Concrete Creep Modeling: Application to Slabs on Ground

Ya Wei, M.ASCE 1; Siming Liang 2; and Jingsong Huang 3

Abstract: Moisture warping of concrete slab occurs as a long process due to the slow transportation nature of moisture in concrete materials.
Creep, as a time-dependent property of concrete, will inevitably affect the mechanical responses of slabs on ground during this process. This
study considered the effect of concrete creep in pavement slab for analyzing moisture warping, warping stress, and the total stress under traffic
Downloaded from ascelibrary.org by University of Central Florida on 11/27/20. Copyright ASCE. For personal use only; all rights reserved.

loads. A method is proposed to calculate the time-dependent warping deformation and stress in concrete slab on ground. This method uses the
sequential coupled hygromechanical finite-element analysis and incorporates a concrete creep model which is capable of considering the
relative humidity effect. The effect of environmental relative humidity; the design parameters, including concrete type, slab size, thickness,
and bonding condition with the base; and the age of concrete exposed to drying on the warping and stress of slabs were investigated
considering the creep effect. It was found that concrete creep has a significant effect on the warping and stress of slabs. It is important to
consider creep effect in designing and analyzing deformation and stress generated in concrete pavement slabs, particularly when a moisture
gradient exists in slabs. DOI: 10.1061/(ASCE)EM.1943-7889.0001663. © 2019 American Society of Civil Engineers.
Author keywords: Creep effect; Cement concrete pavement; Moisture warping; Relative humidity; Sequential coupled hydromechanical
analysis.

Introduction the base layer, and the adjacent slabs, and thus stress is generated in
the warped slab. When combined with traffic loads, accelerated
Creep, as a time-dependent property, is important to cement con- fatigue failure might occur in such a slab, manifesting as bottom-
crete because it affects the magnitude and distribution of deforma- up transverse cracking or top-down corner cracking.
tion and stress in structures, including concrete pavement slabs on Moisture warping is a process related to the formation of mois-
ground. Concrete slab on ground is a special type of structure with ture gradient along the depth of a concrete slab from external drying
most of its surfaces in contact with the environment. This subjects at the top surface and moisture absorption at the bottom surface,
concrete slabs to severe influence from the environmental temper- and thus is a very slow phenomenon. During this prolonged warp-
ature and relative humidity (RH). Moisture is one of the main ing process, concrete creep no doubt has a great impact on the
factors that influence the mechanical responses of cement concrete warping deformation and the stress generated. Therefore, concrete
pavement (Poblete et al. 1990; Suprenant 2002; Wei and Hansen creep needs to be considered during the mechanical response evalu-
2011; Lederle and Hiller 2012; Liang and Wei 2018). Moisture ation of concrete pavement. However, investigations of the creep
warping of portland cement concrete pavement slabs due to the effect on the mechanical responses of pavements are very limited.
change of relative humidity of concrete is a well-recognized issue Yeon et al. (2012) used the creep coefficient adopted from CEB-FIP
which has a significant influence on the long-term mechanical re- Model Code 1990 (CEB-FIP 1993) to analytically calculate the
sponses of concrete pavements. When the bottom part of a concrete average stress at the middepth of a slab. Lim et al. (2014) predicted
slab has a greater moisture content than the top part of the slab, a the average tensile stress generated in concrete pavement using the
negative moisture gradient will be induced. The bottom part of the creep properties measured from a ring test. Both studies found
concrete slab will experience more expansion (or less shrinkage) that the tensile stress can be relaxed by up to 50% compared with
than the top part, resulting in an upward warping. Conversely, when the elastic stress. Therefore, creep plays a very critical role in the
the bottom part of a concrete slab has a lower moisture content than mechanical responses of pavement. However, neither study inves-
the top part of the slab, a positive moisture gradient will occur tigated the time-dependent deformation and stress by considering
and the top part of the slab will experience more expansion (or less concrete creep as a function of relative humidity of concrete, nor
shrinkage) than the bottom part, resulting in a downward warping were the warping deformation and stress evaluated by considering
(Ceylan et al. 2016). Normally, the warping deformation of a con- various influencing factors. A more direct and in situ condition
crete slab is restrained to a certain degree by the self-weight of slab, simulation needs to be performed to evaluate the creep effect on
warping and stress in slabs on ground.
1
Associate Professor, Key Laboratory of Civil Engineering Safety This study investigated the creep effect on the mechanical re-
and Durability of China Education Ministry, Dept. of Civil Engineering, sponses of pavements, including the warping deformation, warping
Tsinghua Univ., Beijing 100084, China (corresponding author). Email: stress, and the total stress including traffic loads. A coupled sequen-
[email protected] tial hygromechanical finite-element analysis (FEA) is proposed to
2
Postdoctoral Fellow, Dept. of Civil Engineering, Tsinghua Univ., model slab on ground in a three-dimensional (3D) manner consid-
Beijing 100084, China. ering the creep effect by incorporating the microprestress solidifi-
3
Graduate Student, Dept. of Civil Engineering, Tsinghua Univ., Beijing
cation (MPS) theory–based creep model and considering the
100084, China.
Note. This manuscript was submitted on June 11, 2018; approved on relative humidity effect. Both the moisture transportation model
March 13, 2019; published online on July 30, 2019. Discussion period open and the creep model were calibrated using measured data. The creep
until December 30, 2019; separate discussions must be submitted for effect on slabs was investigated for different slab sizes, thicknesses,
individual papers. This paper is part of the Journal of Engineering conditions of bonding with the base, types of concrete, ages when
Mechanics, © ASCE, ISSN 0733-9399. exposed to drying, and environmental relative humidity. This study

© ASCE 04019078-1 J. Eng. Mech.

J. Eng. Mech., 2019, 145(10): 04019078

 normally 7 days

normally >7 days

Downloaded from ascelibrary.org by University of Central Florida on 11/27/20. Copyright ASCE. For personal use only; all rights reserved.

Fig. 1. Creep effect on the mechanical responses of slabs on ground at early age and later age.

is expected to provide a methodology for the mechanical response opening is transient over a day because it is mainly caused by the
evaluation of concrete pavement considering the creep effect. daily temperature variation in the slab. Long-term joint opening
will not remain constant due to the creep effect of concrete. On
the other hand, moisture warping is a long process compared with
Concrete Creep in Slab the temperature effect, because this deformation originates from
nonuniform relative humidity distribution which is caused by the
Creep Effect on Mechanical Responses of Concrete very slow moisture transportation process powered by diffusion
Slab on Ground and capillary absorption (Wei et al. 2013). For concrete pavement
slabs, the influential depths of top-surface drying and bottom-
Creep exists throughout the entire life of concrete structures as long surface absorption are relatively fixed due to the very slow trans-
as stress exists in concrete. For concrete slab on ground, the effect portation process; the reduction of relative humidity mostly occurs
of creep manifests in both the early-age hardening stage and the at the top portion of the slab. Moreover, concrete creep is signifi-
hardened service stage. The mechanical responses of concrete slabs cantly affected by the moisture conditions in the concrete (Cagnon
on ground that are affected by concrete creep at these two stages are et al. 2015). Any variation of relative humidity in the concrete will
summarized in Fig. 1. enhance the creep behavior (Bažant et al. 1997). Considering the
During the early-age stage, when creep is normally pronounced, long process and the relative humidity–related nature of moisture
creep reduces the stress generated from restrained deformation warping, this study applied a creep model to a concrete slab to
(Wei and Hansen 2012). It might reduce the compressive stress evaluate the warping deformation, warping stress, and the total
from restrained expansion due to the heat of cement hydration stress when traffic load is involved.
or reduce the tensile stress generated from restrained shrinkage
due to the temperature decrease and the self-desiccation of con-
crete. At the same time, creep promotes the earlier generation of Creep Model
the tensile stress which occurs at a higher temperature relative The microprestress solidification theory–based creep model (Bažant
to the first zero-stress temperature (Schindler and McCullough et al. 1997) is one of the most advanced models for creep prediction
2002). Further temperature decrease or shrinkage will cause greater of cement-based materials. Both temperature and relative humidity
tensile stress and eventually cracking. On the other hand, the crack- can be considered in this model. According to Bažant et al., the MPS
ing time might be delayed by the creep effect of reducing the mag- theory is composed of two parts: the solidification theory part as-
nitude of tensile stress, which consequently delays the saw-cutting sumes that the creep aging is the result of the volume fraction increase
time as well. Therefore, the window for saw-cutting time requires of the nonaging hydration products; the microprestress theory part is
reevaluation if the creep effect is considered. Concrete creep also introduced to account for the temperature and relative humidity
affects the built-in curl in a slab during the hardening process. effect. Microprestress is the normal stress acting across the slip
Built-in curl refers to the temperature gradient in the hardening slab planes represented by the hindered adsorbed layers in the nanostruc-
at the time of final set. A positive built-in curl temperature gradient ture of cement paste. Changes in temperature and relative humidity
often occurs in hot weather or early-morning paving, which often create an imbalance in microprestress, which can reduce the bonding
results in an effective negative temperature gradient and excessive between the microstructure layers and thus increase the creep rate.
upward curl at later age (Hansen et al. 2006). In contrast, a negative In MPS model, the viscoelastic strain εev is assumed to originate
built-in curl temperature gradient often is generated in cold-weather in the solid gel of calcium silicate hydrates, and can be written
paving, which results in excessive downward curl at later age. (Bažant et al. 2004)
However, these built-in curls can be weakened by the creep of
concrete, which is normally recognized as very pronounced at early γ̇ðtÞ
ages (Wei et al. 2013). Yeon et al. (2012) conducted numerical ε̇ev ðtÞ ¼ ð1Þ
vðtÞ
analysis to study the effect of built-in curl on the stress development
of concrete pavement slabs and found that the built-in curl has little
where γðtÞ ¼ ∫ t0 Φðt − τ Þσ̇ðτ Þdτ = viscoelastic strain in the
influence on the long-term stress condition of slabs. P
The effect of creep on the mechanical responses of slabs during cement gel, where Φðt − t 0 Þ ¼ q2 lnð1 þ ξ n Þ ¼ A0 þ Ni¼1 Ai f1 −
the service stage manifests as joint opening, warping deformation, exp½−ðt − t 0 =Þτ i g, where ξ ¼ ðt − t 0 Þ=λ0 , λ0 ¼ 1d, τ 1 ¼ 10−4 d,
warping stress, and the total stress if traffic load is involved (Fig. 1). τ i ¼ 10τ i−1 (i ¼ 2; 3; : : : ; 10), A0 ¼ 0.2794q2 , and Ai ¼
Joint opening mostly affects the load transfer efficiency between 0.2072fð3τ i Þ0.1 =½1 þ ð3τ i Þ0.1 gq2 (i ¼ 1; 3; : : : ; 10); and vðtÞ =
the slabs, which eventually affects the stress generated in slab. Joint volume fraction of hydration products, which causes the

© ASCE 04019078-2 J. Eng. Mech.

J. Eng. Mech., 2019, 145(10): 04019078

 Moisture Diffusion Analysis Quasi-static Structural Analysis

H Hs
[ D( H ) H] shrinkage MPS creep
t t
sh f H J t, t ' g q1, q2 , q3 , q4 , c, k1, t , t '

User Subroutine UMATHT

User Subroutine UMAT

Relative humidity
RH H Loading Condition: self-weight
0.95 Boundary Condition: - D( H ) H Henv result
0.85
0.35 n
0.75 warping (mm)
0.65 1.20 1.20
0.55 1.00
Downloaded from ascelibrary.org by University of Central Florida on 11/27/20. Copyright ASCE. For personal use only; all rights reserved.

0.45 0.80
0.35
0.35 0.60
RH 0.35 0.40 1.20
0.95 0.20
increase 0.95 0.00

1.20 1.00 0.60 0.40 0.20

0.00
0.80
0.35

Y
0.95
X 1.20
Z

RH distribution along Loading Condition: corner traffic load at 40 days

concrete slab

Fig. 2. Procedure of sequential coupled hygromechanical finite-element analysis of moisture warping and stress in concrete pavement considering the
creep effect.

viscoelastic strain of concrete, and is only partially recoverable, Creep Modeling: Application to Slabs
where vðtÞ−1 ¼ ðλ0 =tÞm þ αc , αc ¼ q3 =q2 , n ¼ 0.1, and m ¼ 0.5.
The viscous strain is completely nonrecoverable, which can be Numerical Procedure: Sequential Coupled
modeled by a dashpot and formulated as (Bažant et al. 2004) Hygromechanical Analysis
σðtÞ A sequential coupled hygromechanical finite-element analysis was
ε̇f ðtÞ ¼ ð2Þ
ηðSÞ employed to predict the moisture warping and stress in concrete
slab considering the creep effect. Sequential coupled hygrome-
where ηðSÞ = viscosity of the dashpot, and is expressed as a power chanical analysis was performed by first solving the pure moisture
function of microprestress S, ηðSÞ ¼ 1=ðcbSb−1 Þ, where b ¼ 2, diffusion problem and then reading the relative humidity solution
and c is a parameter that reflects the influence of the microprestress into a quasi-static structural model for stress and strain analysis
on the creep property. The microprestress forms upon the formation (Fig. 2).
of the microstructure and relaxes over time. However, changes of Due to the similarity of the heat transfer model and the moisture
relative humidity and temperature increase microprestress. The diffusion model, the module of sequential coupled thermal-stress
evolution of microprestress can be assumed to follow a Maxwell- analysis in ABAQUS software version 6.11 was used to conduct
type rheological model the sequential coupled hygromechanical analysis. This section
ṠðtÞ SðtÞ ṡðtÞ compares the heat transfer model and the moisture diffusion model
þ ¼ ð3Þ normally used in concrete. The differential functions and the boun-
CS ηðSÞ CS
dary conditions of these two models are quite similar to each other,
where ½ṡðtÞ=CS = time rate of Maxwell model strain so it was possible to employ the module of sequential coupled
  thermal-stress analysis in ABAQUS to conduct the sequential
 Ḣ  coupled hygromechanical analysis of early-age concrete.
ṡðtÞ ¼ k1 Ṫ lnðHÞ þ T  ð4Þ
H

The change of microprestress corresponding to the changing of

temperature and humidity can be expressed as
  Table 1. Mixture proportions of concretes
 Ḣ
Ṡ þ ψS c0 S2 ¼ k1 Ṫ lnðHÞ þ T  ð5Þ Mixture
H
Parameters 3O 4O
where ψS = adjusting factor of the relaxation rate; and c0 ¼ Water:cement ratio 0.3 0.4
bc=½q4 ðb − 1Þ, where q4 is a parameter that reflects the viscous Cement (kg=m3 ) 810 698
compliance term of concrete creep. Water (kg=m3 ) 243 279
The preceding microprestress solidification theory–based creep Coarse aggregate (kg=m3 ) 795 795
model was used to simulate the moisture warping and stress in a Sand (kg=m3 ) 530 530
Super plasticizer (kg=m3 ) 1.8 —
concrete slab by using the finite-element method.

© ASCE 04019078-3 J. Eng. Mech.

J. Eng. Mech., 2019, 145(10): 04019078

 Table 2. Calibrated parameters of moisture diffusion model and MPS where T = temperature of concrete; ρ = density of concrete; ch =
theory–based creep model for 3O and 4O concretes specific heat capacity of concrete; kT ðTÞ = thermal conductivity of
Mixture concrete; ρch ð∂T Q =∂tÞ = generated heat of hydration per unit
Model Parameter 3O 4O
volume; β = convective film coefficient; and ∂T=∂n = temperature
gradient in direction n of the concrete surface.
Moisture diffusion model D0 ½10−5 =ðm2 =dÞ 7.36 9.23 For moisture diffusion, the following governing equation was
α 0.042 0.048 used because it can capture the relative humidity development
β 1.5 1.1
of early-age concretes well (Kang et al. 2012; Oliveira et al. 2015;
Hc 0.905 0.980
λ (m2 =d) 0.020 0.014
Nguyen et al. 2018)
MPS theory–based q1 (10−6 =MPa) 6.7 8.2
creep model q2 (10−6 =MPa) 34.2 61.3 ∂H ∂Hs
q3 (10−6 =MPa) 49.2 39.2 ¼ ∇ · ½DðHÞ · ∇H − ð8Þ
q4 (10−6 =MPa) 4.5 13.7
∂t ∂t
Downloaded from ascelibrary.org by University of Central Florida on 11/27/20. Copyright ASCE. For personal use only; all rights reserved.

c (10−14 =MPa2 =day) 7.0 6.7

k1 (MPa=K) 2.5 3.4 The boundary condition is

∂H
For heat transfer in concrete, the governing equation is −DðHÞ ¼ λðH − H env Þ ð9Þ
∂n
∂T ∂T Q
ρch ¼ ∇ · ½kT ðTÞ · ∇T þ ρch ð6Þ
∂t ∂t where H = concrete relative humidity; Hs = relative humidity due
The corresponding boundary condition is to self-desiccation; DðHÞ = diffusion coefficient as a function of
concrete relative humidity; λ = surface diffusion coefficient;
∂T ∂H=∂n = relative humidity gradient in direction n of the concrete
−kT ðTÞ ¼ βðT − T env Þ ð7Þ
∂n drying surface; and Henv = environmental relative humidity.

600
Symbol: experimental
Line: predicted
500
3O
Beam deflection (µm)

400 sealed
top and bottom surfaces P+G
exposed to RH env =40%
300

200 loaded beam

G
100

0 unloaded beam
0 10 20 30 40 50
(a) Age (days)

800
Symbol: experimental
Line: predicted

600 4O
Beam deflection (µm)

sealed
top and bottom surfaces
P+G
exposed to RH env =40%
400

loaded beam

200 G

0 unloaded beam
0 10 20 30 40 50
(b) Age (days)

Fig. 3. Comparison of the measured and predicted beam deflection development in concrete beams under different drying conditions (sealed curing,
and exposed to drying at the age of 7 days after sealed curing): (a) 3O concrete; and (b) 4O concrete.

© ASCE 04019078-4 J. Eng. Mech.

J. Eng. Mech., 2019, 145(10): 04019078

 Establishing Relationship between Relative Humidity The user subroutine UMATHT in the heat analysis of ABAQUS
and Free Shrinkage was used to simulate the moisture diffusion process, because this
subroutine can define not only the diffusion coefficient as a func-
In the sequential coupled finite-element analysis, it was essential to
tion of relative humidity but also the decrease of relative humidity
first establish the relationship between the relative humidity and the
over time caused by the self-desiccation of concrete. To calibrate
free shrinkage of concrete. The change of relative humidity causes
and validate the moisture diffusion model, the relative humidity
shrinkage or expansion of concrete, consequently causing warping
was measured at the middle depth of specimens under the sealed
deformation in slabs.
curing condition and exposure to drying at ages of 7 and 28 days.
Two types of concrete were measured experimentally, which are
The moisture diffusion model was calibrated by matching the
denoted 3O and 4O. The mixture proportions of these two con-
predicted and measured relative humidity of the specimens under
cretes are listed in Table 1. The relative humidity and the shrinkage sealed curing and exposure to drying at the age of 7 days. The
deformation of the concrete under the sealed-curing condition were calibrated parameters of the moisture diffusion model are listed in
measured on a prism specimen by a shrinkage rig. The specimen Table 2. The calibrated parameters were then validated by predict-
Downloaded from ascelibrary.org by University of Central Florida on 11/27/20. Copyright ASCE. For personal use only; all rights reserved.

was 38 mm high, 100 mm wide, and 1,000 mm long. The shrinkage ing the relative humidity of the specimen exposed to drying at the
deformation was measured by LVDTs. At the same time, the inter- age of 28 days. The predicted result was in good agreement with the
nal relative humidity of concrete was measured at the middle depth measured result, with a deviation of less than 6% (Liang and
of the specimen using relative humidity sensors. The tests were ini- Wei 2019), indicating the validity of the diffusion model and its
tiated after the final set time. More details about the measurement calibrated parameters. The calibrated moisture diffusion model
of the internal relative humidity and the shrinkage deformation from the measured relative humidity results were used to predict the
were given by Wei et al. (2016a). Two specimens were measured relative humidity development within the concrete beam before
for each case and the average was reported. calibrating the creep model of concrete with the same material
A linear relationship between the relative humidity and the free proportion.
shrinkage of concrete was found from the measurement. By linear
regression analysis of the relation between the free shrinkage and
the relative humidity (εsh ¼ ka × H þ ε0 , where εsh is the free Calibration of Creep Model
shrinkage, ka is the shrinkage coefficient, H is the relative humid- The creep parameters in the MPS model were calibrated by match-
ity, and ε0 is the initial shrinkage), the shrinkage coefficient ka was ing the predicted creep-induced beam deflection with the measured
determined as 1.663 × 10−3 and 1.895 × 10−3 for 3O and 4O con- one. The creep-induced beam deflection was measured using a
cretes, respectively (Liang and Wei 2019). The obtained shrinkage four-point bending configuration on concrete beams under sealed
coefficient was used to calculate the shrinkage strain of the concrete and symmetric drying conditions. For the sealed case, the beams
slab caused by the changes of relative humidity in the sequential were sealed with three layers of self-adhesive aluminum foil on all
coupled hygromechanical finite-element analysis. The preceding faces. For the drying case, the top and bottom surfaces were ex-
relative humidity–shrinkage relationship was obtained from the posed to the environment with relative humidity of 40% starting
results of the sealed specimens, because the uniform and unique from the age of 7 days. All the creep tests were conducted in a
shrinkage and relative humidity can be measured in the sealed- room with a temperature of 23°C  2°C. Two types of concrete
cured condition. were used to make concrete beam, which were the same as those

traffic load applied at the age of 40 days non-critical loading position

on warped slab
critical loading position on
moisture warped slab

top surface exposed

to environment
spring

(a)
tensile
F F
tensile

l l

compressive compressive compressive

tensile
unbonded case bonded case

connection between slab element and spring element

(b)

Fig. 4. Finite-element model of cement concrete pavement.

© ASCE 04019078-5 J. Eng. Mech.

J. Eng. Mech., 2019, 145(10): 04019078

 Table 3. Summary of input parameters in finite-element analysis beam under sealed condition, unloaded beam under drying condi-
Index Input parameters tion, and loaded beam under drying condition). The predicted beam
deflections for validation are plotted in Fig. 3. The deflection of
Mixture proportion 3O and 4O
Slab size 4 × 4 m and 1 × 1 m
concrete beams under the sealed and the drying conditions was pre-
Slab thickness 15 and 18 cm for 4 × 4-m slab; dicted exactly by the calibrated MPS creep model, which suggests
24 and 30 cm for 1 × 1-m slab that the MPS theory–based creep model is capable of characterizing
Environmental RH 25%, 50%, and 75% the basic and drying creep of concrete subjected to bending in ad-
Age when exposed to drying 7 and 28 days dition to characterizing the creep under tension and compression.
Subgrade type Unbonded and bonded Moreover, the deflection of the sealed and the drying beams can be
predicted by using the same set of parameters as long as the mixture
proportions are the same. This suggests that the creep parameters
used for shrinkage and relative humidity measurements. The beams in Table 2 are the intrinsic creep parameters which are mixture-
Downloaded from ascelibrary.org by University of Central Florida on 11/27/20. Copyright ASCE. For personal use only; all rights reserved.

were 50 mm high, 50 mm wide, and 1,220 mm long. The creep- dependent only. Thus they can be used to predict the creep be-
induced deflection was measured on both unloaded (with self- havior of concrete members exposed to different environmental
weight only) and loaded (self-weight and external load) beams. All conditions.
tests started at the age of 7 days, and the measuring duration was
about 40 days.
Model of Slab on Ground in Finite-Element Analysis
The creep-induced deflection of concrete beam under loaded
Incorporating Creep Effect
and sealed conditions was used to calibrate the MPS creep model
to obtain the creep parameters listed in Table 2, and these creep To evaluate the creep effect on the mechanical responses of con-
parameters were then validated by predicting the creep-induced de- crete pavement in terms of warping deformation and the stress gen-
flection of beams under three other testing conditions (unloaded erated in slabs, two types of analysis were conducted in ABAQUS
Warping deformation (mm)

RHenv=25%, no creep
RHenv=25%, with creep

RHenv=50%, no creep
RHenv=50%, with creep

RHenv=75%, no creep
RHenv=75%, with creep

Slab width (m)

Slab length (m)

(a)
Warping deformation (mm)

RHenv=25%, no creep

RHenv=25%, with creep

RHenv=50%, no creep

RHenv=50%, with creep

RHenv=75%, no creep
RHenv=75%, with creep

Slab width (m)

Slab length (m)

(b)

Fig. 5. 3D warping deformation of slabs of 4O concrete under different drying conditions considering or not considering the creep effect:
(a) 1-m × 1-m × 15-cm slab; and (b) 4-m × 4-m × 24-cm slab.

© ASCE 04019078-6 J. Eng. Mech.

J. Eng. Mech., 2019, 145(10): 04019078

 for comparison: elastic analysis (with no creep), and viscoelastic The concrete pavement structure [Fig. 4(a)] was modeled as
analysis (with creep). ABAQUS version 6.11 was employed to nu- a slab supported by a Winkler foundation. The bonding between
merically analyze the creep effect on the moisture warping and the slab and the base was simulated as bonded and unbonded
stresses developments in concrete slab by incorporating the mois- conditions. When the slab was unbonded with the base, the spring
ture transport model and the MPS creep model. To predict the time- elements could only bear compressive load, whereas in the bonded
dependent warping of concrete slabs caused by moisture transport case, the spring elements could bear both compressive and tensile
as well as creep behavior, a constitutive relation considering the load [Fig. 4(b)]. The two bonding cases are actually ideal situa-
shrinkage and creep properties of concrete in the quasi-static struc- tions, and the actual concrete pavement structure is mostly be-
tural analysis should be provided. Because the temperature during tween the bonded and unbonded conditions. The Poisson’s ratio
all measurements was kept constant, the temperature effect on con- of the concrete was 0.2. The modulus of subgrade reaction was
crete creep was not considered. The three-dimensional stress incre- 68 MPa=m.
ment and the strain increment relation is Two slab sizes, 4 × 4 m and 1 × 1 m, were analyzed to inves-
 tigate the behavior of large and small slabs, respectively. The slab
Downloaded from ascelibrary.org by University of Central Florida on 11/27/20. Copyright ASCE. For personal use only; all rights reserved.

X
10
ð1 − λn ÞΔγin
Δσnþ1 ¼ E Q eff −1
Δεnþ1 − i thicknesses were 15 and 18 cm for the small slab (1 × 1 m), and the
i¼1
vnþ0.5 thicknesses were 24 and 30 cm for the large slab (4 × 4 m). Three
 environmental relative humidities (25%, 50%, and 75%) were
ψnþ0.5 Δtnþ1
− Δσn − Δεsh ð10Þ analyzed in the finite-element analysis to evaluate the effect of envi-
ηnþ0.5 nþ1
ronmental relative humidity effect on the mechanical responses of
where Eeff = effective elastic modulus, Eeff ¼ 1=ðq1 þ A0 þ slab. The input parameters are summarized in Table 3.
P10 The procedure of the sequential coupled FEA was as follows.
A þ A Þ, where A ¼ ½A0 þ i¼1 Ai ð1 − λni Þ=vnþ0.5 , and
ev f ev
The moisture diffusion analysis was conducted to obtain the RH
Af ¼ ½ψnþ0.5 Δtnþ1 =ηnþ0.5 ; Δεshnþ1 ¼ ka ðΔH nþ1 ;ΔH nþ1 ;ΔH nþ1 ; field of the concrete slab. During the moisture diffusion analysis,
0;0;0ÞT , where ka = shrinkage coefficient, which denotes the the top surface of the slab was assumed to be exposed to the envi-
shrinkage strain increment caused by unit RH increment; and ronment with different relative humidities and the temperature
2 3 was assumed to be constant. A DC3D20 element was used in the
1 −V −V 0 0 0
6 −V 7 moisture diffusion analysis. The effect of RH on the mechanical
6 1 −V 0 0 0 7 responses of concrete pavement slabs was evaluated by structural
6 7
6 −V −V 1 0 0 0 7 analysis based on the RH results obtained in the moisture diffusion
6 7
Q¼6 7 analysis. A C3D20 element was used to mesh the concrete slab
6 0 0 0 2ð1 þ VÞ 0 0 7
6 7 during the structural analysis. A single axle load of 100 kN with
6 7
4 0 0 0 0 2ð1 þ VÞ 0 5 a dual wheel was applied at the slab corner [Fig. 4(a)] at the age of
0 0 0 0 0 2ð1 þ VÞ 40 days, which was the most adverse location in the slabs when
combined with an upward moisture warping condition. The contact
where V = Poisson’s ratio of concrete. area of each wheel was 250 × 167 mm.

Table 4. Summary of calculated results of unbonded concrete slabs at age of 40 days

RH (%) Size (m × m) Thickness (cm) Analysis type δ (mm) Swx (MPa) Stx (MPa)
25 1×1 15 Elastic 1.44 14.10 14.31
Viscoelastic 1.25 (13%) 1.15 (92%) 1.53 (89%)
18 Elastic 1.08 15.33 15.41
Viscoelastic 0.93 (14%) 1.45 (91%) 1.75 (89%)
4×4 24 Elastic 9.86 16.88 17.98
Viscoelastic 7.01 (29%) 2.31 (86%) 3.42 (81%)
30 Elastic 6.61 16.32 17.03
Viscoelastic 4.87 (26%) 2.16 (87%) 2.99 (82%)
50 1×1 15 Elastic 0.92 8.66 8.95
Viscoelastic 0.81 (12%) 0.84 (90%) 1.23 (86%)
18 Elastic 0.69 9.48 9.65
Viscoelastic 0.60 (13%) 1.03 (89%) 1.34 (86%)
4×4 24 Elastic 6.07 10.79 11.79
Viscoelastic 4.01 (34%) 1.75 (84%) 2.80 (76%)
30 Elastic 4.06 10.39 11.03
Viscoelastic 2.82 (31%) 1.60 (85%) 2.36 (79%)
75 1×1 15 Elastic 0.38 2.84 3.19
Viscoelastic 0.35 (8%) 0.27 (90%) 0.70 (78%)
18 Elastic 0.29 3.32 3.58
Viscoelastic 0.26 (10%) 0.39 (88%) 0.74 (79%)
4×4 24 Elastic 2.39 4.42 5.22
Viscoelastic 1.28 (46%) 0.95 (79%) 1.94 (63%)
30 Elastic 1.61 4.16 4.73
Viscoelastic 0.96 (40%) 0.83 (80%) 1.55 (67%)
Note: δ = maximum warping deformation; Swx
= maximum warping stress located at the top surface of the pavement slab; and Stx
= maximum total stress
located at the top surface of the pavement slab. Values in brackets are the reduction of warping deformation or stress caused by creep behavior.

© ASCE 04019078-7 J. Eng. Mech.

J. Eng. Mech., 2019, 145(10): 04019078

 Results and Discussions The reduction of the warping stress and the total stress was more
significant than the reduction of the moisture warping from the
Effect of Creep on Moisture Warping and Stresses creep effect. When the creep effect was considered, the warping
stress and the total stress were reduced by about 60%–90%, and all
Concrete creep affects both moisture warping and stresses in slabs. the stresses were below 4.5 MPa. The fatigue limit obtained was
Both elastic analysis without considering the creep effect and vis- 2.0 MPa, which is 45% of the flexural strength based on the PCA
coelastic analysis considering the creep effect were conducted to fatigue model (Huang 2004). The design criterion for slabs is to
calculate moisture warping and stress in slabs. The reduction of ensure that no fatigue damage occurs by limiting the maximum ten-
moisture warping is clearly seen in the 3D plots in Fig. 5. Moisture sile stress to below the fatigue limit of 2.0 MPa for various combi-
warping from the creep effect was reduced by a maximum of 46% nations of slab size and thickness, axle load, and temperature and
compared with the elastic results for all slabs investigated in this moisture gradients. The fatigue limit is plotted in Figs. 6(e and f) for
study (Table 4). the total stress evaluation. When the creep effect is considered,
Warping stress is stress generated in slab by moisture warping fatigue damage is not expected in the small slab under different
Downloaded from ascelibrary.org by University of Central Florida on 11/27/20. Copyright ASCE. For personal use only; all rights reserved.

alone, whereas the total stress is the stress generated in a slab when environmental relative humidities, whereas fatigue damage might
traffic load and moisture warping both exist. The loading position occur in large slab when the slab is exposed to lower environmental
of the traffic load is the critical loading position [Fig. 4(a)]. The relative humidity (Fig. 6).
stress results are plotted in Fig. 6 and summarized in Table 4.
The results of the pure elastic analysis (Fig. 6) indicated that most
Effect of Environmental Relative Humidity on Moisture
of the warping stress and the total stress exceeded the flexural
Warping and Stresses
strength of 4.5 MPa of normal pavement concrete when the envi-
ronmental relative humidity was below 75%. However, these elastic Moisture warping of concrete slabs is caused by the moisture
results are not realistic because the slabs in the field do not fail gradient distributed differentially along the slab depth, which is
when the environmental relative humidity is below 75%. closely related to the environmental relative humidity. Lower

12 h=24cm,no creep 2.0 1 1m slab h=15cm,no creep

4 4m slab
Warping deformation (mm)

Warping deformation (mm)

h=24cm,with creep h=15cm,with creep
10
h=30cm,no creep 1.5 h=18cm, no creep
8 h=30cm,with creep h=18cm,with creep
6 1.0

4
0.5
2

0 0.0
15 25 35 45 55 65 75 85 15 25 35 45 55 65 75 85
(a) Environmental RH (%) (b) Environmental RH (%)

20 h=24cm,no creep 20 1 1m slab h=15cm,no creep

4 4m slab
h=24cm,with creep h=15cm,with creep
Warping stress (MPa)

Warping stress (MPa)

16 h=30cm,no creep 16
h=18cm, no creep
h=30cm,with creep h=18cm,with creep
12 12

8 8

4 4

0 0
15 25 35 45 55 65 75 85 15 25 35 45 55 65 75 85
(c) Environmental RH (%) (d) Environmental RH (%)

20 4 4m slab h=24cm,no creep 20 1 1m slab h=15cm,no creep

h=24cm,with creep h=15cm,with creep
16 h=30cm,no creep 16
Total stress (MPa)

h=18cm, no creep
Total stress (MPa)

h=30cm,with creep
12 12 h=18cm,with creep

8 8

4 4

0 0
15 25 35 45 55 65 75 85 15 25 35 45 55 65 75 85
(e) Environmental RH (%) (f) Environmental RH (%)

Fig. 6. Analysis of 4 × 4-m and 1 × 1-m 4O concrete slabs with top surface exposed to environmental RH = 25%, 50%, and 75%: (a and b) calculated
maximum warping deformation; (c and d) calculated maximum warping stress; and (e and f) calculated maximum total stress.

© ASCE 04019078-8 J. Eng. Mech.

J. Eng. Mech., 2019, 145(10): 04019078

 Sx, max=1.15 Sx,max=0.84 Sx,max =0.27
Downloaded from ascelibrary.org by University of Central Florida on 11/27/20. Copyright ASCE. For personal use only; all rights reserved.

(a) (b) (c)

Sx,max=1.23 Sx,max=0.70
Sx,max=1.53

(d) (e) (f)

Fig. 7. Stress at top surface of 1-m × 1-m × 15-cm 4O concrete by slab considering creep effect (unit: MPa): (a) warping stress contour at
RHenv ¼ 25%; (b) warping stress contour at RHenv ¼ 50%; (c) warping stress contour at RHenv ¼ 75%; (d) total stress contour at
RHenv ¼ 25%; (e) total stress contour at RHenv ¼ 50%; and (f) total stress contour at RHenv ¼ 75%.

environmental relative humidity causes extensive drying at the top respectively. The maximum warping stress appeared at the middle
portion of a concrete slab on ground, and thus a more severe mois- of slab, and the total stress appeared at the slab edge 1=3 of the
ture gradient is formed within the slab depth because the moisture distance from the slab corner when the axle load was applied at
contents at the middle and the bottom portions vary little along the the critical loading position [Fig. 4(a)].
slab depth (Wei et al. 2016b).
The effect of environmental relative humidity (RHenv ) on the
magnitude of moisture warping was investigated for RHenv ¼ 25%, Effect of Slab Size and Thickness on Moisture Warping
50%, and 75%, covering the major range of environmental relative and Stresses
humidity in the field. The 3D warping deformations are shown Slab size influences the warping deformation of concrete slabs sig-
in Figs. 5(a and b) for small (1 m × 1 m × 15 cm) and large nificantly. Warping deformation decreased with decreasing slab
(4 m × 4 m × 24 cm) slabs, respectively. Moisture warping is very size in both the elastic analysis and viscoelastic analysis (Fig. 5).
sensitive to the environmental relative humidity. Lower environ- For the slab sizes investigated in this study, the maximum moisture
mental relative humidity results in greater moisture warping. A warping of small slabs did not exceed 1.5 mm, whereas it was
linear relationship with a negative slope was found between RHenv 10 mm for large slabs. The creep effect on reducing moisture
and the maximum warping deformation [Figs. 6(a and b)]. Exten- warping was more significant in large slabs than in small slabs.
sive warping was found for the RHenv ¼ 25% case, particularly for Moreover, slab thickness also affected the warping deformation
large slabs, with the maximum warping reaching 10 mm. of concrete slabs; thinner slabs had greater moisture warping than
Similar to warping deformation, both warping stress and the did thick slabs. Warping deformation decreased with increasing
total stress increase with the lowering of environmental relative slab thickness in both the elastic and the viscoelastic analyses. The
humidity, and an inverse liner relationship also was found between thickness effect on reducing moisture warping was pronounced in
the stresses and the environmental relative humidity [Figs. 6(c–f)]. large slabs.
Although the stress increased with the lowering of environmental The significant moisture warping in large slabs was expected
relative humidity, the stress distributions were not affected by the to cause greater warping stress and total stress when traffic load
environmental relative humidity. The stress distribution patterns of was involved; however, although the warping deformation of the
the warping stress and the total stress are shown in Figs. 7 and 8 for large slab was much greater than that of the small slab [Figs. 5
small (1 m × 1 m × 15 cm) and large (4 m × 4 m × 24 cm) slabs, and 6(a and b)], the warping stress and the total stress of the large

© ASCE 04019078-9 J. Eng. Mech.

J. Eng. Mech., 2019, 145(10): 04019078

 Sx,max=2.31 Sx,max=1.75 Sx,max=0.95
Downloaded from ascelibrary.org by University of Central Florida on 11/27/20. Copyright ASCE. For personal use only; all rights reserved.

(a) (b) (c)

Sx,max=3.42 Sx,max=2.80 Sx,max=1.94

(d) (e) (f)

Fig. 8. Stress at top surface of 4-m × 4-m × 24-cm 4O concrete slab considering creep effect (unit: MPa): (a) warping stress contour at RHenv ¼ 25%;
(b) warping stress contour at RHenv ¼ 50%; (c) warping stress contour at RHenv ¼ 75%; (d) total stress contour at RHenv ¼ 25%; (e) total stress
contour at RHenv ¼ 50%; and (f) total stress contour at RHenv ¼ 75%.

7
3O
6 4O
Maximum warping deformation (mm)

0
4m×4m×24cm, 4m×4m×24cm, 4m×4m×30cm, 4m×4m×30cm, 1m×1m×15cm, 1m×1m×15cm, 1m×1m×18cm, 1m×1m×18cm,
no creep with creep no creep with creep no creep with creep no creep with creep

Fig. 9. Maximum warping deformation of 3O and 4O concrete slabs with top surface exposed to environment with RH of 50%.

© ASCE 04019078-10 J. Eng. Mech.

J. Eng. Mech., 2019, 145(10): 04019078

 slab were only slightly greater than those of the small slab, regard- with water:concrete (w=c) ratios of 0.3 (denoted 3O) and 0.4
less of whether creep was considered or not. This suggests that (denoted 4O) were investigated. The 3O concrete caused relatively
small slabs are very effective to reduce warping deformation to im- lower moisture warping compared with the 4O concrete, regardless
prove the flatness of roads, and at the same time, small slabs can of slab size, thickness, and the creep effect (Fig. 9). This is because
reduce the stress generated in slabs to some extent. Therefore, it is the 3O concrete had a lower water transport coefficient, and the pro-
recommended to adopt small slab in dry areas where the moisture cess of moisture gain from or release to the environment was slower
gradient in slabs on ground might be very intensive. than that of the 4O concrete slab. This caused a less severe moisture
gradient and thus less warping deformation in the 3O concrete slab.
The difference of moisture warping between the 3O and 4O concrete
Effect of Mixtures on Moisture Warping and Stress
slabs was reduced when the creep effect was considered.
The effects of mixtures on the magnitude of warping, warping stress, Unsurprisingly, the 3O concrete slab had less warping stress
and the total stress are shown in Figs. 9–11. Two concrete mixtures (Fig. 10) and total stress (Fig. 11) in the elastic analysis due to
Downloaded from ascelibrary.org by University of Central Florida on 11/27/20. Copyright ASCE. For personal use only; all rights reserved.

12
3O

4O
10
Maximum warping stress (MPa)

0
4m×4m×24cm, 4m×4m×24cm, 4m×4m×30cm, 4m×4m×30cm, 1m×1m×15cm, 1m×1m×15cm, 1m×1m×18cm, 1m×1m×18cm,
no creep with creep no creep with creep no creep with creep no creep with creep

Fig. 10. Maximum warping stress (Sx ) of 3O and 4O concrete slabs with top surface exposed to environment with RH of 50%.

12

3O

10 4O
Maximum total stress (MPa)

0
4m×4m×24cm, 4m×4m×24cm, 4m×4m×30cm, 4m×4m×30cm, 1m×1m×15cm, 1m×1m×15cm, 1m×1m×18cm, 1m×1m×18cm,
no creep with creep no creep with creep no creep with creep no creep with creep

Fig. 11. Maximum total stress (Sx ) of 3O and 4O concrete slabs with top surface exposed to environment with RH of 50%.

© ASCE 04019078-11 J. Eng. Mech.

J. Eng. Mech., 2019, 145(10): 04019078

 the less severe moisture warping compared with the 4O concrete. warping and stresses. Because concrete dries fast at early age,
However, the viscoelastic analysis indicated slightly greater warp- the exposure time at the age of 7 and 28 days was evaluated. 4O
ing stress and total stress for the 3O concrete slab than for the 4O concrete was selected for evaluation because the moisture transport
concrete slab. This is because the 4O concrete had greater creep was faster in 4O concrete than that in 3O concrete. Early age drying
than the 3O concrete, which reduced more stress in the 4O concrete caused slightly higher moisture warping in slabs, regardless of slab
slab than in the 3O concrete slab. Therefore, 4O concrete might be size, environmental relative humidity, and the creep effect (Fig. 12),
better than 3O concrete in terms of larger creep to reduce the risk of
whereas the warping stress (Fig. 13) and the total stress (Fig. 14)
fatigue failure under repetitive traffic loads.
were almost unaffected by the drying age. This suggests that the
first 7 days of curing after placement are sufficient to prevent ex-
Effect of Drying Age on Moisture Warping and Stress cessive moisture warping and stress, because 7-day curing achieved
The age at which a concrete slab is exposed to drying is also im- the same effect as that of 28-day curing in terms of warping stress
portant in terms of its influence on the magnitude of moisture and total stress.
Downloaded from ascelibrary.org by University of Central Florida on 11/27/20. Copyright ASCE. For personal use only; all rights reserved.

12
exposed at 7 days

exposed at 28 days
10
Maximum warping deformation (mm)

0
RH=25%, no creep RH=25%, with creep RH=50%, no creep RH=50%, with creep RH=75%, no creep RH=75%, with creep

Fig. 12. Maximum warping deformation of 4 × 4 × 24-m slab with top surface exposed to different drying conditions at ages of 7 and 28 days for 4O
concrete.

20
exposed at 7 days

exposed at 28 days
16
Maximum warping stress (MPa)

12

0
RH=25%, no creep RH=25%, with creep RH=50%, no creep RH=50%, with creep RH=75%, no creep RH=75%, with creep

Fig. 13. Maximum warping stress (Sx ) of 4 × 4 × 24-m slab with top surface exposed to different drying conditions at ages of 7 and 28 days for 4O
concrete.

© ASCE 04019078-12 J. Eng. Mech.

J. Eng. Mech., 2019, 145(10): 04019078

 20
exposed at 7 days

exposed at 28 days
16
Maximum total stress (MPa)

12

8
Downloaded from ascelibrary.org by University of Central Florida on 11/27/20. Copyright ASCE. For personal use only; all rights reserved.

0
RH=25%, no creep RH=25%, with creep RH=50%, no creep RH=50%, with creep RH=75%, no creep RH=75%, with creep

Fig. 14. Maximum total stress (Sx ) of 4 × 4 × 24-m slab with top surface exposed to different drying conditions at ages of 7 and 28 days for 4O
concrete.

8
unbonded pavement
Maximum warping deformation (mm)

7 bonded pavement
6

0
h=24cm, no creep h=24cm, with creep h=30cm, no creep h=30cm, with creep

Fig. 15. Maximum warping deformation of 4 × 4-m unbonded and bonded pavement slabs with top surface exposed to environmental RH of 50% at
age of 7 days for 4O concrete.

Effect of Bonded and Unbonded Bases on Moisture in numerical analysis of the mechanical responses of pavement,
Warping and Stress such as warping deformation and the related stress in concrete slab.
The bonding type between the concrete slab and the base affects the
warping deformation and the stress. Figs. 15–17 show the effect of Conclusions
bonded and unbonded conditions on the warping deformation,
warping stress, and total stress of 4 × 4-m slabs, respectively. The A method of analysis was proposed to study the time-dependent
unbonded slab developed greater warping than the bonded slabs warping and stress development in concrete pavement slabs on
regardless of the slab thickness and the creep effect. This is because ground. This method incorporates the creep model into the sequen-
an unbonded slab can deform more freely than a bonded slab under tial coupled hygromechanical finite-element analysis considering
a moisture gradient. the relative humidity effect. The numerical results demonstrated
Opposite to the results of warping deformation, the warping the capability of this method to evaluate the mechanical responses
stress and the total stress in bonded slabs were greater than those of concrete slabs on ground when the creep effect is considered.
in unbonded slabs. This might be due to the restraint from the base It was found that concrete creep has a significant effect on
under the bonded condition. The difference of stress between the reducing the warping deformation and the stresses generated in
bonded and unbonded conditions was reduced when the creep ef- concrete pavement slabs. The moisture warping was reduced by
fect was considered. Moreover, the extremely high elastic warping a maximum of 46% and the warping stress and the total stress
stress and the total stress in slabs were reduced if creep effect was were reduced by a maximum of 60%–90% compared with the
considered. This suggests that the creep effect should be considered elastic results for the concrete investigated in this study. A linear

© ASCE 04019078-13 J. Eng. Mech.

J. Eng. Mech., 2019, 145(10): 04019078

 16
unbonded pavement
bonded pavement

Maximum warping stress (MPa)

12

4
Downloaded from ascelibrary.org by University of Central Florida on 11/27/20. Copyright ASCE. For personal use only; all rights reserved.

0
h=24cm, no creep h=24cm, with creep h=30cm, no creep h=30cm, with creep

Fig. 16. Maximum warping stress (Sx ) of 4 × 4-m unbonded and bonded pavement slabs with top surface exposed to environmental RH of 50% at
age of 7 days for 4O concrete.

16
unbonded pavement
bonded pavement
Maximum total stress (MPa)

12

0
h=24cm, no creep h=24cm, with creep h=30cm, no creep h=30cm, with creep

Fig. 17. Maximum total stress (Sx ) of 4 × 4-m unbonded and bonded pavement slabs with top surface exposed to environmental RH of 50% at age of
7 days for 4O concrete.

relationship with a negative slope was found between the environ- J. Eng. Mech. 130 (6): 691–699. https://doi.org/10.1061/(ASCE)0733
mental relative humidity and the warping deformation, the warping -9399(2004)130:6(691).
stress, and the total stress in slabs on ground. The creep effect on Bažant, Z. P., A. B. Hauggaard, S. Baweja, and F. J. Ulm. 1997. “Micro-
reducing moisture warping was more significant in large slabs than prestress-solidification theory for concrete creep. I: Aging and drying
effects.” J. Eng. Mech. 123 (11): 1188–1194. https://doi.org/10.1061
in small slabs. Curing for 7 days is sufficient to prevent excessive
/(ASCE)0733-9399(1997)123:11(1188).
warping and stresses, compared with 28 days of curing.
Cagnon, H., T. Vidal, A. Sellier, X. Bourbon, and G. Camps. 2015. “Drying
The results of this study prove the importance of considering the creep in cyclic humidity conditions.” Cem. Concr. Res. 76 (Oct): 91–97.
concrete creep effect when analyzing moisture warping and stresses https://doi.org/10.1016/j.cemconres.2015.05.015.
generated in concrete pavement slabs. The findings of this study are CEB-FIP (Comité Euro-International du Béton-Fédération International de
expected to provide a numerical methodology to the current design la Précontrainte). 1993. CEB-FIP Model Code 1990. London: Thomas
and mechanical response evaluation of concrete pavement in the Telford.
field, particularly for concrete pavement slab design in dry climate Ceylan, H., R. F. Steffes, K. Gopalakrishnan, S. Kim, S. Yang, and
areas. K. Zhuang. 2016. Development and evaluation of a portable device
for measuring curling and warping in concrete pavements. Trans
Project Reports. Ames, IA: Iowa State Univ.
Acknowledgments Hansen, W., Y. Wei, D. L. Smiley, Y. Peng, and E. A. Jensen. 2006.
“Effects of paving conditions on built-in curling and pavement perfor-
The authors thank the support from National Natural Science mance.” Int. J. Pavement Eng. 7 (4): 291–296. https://doi.org/10.1080
Foundation of China under Grant Nos. 51578316 and 51778331. /10298430600798952.
Huang, Y. H. 2004. Pavement analysis and design. 2nd ed. Upper Saddle
River, NJ: Pearson.
References Kang, S. T., J. S. Kim, Y. Lee, Y. D. Park, and J. K. Kim. 2012. “Moisture
diffusivity of early age concrete considering temperature and porosity.”
Bažant, Z. P., G. Cusatis, and L. Cedolin. 2004. “Temperature effect KSCE J. Civ. Eng. 16 (1): 179–188. https://doi.org/10.1007/s12205-012
on concrete creep modeled by microprestress-solidification theory.” -1445-4.

© ASCE 04019078-14 J. Eng. Mech.

J. Eng. Mech., 2019, 145(10): 04019078

 Lederle, R. E., and J. E. Hiller. 2012. “New warping and differential drying Schindler, A., and B. Frank McCullough. 2002. “Importance of concrete
shrinkage models for jointed plain concrete pavements derived with temperature control during concrete pavement construction in hot
nonlinear shrinkage distribution.” Transp. Res. Rec. 2305 (1): 3–13. weather conditions.” Transp. Res. Rec. 1813 (1): 3–10. https://doi
https://doi.org/10.3141/2305-01. .org/10.3141/1813-01.
Liang, S., and Y. Wei. 2018. “Modelling of creep effect on moisture Suprenant, B. A. 2002. “Why slabs curl—Part II.” Concr. Int. 24 (4):
warping and stress developments in concrete pavement slabs.” Int. J. 59–64.
Pavement Eng. 19 (5): 429–438. https://doi.org/10.1080/10298436 Wei, Y., X. Gao, and W. Hansen. 2013. “Influential depth by water absorp-
.2017.1402595. tion and surface drying in concrete slabs.” Transp. Res. Rec. 2342 (1):
Liang, S., and Y. Wei. 2019. “Methodology of obtaining intrinsic 76–82.
creep property of concrete by flexural deflection test.” Cem. Concr. Wei, Y., W. Guo, and X. Zheng. 2016a. “Integrated shrinkage, relative
Compos. 97 (Mar): 288–299. https://doi.org/10.1016/j.cemconcomp humidity, strength development, and cracking potential of internally
.2019.01.003. cured concrete exposed to different drying conditions.” Drying Technol.
Lim, J. S., S. H. Kim, and J. H. Jeong. 2014. “Testing and analysis of 34 (7): 741–752. https://doi.org/10.1080/07373937.2015.1072549.
viscoelastic characteristics of solidifying concrete pavement slabs.” Wei, Y., and W. Hansen. 2011. “Characterization of moisture transport and
Downloaded from ascelibrary.org by University of Central Florida on 11/27/20. Copyright ASCE. For personal use only; all rights reserved.

KSCE J. Civ. Eng. 18 (4): 1063–1071. https://doi.org/10.1007/s12205 its effect on deformations in jointed plain concrete pavement.” Transp.
-014-0119-9. Res. Rec. 2240 (1): 9–15. https://doi.org/10.3141/2240-02.
Nguyen, H., S. Rahimi-Aghdam, and Z. P. Bažant. 2018. “Time lag in Wei, Y., and W. Hansen. 2012. “Tensile creep behavior of concrete subject
measuring pore humidity in concrete by a gage in finite cavity.” Mater. to constant restraint at very early ages.” J. Mater. Civ. Eng. 25 (9):
Struct. 51 (1): 18. https://doi.org/10.1617/s11527-018-1143-x. 1277–1284. https://doi.org/10.1061/(ASCE)MT.1943-5533.0000671.
Oliveira, M. A., M. Azenha, and P. B. Lourenço. 2015. “Simulation of Wei, Y., S. Liang, and X. Gao. 2016b. “Numerical evaluation of moisture
humidity fields in concrete: Experimental validation and parameter warping and stress in concrete pavement slabs with different water-to-
estimation.” J. Adv. Concr. Technol. 13 (4): 214–229. https://doi.org/10 cement ratio and thickness.” J. Eng. Mech. 143 (2): 04016111. https://
.3151/jact.13.214. doi.org/10.1061/(ASCE)EM.1943-7889.0001180.
Poblete, M., A. Garcia, J. David, P. Ceza, and R. Espinosa. 1990. “Moisture Yeon, J. H., S. Choi, S. Ha, and M. C. Won. 2012. “Effects of creep and
effects on the behavior of PCC pavements.” In Proc., 2nd Int. Workshop built-in curling on stress development of Portland cement concrete
on the Theoretical Design of Concrete Pavements. Hertogenbosch, pavement under environmental loadings.” J. Transp. Eng. 139 (2):
Netherlands: Vereniging Nederlandse Cementindustrie. 147–155. https://doi.org/10.1061/(ASCE)TE.1943-5436.0000451.

© ASCE 04019078-15 J. Eng. Mech.

J. Eng. Mech., 2019, 145(10): 04019078