term (gel isotherm) represents the
early-age concrete. can be usedbound
physically as an inverse
(adsorbed) measure
waterofand
the the
brittleness
second
of the evaporable water we (capillary water, water of the concrete. The smaller the
term (capillary isotherm) represents the capillary characteristic
vapor, and adsorbed water) and the non-evaporable length,
water. Thisthe more brittle isthevalid
expression material, andlow
only for vicecontent
versa
(chemically bound) water wn (Mills 1966, (Hordijk
of SF. The coefficient G1 represents the amount &
et al. 1989, Karihaloo 1995, De Schutter of
Pantazopoulo & Mills 1995). It is reasonable to Taerwe
water per1997). The characteristic
unit volume held in the gel poreslchatis100%
length also
assume that the evaporable water is a function of the length
relative of the FPZ.
humidity, and itFor canmost lch ap-
concretes,(Norling
be expressed
relative humidity, h, degree of hydration, αc, and pears
Mjornell to 1997)
rangeasbetween 200 mm and 400 mm
degree of silica fume reaction, αs, i.e. we=we(h,αc,αs) (Hordijk et al. 1989), but can sometimes have values
= age-dependent sorption/desorption isotherm greater than 1000 mm s(Hillerborg 1985). Other typi-
(Norling Mjonell 1997). Under this assumption and G (αvalues = k cthe (5)
by substituting Equation 1 into Equation 2 one
cal c s vg c + k vg α s s length of mature-age
α ) for
, c
α characteristic
concrete and other materials are listed in Table 1.
1

obtains
where1.kTypical
Table
c
vg and k vg are material parameters. From the
s
characteristic lengths of mature concrete and
∂w ∂h maximum
materialsamount of water
1995). per unit volume that can
− e + ∇ • ( D ∇h) = ∂we ∂w
α&c + e α&s + w&n (3)
other
fillMaterial
(Karihaloo
all pores (both capillary pores and gel pores), (mm) one
∂h ∂t h ∂α ∂α can calculate K as one obtains
lch

c s Glass 1 10 -6

Figure 2. Stress distribution and softening curve: (a, b) cohe- Cement paste densified by silica fume 1
where
sive crack e/∂h is
∂wmodel forthe slope of the
ductile-brittle sorption/desorption
materials; (c, d) cohesive Hardened cement paste ⎡
⎢ ⎜ g α −5α - ⎟15
⎛ ∞
c c h⎥
⎞ ⎤
10

isotherm
crack (also
model for called materials;
quasi-brittle moisture (e, f)capacity). The
crack band model Mortar w − α c s + α s s − G ⎢ − e
0.188 0.22
⎝
100 -⎠200
1 ⎥
1

governing equation
materials(Equation 3) must be completed ⎦ (6)
0 1
for quasi-brittle (Bazant 2002). ⎢ ⎥
K High strength concrete (50-100 MPa)
(α α ) =
⎣ 150 - 300
by appropriate boundary and initial conditions. 1 c s concrete (agg. size
Normal
,
smaller
⎜ g α − α ⎟h
⎛ ∞ ⎞ than 200 - 500
The area underbetween
relation the entire
the softening
amount ofstress- crack
evaporable 20 mm) e ⎝ c c⎠ −
10
1

opening curve (Fig. 2d) represents the total energy

1

water and relative humidity is called ‘‘adsorption Dam concrete (agg. size of 38 mm) 700
dissipated
isotherm” by if fracture
measuredperwithunit area of crack,
increasing and is
relativity The material parameters kcvg and ksvg and g1 can
defined
humidityasand the fracture energy,
‘‘desorption GF. Thatinis,the opposite
isotherm” be Itcalibrated by fitting
is noticeable from experimental data relevant
Table 1 that normal concreteto
case. Neglecting their difference (Xi et al. 1994), in isfreeless(evaporable) waterstrength
brittle than high contentconcrete
in concrete
but moreat
the following,
cw ‘‘sorption isotherm” will be used with variousthan
brittle agesdam
(Di Luzio & Cusatis
concrete, which2009b).
comprises larger
GF = ∫ σ (wto)dw
reference both sorption and desorption conditions. (1) aggregates. That is, the brittleness decreases with in-
By the 0 way, if the hysteresis of the moisture creasing aggregateevolution
2.2 Temperature size but increases with increasing
isotherm would be taken into account, two different concrete strength. Also, a comparison of the data for
where
relation, wFevaporable
is the crackwater
opening at zerohumidity,
vs relative stress. This is
must Note that, cement
hardened at early paste,
age, since
pastethedensified
chemical by reactions
silica
obviously also the external
be used according energy
to the sign of therequired
variation to create
of the associated
fume, with strength
and high cement concrete
hydrationwithand that
SF for
reaction
nor-
and fully break
relativity a unit The
humidity. surfaceshape
area of crack. The GF
the sorption are exothermic, the temperature field
mal strength concrete indicates that the more com-is not uniform
of matureforconcrete,
isotherm determinedbybymany
HPC is influenced eitherparameters,
direct or for non-adiabatic
pact systems
the microstructure of even if thetheenvironmental
the mix, more brittle
indirect
especially tensile tests,
those that normally ranges between
influence extent and rate of the 40 temperature
the material. is constant. Heat conduction can be
J/m
chemical
2
and 150 J/m (Elfgren
reactions2
and, in1989,
turn,Karihaloo
determine1995).pore described
Despite in theconcrete,
successfulatapplication
least for temperature
of fracture me- not
Itstructure
is influenced
and poreto size
varying degrees(water-to-cement
distribution by both micro- exceeding 100°C (Bažant & Kaplan
chanics to study the fracture behavior of mature con- 1996), by
structural
ratio, cement and environmental/testing
chemical composition, factors, includ-
SF content, Fourier’s
crete, its law, which reads
application to the early-age cracking of
ing
curing the time
aggregate size andtemperature,
and method, quality, water-cement
mix additives, ra- concrete is still in its infancy, with very limited lit-
tio,
etc.).age In of
the concrete,
literature and loading
various rate (Petersson
formulations can be erature
q = − λ ∇available.
T A summary of relevant literature (7)
1980a,
found tob, describe
Wittmanntheetsorption
al. 1987,isotherm
Karihaloo 1995).
of normal on the fracture behavior of early-age concrete is pre-
Generally, the higher the strength of
concrete (Xi et al. 1994). However, in the present the concrete, sented
where inq Table is the2 (Dao
heat etflux,
al. 2009).
T is In thetheabsolute
often-
the
paper higher
the the fracture energy.
semi-empirical This is in
expression stark con-
proposed by mentioned state-of-the-art report on the
temperature, and λ is the heat conductivity; in thisproperties of
trast
Norling to elastic-plastic
Mjornell (1997) metalsiswhose
adoptedfracture energyit
because concrete at early ages (RILEM committee 42-CEA

Proceedings of FraMCoS-7, May 23-28, 2010

 J = − D ( h , T ) ∇h
Table 2. Summary of literature on fracture behavior of early-age concrete (Dao et al. 2009).
Study Testing Conclusions
(Ostergaard et al. 2004) Wedge-splitting test to determine stress-crack Tensile strength, Young’s modulus, and fracture
opening relationships of two high-performance The proportionality
energy were found to increase with age.coefficient
However, D(h,T)
moistureinpermeability
concrete mixes with w/c of 0.31 and 0.48 at ages ductility expressed and it length
terms of characteristic is a nonlinea
between 8 h and 28 days. was foundoftothe relative
decrease withhumidity
age. h and temperature
(Zollinger et al. 1993) Three-point bending test of concrete aged be- Increases & in both
Najjar critical stressThe
1972). intensity factor mass
moisture and balanc
tween 12 h and 28 days. critical effective crack
that observed, length
the variation or process
in time zone size
of the water mas
with age were indicating that early-age
volume of concrete (water
concrete is more brittle than mature concrete. content w) be eq
(Kim et al. 2004) divergence of the moisture
Wedge-splitting test for concrete of ages from 1 Both critical stress intensity factor and fracture flux en-
J
day to 28 days. ergy increase with age, especially at early ages,
and converge∂wto a limit at 28 days. Similar trends
were noted− by = ∇ • J et al. 1993).
∂t (Zollinger
(De Schutter & Taerwe Three-point bending test of unnotched prisms of Results similar to those of most other researchers
1997) 150 mm x 150 mm x 600 mm; span 500 mm; were observed.
concrete aged between 1 day and 28 days The water content w can be expressed a
(Morris & Dux 2005) Using LEFM to estimate fracture energy of ce- of thewas
Fracture energy evaporable
estimated aboutwater we (capillary wa
73 N/m,
ment mortars aged up to 8 h from mixing, by which suggested that plastic cracking
vapor, and adsorbed water) of cement
and the non-e
measuring the total suction and assuming crack mortar involves a significant zone of plastic strain-
depth. ing or microcracking adjacent to the crack tip. wn (Mil
(chemically bound) water
Pantazopoulo & Mills 1995). It is reas
1981), nothing can be found concerning fracture en- that are crucial for assumethe that the
study of evaporable
concrete at water
these is a fu
ergy. Even at a relatively recent major international relative
very early ages aredegreedescribedhumidity,
in Dao h, al.
degree of hydration
conference concerning early-age concrete (Kovler et of silica fumeet reaction,
(2009). αs, i.e. we=w
al. 2004), little attention was paid to the experimen- = age-dependent sorption/desorption
tal study of the evolution of the softening behavior. 2 STRESS-SEPARATION (Norling Mjonell 1997). Under this assum
RELATIONSHIPS
The dearth of available literature on the fracture be- by substituting Equation 1 into Equati
havior of early-age concrete is principally the result Various models, obtains
including linear, bi-linear and non-
of difficulties in the experimental determination of linear functions (Hillerborg et al. 1976, Gopalarat-
∂w ∂h ∂w we
the tensile stress-separation curves of early-age con- nam & Shah 1985, − e
Reinhardt 1985, Foote
+ ∇ • ( D ∇h ) = e α&et+ ∂al. α& + w
crete (Dao et al. 2009). Consequently, a new test ap- 1986, Liaw et al. 1990, ∂h ∂Karihaloo
t h 1995, ∂Elices
α
c
c
et ∂al.
α s
s
paratus and experimental procedures have been de- 2002), have been proposed for the stress-separation
veloped at the University of Queensland (Dao 2007), relationships of wheremature∂wconcrete. isonthe Based
slope on these
of the sorption/
which have enabled the collection of reliable data on models for matureisothermconcretee/∂hand
(also the
calledearly-age
moisture ten- capac
the complete tensile stress-displacement behavior of sile strengths andgoverning
fracture equation
energies (Equation
in (Dao et3) al. must be
concrete specimens at ages of 1.5 h or more after 2009) (Fig. 4), the by following
appropriate models
boundaryfor the
and soften-
initial conditi
mixing. The variation of the fracture energy with the ing behavior of concrete Thecases,at verybetween
relation early ages
the have of e
amount
age and the tensile strength of early-age concrete are been developed. In all
water and best fits were obtained
relative humidity is called ‘‘
plotted in Figures 3-4, respectively. Other properties by ordinary least-squares
isotherm”methods.
if measured with increasing
80 humidity and ‘‘desorption isotherm” in th
Mix 1
case. Neglecting their difference (Xi et al.
60 Mix 2

the following, ‘‘sorption isotherm” will be

)
Mix 3

2
m
/J 40 reference to both sorption and desorption c
(F
G 20
By the way, if the hysteresis of the
isotherm would be taken into account, two
0 relation, evaporable water vs relative humi
0:00 1:30 3:00
4:30 6:00 7:30 be used according to the sign of the varia
Age (h) relativity humidity.
Figure 5. Idealized stress-separation relationships. The shape of the
Figure 3. Fracture energy versus age of concrete. isotherm for HPC is influenced by many p
especially those that influence extent and
80 2.1 Linear modelchemical reactions and, in turn, determ
structureforandlinear
pore size distribution (water-
Mix 1

Mix 2
The softening function stress-separation
behavior (Fig. 5a)ratio,is cement chemical composition, SF
60

2
) Mix 3

m
J/( 40
curing time and method, temperature, mix
G
F

⎛ w ⎞
etc.). In the literature various formulatio
σ = ft ⎜⎜1 − ⎟⎟ found to describe the sorption(3) isotherm
20

0
⎝ w c ⎠ concrete (Xi et al. 1994). However, in th
0 20 40 60 80 100
paper the
The fracture energy semi-empirical
defined in Equationexpression
1 thus pro
Tensile strength (kPa)
becomes Norling Mjornell (1997) is adopted b
Figure 4. Fracture energy versus tensile strength of concrete.
Proceedings of FraMCoS-7, May 23-28, 2010
J = − Dw (h, T )∇h
c
ft wc (1) explicitly are
functions accounts for the evolution of hydration
investigated:
GF = ∫ σ (w)dw = (4) reaction and SF content. This sorption isotherm
The0 proportionality 2
coefficient D(h,T) is called reads ⎛ w ⎞m
moisture permeability and it is a nonlinear function σ = ft ⎜⎜1 − ⎟⎟ (8)
of The best-fithumidity
the relative value forh and wc obtained
temperature forTthe data
(Bažant ⎝ wc ⎠ ⎡ ⎤
plotted in Figure 4 is 1.31 mm, about
& Najjar 1972). The moisture mass balance requires two orders of ⎢
we (h, α c , α s ) = Gn1 (α c , α s )⎢1 − 1 ⎥
+
magnitude higher than typical values
that the variation in time of the water mass per unitof between ⎡ ⎛ w⎞ ⎤ 10(g α
∞ ⎥
− α c )h ⎥
0.01
volumemmofand 0.02 mm
concrete forcontent
(water mature wconcrete
) be equal(Hiller-
to the σ = ft ⎢1 − ⎜⎜ ⎟⎟ ⎥ ⎢
⎣ en 1 c
where 0 < < 1 ⎦ (9)
(4)
borg et al. 1976).
divergence of the Itmoisture
is, moreover,
flux J significantly small- ⎢
⎣ ⎝ wc ⎠ ⎥
⎦ ⎡ 10(g α ∞ − α )h ⎤
er than the value of about 5 mm obtained from the K1 (α c , α s ) e⎢ 1 c c − 1⎥
experimental
∂ = ∇ • J study of early-age concrete by Dao(2)et Combining Equations 5 and 9 with Equation 1
⎢ ⎥
− w
al. ∂(Dao et al. 2009). Consequently, the linear model, ⎣ ⎦
while adopting the experimentally determined value
which is also inconsistent with the corresponding ex-
t

of 5 mmthe(Dao
where firstetnterm
al. 2009)
(gel for wc, gives
isotherm) best-fit val-
represents the
perimental
The waterstress-separation
content w can berelationships,
expressed as istheinap-
sum ues for m and of 6.647 and 0.151, respectively.
propriate for concrete at very
of the evaporable water we (capillaryearly ages. water, water physically
The latter bound (adsorbed)
is closeisotherm) water
to the corresponding and the
value second
of n for
vapor, and adsorbed water) and the non-evaporable term
mature (capillary
concrete, represents
which isranges the capillary
(chemically bound) water wn (Mills 1966, water. This
(Reinhardt expression
1985). valid between 0.2 and
only for low 0.4
content
2.2 Bi-linear model of SF. The coefficient G1 represents the amount of
Pantazopoulo & Mills 1995). It is reasonable to water per unit volume held in the gel pores at 100%
The
assume softening
that thefunction
evaporable for water
the bi-linear stress-
is a function of
separation behavior (Fig. 5b) is
relative humidity, h, degree of hydration, αc, and 3relative
SUMMARY
Mjornell
humidity,
1997) as ANDandCONCLUSIONS
it can be expressed (Norling
degree of silica fume reaction, αs, i.e. we=we(h,αc,αs)
σ= = age-dependent sorption/desorption
for 0 ≤ w ≤ w1 isotherm
(5) In this paper,c the fracture mechanics of concrete
w
f − ( f − σ1 ) s
(Norling Mjonellw1997).
t t

1
Under this assumption and structures
G (α
c α ,
s ) hasvgbeen
= k α
c c +briefly
k vg α
s sreviewed. This has high- (5)
by substituting Equation 1 into Equation 2 one lighted clearly the limitations of the application of
1

obtains
and fracture cmechanics to the early-age cracking of con-
crete, k vg anditsksvgsuccessful
wheredespite are material parameters.
application From
in the the
study
∂w ∂h σ ( w − w ) ∂w ∂w maximum
of the fracture amount of water
behavior per unit
of mature volume that can
concrete.
= we1 α
σ− =∂hσe 1 ∂−t +1∇ • ( Dh1∇h) for &
< e & &
cw+≤ wc α s + wn (6)
(3) fillMathematical
all pores (bothmodels capillary pores and gel pores), one
for the stress-separation re-
( wc − w1 ) ∂ α
c ∂α
s can calculate K as one obtains
lationship of early-age concrete have been derived
1
based on recent experimental data for concrete at
The ∂w
where fracture
e/∂h isenergy (Equation
the slope of the 1) is thus
sorption/desorption ages of 1.5 h or more after mixing. ⎡
⎢
It⎛⎜⎝ ghas
10 α ∞ −been
c α ⎞⎟h ⎤⎥dem-
c⎠ ⎥
isotherm (also called moisture capacity). The onstrated that w − notαall
0.188
c s + existing
0.22α s − G stress-separation
s ⎢ −e1
1
rela-
governing equation (Equation tionships for mature concrete are applicable to (6)
⎦ con-
w3)1) +must be completed
0 1
⎢ ⎥

GF appropriate
by = ( ft − σ 1 )boundary
1
w1 + 1 σ1(and wc −initial 1w1
σconditions. (7) K (α c α s ) =
crete ,
at very early ages, and
⎣
the revised models
⎜ g α − α ⎟h
1 ⎛ ∞ ⎞
2 2
The relation between the amount of evaporable proposed in this papere should
10
⎝ 1c be c ⎠ preferred.
− 1 The bi-
water and relative
Adopting humidity is determined
the experimentally called ‘‘adsorption
value of linear model appears to be the best of the three mod-
isotherm” if measured with increasing
5 mm (Dao et al. 2009) for wc, the best-fit values relativity
for els The
investigated. However, further
material parameters kcvg andwork ksvg isand
beingg1car- can
humidity and ‘‘desorption isotherm”
w1 are 0.808 mm, 0.558 mm, and 0.308 mm for val- in the opposite ried out by thebyauthors
be calibrated to collect more
fitting experimental datadata and re-
relevant to
case.ofNeglecting
ues σ1 of 0.1ft,their0.15fdifference (Xi et al. 1994), in
t, and 0.2ft respectively. This
fine
free the(evaporable)
revised stress-separation
water content relationships.
in concrete at
the following, ‘‘sorption
range of values for σ1 is consistentisotherm” willwith
be used
bothwith
the various ages (Di Luzio & Cusatis 2009b).
reference to both sorption and desorption
theoretical model (Fig. 5b) and the experimental conditions.
By the way, ifrelationships
stress-separation the hysteresis of the concrete
for early-age moisture REFERENCES
2.2 Temperature evolution
isotherm by
obtained would(Daobeettaken into account,
al. 2009). Obviously, two when
different
σ
relation, evaporable water vs relative
equals zero, the model becomes linear (Equations humidity, must31 Note that,
Bazant, at early
Z.P. 2002. age, fracture
Concrete since the chemical
models: testingreactions
and prac-
be used according tice. Engineering Fracture Mechanics/Fracture of Con-
and 4). For values toof theσ1 sign
greaterof thethanvariation
0.25ft, wofc be-
the associated with cement hydration and SF reaction
crete and Rock 69(2): 165-205.
relativity humidity. The
comes very small or even negative. shape of the sorption are exothermic,
Bazant, the temperature
Z.P. 2003. Reminiscences fielddecades
on four is notofuniform
struggle
isotherm
These for HPCindicate
results is influenced
that the by many parameters,
bi-linear model, forand
non-adiabatic systemsdamage
progress in softening even ifand
thesize
environmental
effects. Con-
especially those that influence
which has the virtue of simplicity, can modelextent and rate of the
the temperature
crete Journalis(Tokyo)
constant.
40(2): Heat
16-28. conduction can be
chemical reactions and, in turn,
stress-separation curve sufficiently accurately for determine pore described
Bazant, Z.P. in concrete,
& Oh, at Crack
B.H. 1983. least band
for temperature not
theory for fracture
structure purposes.
and pore size distribution (water-to-cement of concrete.100°C
exceeding Materials and Structures/Materiaux
(Bažant & Kaplan 1996), et Con-
by
practical structions 16(93): 155-177.
ratio, cement chemical composition, SF content, Fourier’s
Dao, V.N.T.law,
2007.which readscracking of concrete. Brisbane:
Early-age
curing time and method, temperature, mix additives, University of Queensland, Australia.
2.3
etc.).Non-linear model various formulations can be
In the literature q = −V.N.T.
Dao, λ ∇T et al. 2009. Tensile properties of early-age con-
(7)
found to describe
A number of non-linear the sorption
functions isotherm of normal
for the stress- crete. ACI Materials Journal 106(6): 1-10.
concrete (Xi
separation et al. 1994).
behavior (Fig. 5c) However,
have been in the present
suggested De Schutter, G. & Taerwe, L. 1997. Fracture energy of con-
where is ages.
the Materials
heat flux, T is the absolute
paper the semi-empirical crete atq early and Structures/Materiaux et
(Gopalaratnam & Shah 1985,expression
Reinhardtproposed
1985, Foote by temperature, and λ is the
Constructions 30(2): 67-71. heat conductivity; in this
et al. 1986). Here, the following two non-linearit
Norling Mjornell (1997) is adopted because

Proceedings of FraMCoS-7, May 23-28, 2010

Elfgren, L. (1989). Fracture Mechanics of Concrete Struc- Liaw, B.M. et al. 1990. − D (h, T )nonlinear
J =Improved ∇h model for concrete
tures: From Theory to Applications (Report of the Techni- fracture. Journal of Engineering Mechanics 116(2): 429-
cal Committee 90-FMA Fracture Mechanics to Concrete - 445.
Applications). London: Chapman and Hall. Morris, P.H. & Dux, P.F.The proportionality
2005. Suctions, fracturecoefficient
energy, and D(h,T)
Elices, M. et al. 2002. The cohesive zone model: Advantages, plastic cracking ofmoisture permeability
cement mortar andACI
and concrete. it isMate-
a nonlinea
limitations and challenges. Engineering Fracture Mechan- of the
rials Journal 102(6): relative humidity h and temperature
390-396.
ics 69(2): 137-163. Neville, A.M. 1959a.&Some aspects
Najjar of the
1972). Thestrength of concrete
moisture mass balanc
Foote, R.M.L. et al. 1986. Crack growth resistance curves in - Part 1. Civil Engineering
that the (London)
variation 54(639):
in time 1153-1156.
of the water mas
strain-softening materials. Journal of the Mechanics and Neville, A.M. 1959b. Some aspects of the strength of concrete
Physics of Solids 34(6): 593-607. - Part 2. Civil Engineering (London) 54(640): 1308-1310. w) be eq
volume of concrete (water content
Gopalaratnam, V.S. & Shah, S.P. 1985. Softening response of Neville, A.M. 1959c.divergence
Some aspectsofofthe themoisture
strength offlux J
concrete
plain concrete in direct tension. ACI Journal 82(3): 310- - Part 3. Civil Engineering (London) 54(641): 1435-1439.
323. Ostergaard, L. et al. 2004.
∂w =Early-age stress-crack opening rela-
Griffith, A. A. 1920. The phenomena of rupture and flow in tionships for high−performance
∇ • J concrete. Cement and Con-
solids. Philosophical Transactions of the Royal Society of crete Composites 26(5):∂t 563-572.
London Series A 221: 163-198. Otsuka, K. & Date, H. 2000. Fracture process zone in concrete
Hillerborg, A. 1985. Results of three comparative test series The waterFracture
tension specimen. Engineering contentMechanics
w can be65(2): expressed a
for determining the fracture energy GF of concrete. Materi- 111-131. of the evaporable water we (capillary wa
als and Structures/Materiaux et Constructions 18(107): Petersson, P.E. 1980a. Fracture
vapor, energy
andConcrete of concrete:
adsorbed water)Method
and 79- of non-e
the
407-413. determination. Cement and Research
(chemically bound) water wn (Mil 10(1):
Hillerborg, A. et al. 1976. Analysis of crack formation and 89.
crack growth in concrete by means of fracture mechanics Petersson, P.E. 1980b.Pantazopoulo
Fracture energy & ofMills 1995).
concrete: It is reas
Practical
and finite elements. Cement and Concrete Research 6(6): performance and experimental results. Cement and Con- is a fu
assume that the evaporable water
773-781. relative
crete Research 10(1): 91-101.humidity, h, degree of hydration
Hordijk, D.A. et al. 1989. Material properties. In L. Elfgren Reinhardt, H.W. 1985. degree ofsoftening
Crack zone in
silica fume plain concrete
reaction, αs, i.e. we=w
(ed), Fracture Mechanics of Concrete Structures: From under static loading.
= Cement and
age-dependent Concrete Research 15(1):
sorption/desorption
Theory to Applications (Report of the Technical Committee 42-52.
90-FMA Fracture Mechanics to Concrete - Applications), RILEM committee 42-CEA(Norling 1981.Mjonell 1997).
Properties of setUnder
concretethisat assum
67-127. London: Chapman and Hall. by substituting
early ages: State-of-the-art Equationand1 Struc-
report. Materials into Equati
Kaplan, M.F. 1961. Crack propagation and the fracture of con- tures/Materiaux etobtains
Constructions 14(84): 399-451.
crete. Journal of the American Concrete Institute 58(11): Wittmann, F.H. et al. 1987. Influence of age of loading, water-
591-610. cement ratio and rate∂wof loading on fracture energy ∂w
of con-
∂w
e ∂h e e α& + w
Karihaloo, B.L. 1995. Fracture Mechanics and Structural
Concrete. Harlow: Longman.
crete. Materials and−
20(116): 103-110. ∂h ∂t
Structures/Materiaux
+ ∇ • ( D ∇h ) =
h
et
∂α
& c +
Constructions
α
∂α s

Kim, J.-K. et al. 2004. Fracture characteristics of concrete at c

Zollinger, D.G. et al. 1993. Fracture toughness of concrete at
early ages. Cement and Concrete Research 34(3): 507-519.
Kovler, K. et al. 2004. Early age concrete-properties and per-
formance. Cement and Concrete Composites 26(5): 413-
415.
s
early ages. ACI Materials Journal 90(5): 463-471.
where ∂we/∂h is the slope of the sorption/
isotherm (also called moisture capac
governing equation (Equation 3) must be
by appropriate boundary and initial conditi
The relation between the amount of e
water and relative humidity is called ‘‘
isotherm” if measured with increasing
humidity and ‘‘desorption isotherm” in th
case. Neglecting their difference (Xi et al.
the following, ‘‘sorption isotherm” will be
reference to both sorption and desorption c
By the way, if the hysteresis of the
isotherm would be taken into account, two
relation, evaporable water vs relative humi
be used according to the sign of the varia
relativity humidity. The shape of the
isotherm for HPC is influenced by many p
especially those that influence extent and
chemical reactions and, in turn, determ
structure and pore size distribution (water-
ratio, cement chemical composition, SF
curing time and method, temperature, mix
etc.). In the literature various formulatio
found to describe the sorption isotherm
concrete (Xi et al. 1994). However, in th
paper the semi-empirical expression pro
Norling Mjornell (1997) is adopted b

Proceedings of FraMCoS-7, May 23-28, 2010