Fracture Mechanics of Concrete and Concrete Structures -

High Performance, Fiber Reinforced Concrete, Special Loadings and Structural Applications- B. H. Oh, et al. (eds)
ⓒ 2010 Korea Concrete Institute, ISBN 978-89-5708-182-2

An analytical study on the impact of hollow shapes in bi-axial hollow

slabs
J. H. Chung & J. H. Park
Hanyang University, Seoul, Korea
H. K. Choi
Hanyang University, Seoul, Korea
S. C. LEE
SAMSUNG C&T Corporation, Korea
C. S. CHOI
Hanyang University, Seoul, Korea

ABSTRACT: This paper presents optimal hollow sphere shapes in a biaxial hollow slab. To derive optimal hol-
low shapes, numerical simulations using nonlinear Finite Element Methods were executed by the nonlinear fi-
nite element program ‘LUSAS’. Recently, various types of slab systems which can reduce self-weight of slabs
have been studied as the height and width of building structures rapidly increase. A biaxial hollow slab system
is widely known as one of the effective slab system which can reduce self-weight of slab. A biaxial hollow slab
has hollow spheres within slab in order to reduce self-weight of slab. Because of reducing self-weight of slab
by hollow spheres, size of vertical elements like walls and columns can be smaller and slabs span can be longer.
A capacity of biaxial hollow slabs is influenced by the shape and volume of hollow spheres. Therefore, in this
study, several biaxial hollow slabs which have different shapes of hollow spheres were analyzed by using the fi-
nite element method program in order to derive optimal hollow sphere shapes.

1 INTRODUCTION because it made efficient and economical building de-

1.1 Preface
In building, the slab is very important structural
member to make a space. And Slab is one of the
largest member consuming concrete. In a general
way, the slab was designed only to resist vertical
load. However, deflection and vibration of slab are
also considered recently because people are getting
more interest of residential environment. In addition,
when span of the building is increasing, deflection of
slab is more important. Therefore, the slab thickness
is on the increase. The increasing of slab thickness
makes slab heavier, and it leads to increase column
and base size. Thus, it makes building consume more
materials such as concrete and steel. Moreover, the
increasing of weight is harmful for building when
earthquake occur.
To avoid these disadvantages which were caused
by increasing of self-weight of slabs, the biaxial hol-
low slab system, also known as void slab, was sug-
gested. This slab system could optimize the size of
vertical members like walls and columns by lighten-
ing the weight of slabs. Therefore, it got attention
sign possible. A capacity of biaxial hollow slab is in-
fluenced by hollow sphere shapes. However, the re-
searches about hollow sphere shape have been
insufficient. So, in this study, several kinds of hollow
slabs which have different hollow sphere shapes were
analyzed by using finite element method program in
order to derive the optimal hollow sphere shapes and
to verify the impact of hollow shapes in biaxial hol-
low slabs.
1.2 Literature review
A biaxial hollow slab system was developed in 1990s.
In 21C, hollow slab systems which have same con-
cept and different hollow shapes were invented over
USA, Europe and Japan. (See Table 1.)
According to existing hollow slab systems, self-
weight reduction ratio of slab was 25~30%. And
their flexural strength was similar to solid slab.
1.3 Research objective
The aims of this research were to grasp the relation-
ship between hollow sphere shapes and slab’s
 D (The
J = −1.
Table h, T )existing
∇h Hollow slab systems. (1) explicitly
radius and accounts for the
hole diameter. Andevolution
numericalofsimulations
hydration
Sys-
Void Weight Strength reaction
of and SF content.
two-dimensional hollow This
slab sorption
model were isotherm
per-
Name
The proportionality
tem Materialcoefficient
Shape
reduc-
tion D(h,T) is called
FlexureShear reads It is meaningless to perform quantitative
formed.
moisture permeability and it is a30%
Sphere nonlinear function analysis of two-dimensional hollow slab model about
of the Cobiax
Europe relative humidity hEllipse
and temperature 100% 50%
T (Bažant above parameters. Because ⎡ it cannot consider⎤ change
& Najjar 1972). The moisture mass
U-Boot 35%balance
100% requires
45% of hollow shapes and concrete
we (h, α cspheres. ⎢
, α ) = G (α , α ) 1 −
web1 parts ⎥between
+
that thegreevariation in timeCuboid
Fili- Plastic
of the water mass per unit hollow s c its is
1But ⎢ meaningful
10(g α
∞ to qualitative
⎥
− α c )h ⎥ about
volumewide
USA of concrete
Void (water content 25%w) be100%
equal65%
to the analysis of two-dimensional ⎢ e 1 cslab model
⎣ hollow ⎦ (4)
divergence
slab of the
Slab moisture flux J the parameters to find qualitative effects on hollow
⎡ 10(g α ∞ − α )h ⎤
slab such as crack propagation or concentration.
1 c c
Mom-
30% 100% 50% K (α c α s ) e
⎢ , ⎥ − 1
∂
− =EJ
Japan
w slab
∇•J
Styro- Sphere
(2)
1
⎢
⎣
⎥
⎦
foam Ellipse
∂t Void 30% 100% 50%
* Compared to Solid slab
The water content w can be expressed as the sum where the first term (gel isotherm) represents the
of the evaporable water we (capillary water,hollow
water physically bound (adsorbed) water and the second
capacities.
vapor, shape
And development
and adsorbed water)
of the optimal term (capillary isotherm) represents the capillary
sphere was another aimand the research.
of this non-evaporable water. This expression is valid only for low content
(chemically bound) water w
To do these, 3 stages of analytical research
n (Mills 1966,
proc- of SF. The coefficient G1 represents the amount of
Pantazopoulo & Mills 1995). It
ess were performed by using finite element is reasonable
methodto water per unit volume held in the gel pores at 100%
assume that the evaporable
program named ‘LUSAS’. water is a function of relative humidity,
a )Typical and it can beCurve
shape : Load-Deflection expressed (Norling
(Flexural/Shear)
relative humidity, h, degree of hydration, αc, and Mjornell 1997) as
degree of silica
1) Finding out thefume reaction,ofαhollow
parameters s, i.e. weshape.
=we(h,αc,αs)
= age-dependent sorption/desorption
2) Grasping the impact of hollow sphere shapes isotherm
in
(Norling Mjonell 1997). Under this assumption and c α c+ ks α s
G (α c α s ) = k vg (5)
biaxial hollow slabs. ,
c vg s
byDeveloping
3) substitutingtheEquation 1 intosphere
optimal hollow Equation
shape.2 one
1

obtains
where kcvg and ksvg are material parameters. From the
we ∂hPARAMETERS∂wOF
2 ∂THE ∂w maximum amount of water per unit volume that can
e α&HOLLOW e α& +SPHERE
w&n
c +
− SHAPE + ∇ • ( D ∇h ) = (3) fill all pores (both capillary pores and gel pores), one
∂h ∂t h ∂α ∂α s
can
b ) calculate K1 :asLoad-Deflection
one obtains Curve (Flexural/Shear)
c s Corner radius
For finding out the parameters of hollow shape, the
existing e/∂h isshapes
where ∂whollow the slope
wereofcompared
the sorption/desorption
and analyzed
⎡
⎢ 10⎜
⎛
g α c∞ − α c ⎞⎟h ⎤⎥
isotherm
each other.(also
As a called moisture
result, three capacity).
parameters were The
de- w − 0.188 α s + 0.22α s G
c s − ⎢1 − e ⎝ 1 ⎠
⎥

governing equation (Equation 3) must be completed (6)

0 1
⎢ ⎥
rived such as typical shapes, corner radius and hole K (α c α s ) = ⎣ ⎦
by appropriate
diameter. boundary and initial conditions.
(Fig. 1) 1
,
⎛
g αc − αc h
∞ ⎞

The relation between the amount of evaporable

10⎜ ⎟
e ⎝ 1
− ⎠ 1

water and relative humidity is called ‘‘adsorption

isotherm” if measured with increasing relativity The material parameters kcvg and ksvg and g1 can
humidity and ‘‘desorption isotherm” in the opposite bec )calibrated by fitting experimental data relevant to
Hole diameter : Load-Deflection Curve (Flexural/Shear)
case. Neglecting their difference (Xi et al. 1994), in free 2.(evaporable)
Figure water content
2D FEM Load-Deflection Curve. in concrete at
the following, ‘‘sorption
a ) Typical shapes isotherm”
of hollow will bevolume)
sphere (Hollow used with various ages (Di Luzio & Cusatis 2009b).
reference to both sorption and desorption conditions.
By the way, if the hysteresis of the moisture 2.2 Temperature evolution
isotherm would be taken into account, two different
relation, evaporable
b) Corner water vshollow
radius of relative humidity, must
sphere Note that,a)at0 early
mm age, since the chemical
e) 40 mmreactions
be used according to the sign of the variation of the associated with cement hydration and SF reaction
relativity humidity. The shape of the sorption are exothermic, the temperature field is not uniform
isotherm for HPC is influenced by many parameters, for non-adiabatic systems even if the environmental
especially those that influence extent and rate of the temperature
b) 10ismmconstant. Heat conduction
f) 50 mm can be
chemical reactions and, in turn, determine pore described in concrete, at least for temperature not
structure andc) pore size distribution
Hole diameter (water-to-cement
of hollow sphere
Figure 1. The Parameters of hollow sphere shape.
exceeding 100°C (Bažant & Kaplan 1996), by
ratio, cement chemical composition, SF content, Fourier’s law, which reads
curing time and method, temperature, mix additives,
The section geometries of hollow spheres were c) 20 mm g) 60 mm
etc.). In the literature various formulations can be
studied to limit the field of the three-dimensional hol-
q = − λ ∇T (7)
low spheredescribe
found to the sorption
shapes, before isothermof of
the derivation thenormal
three-
concrete (Xihollow
dimensional et al. 1994).
sphere However,
shapes. Toindo thethis,
present
the where q is the heat flux, T is the absolute
paper the
hollow slabssemi-empirical
were modeled in expression proposed
two-dimensional by
plane, temperature,
d) 30and
mmλ is the heat conductivity;
h) 70 mm in this
changing 3 parameters such as typical shapes, cornerit
Norling Mjornell (1997) is adopted because Figure 3. Crack pattern by corner radius.

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

 D (h,flexural
J = −the
method of estimating T )∇h strength. Concrete
section is based on the compatibility of strains and
equilibrium of forces Theacting
proportionality coefficient
on the section at the D(h,T)
a) 0 mm e) 40 mm moisture permeability and it is a nonlinea
stage of failure such as reinforced bar strain reaches
0.002 or concreteofstrain
the relative
reaches humidity h andoftemperature
0.003. Strain con-
& Najjar 1972). The moisture mass balanc
crete and reinforced bar could be derived by using
that the variation in time of the water mas
equations (1) ~ (4).
b) 20 mm f) 50 mm volume of concrete (water content w) be eq
dp divergence of the moisture flux J
εs = (0.003) − 0.003 (1)
c
c) 30 mm − ∂ = ∇•J
w

Figure 4. Crack pattern by hole diameter. fs = ε s E ∂ t

(2)
s
The results of 2D FEM analysis were shown in The water content w can be expressed a
Figure 2, 3 and 4. T = As fu of the evaporable water we (capillary
(3) wa
As shown Figure 2, the parameters of two- vapor, and adsorbed water) and the non-e
dimensional hollow slab model does not affect bend- (chemically bound) water wn (Mil
C = 0.85 fc β1cb Pantazopoulo & Mills 1995). (4) It is reas
'
ing strength and stiffness of slab.
However, Figure 2 shows that the parameters af- assume that the evaporable water is a fu
fect shear strength and stiffness. And Figure 3 shows relative humidity, h, degree of hydration
In target structure, thesilica
degree of height
fumeofreaction,
compressive
αs, i.e. we=w
that the corner radius becomes smaller, the cracks stress block was changed by location. As shown Fig-
are concentrated more. Because of that, the early de- = age-dependent sorption/desorption
ure 6, the height of hollow Mjonell
(Norling sphere had to beUnder
1997). less than
this assum
struction occurs by the progress of cracks in shear 160mm to avoid by development of compressive stress
failure mode. And Figure 4 shows that the hole di- substituting Equation 1 into Equati
block in hollow parts.
obtainsHowever, as considered an-
ameter becomes larger, the shear cracks are pre- chorage of reinforced bar, 140mm was adequate
vented. height of hollow sphere in 250mm thick flat plate
∂w ∂h ∂w ∂w
e + ∇ • ( D module, e & e α& + w
c +
slabs. Considered− architectural ∇h ) = adequate
α
h t
width of hollow sphere was 270mm.
∂ ∂ h ∂α ∂α s

3 SETTING UP THE HOLLOW SPHERE SHAPE c s

3.1 Target structure where ∂we/∂h is the slope of the sorption/

isotherm (also called moisture capac
The hollow slab system is effective when self-weight governing equation (Equation 3) must be
of slab is a high rate. In other words, this system is by appropriate boundary and initial conditi
most effective when it is applied to thick flat plate The relation between the amount of e
slabs. Therefore, target structure system was set up water and relative humidity is called ‘‘
‘wall + flat plate slab’ system which had 250mm isotherm” if measured with increasing
thick slab. (See Fig. 5.) 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
Figure 6. The heightisotherm would
of compressive beblock.
stress taken into account, two
relation, evaporable water vs relative humi
be used according to the sign of the varia
3.3 Hollow sphere shape
relativity humidity. The shape of the
The hollow sphere isotherm
shapes for HPCwith
varied is influenced by many p
the limitation
Figure 5. Target structure system (wall + flat plate slab). especially
of size. And the range those volume
of hollow that influence
was setextent
up and
20% to 45% of one chemical reactions
slab module volume and,(300mm
in turn,x determ
300mm x 250mm). structure and pore
Considered size distribution
the height of hollow (water-
3.2 Size of hollow sphere sphere, a range ofratio,
cornercement
radius ofchemical composition,
hollow sphere was SF
The height of hollow sphere was set up, according to set up 0mm to 70mm. curingAndtimea and
rangemethod,
of holetemperature,
diameter mix
height of compressive stress block in target structure was set up more etc.). In the asliterature
than 30mm, considered various formulatio
concrete
slab. As assuming hollow shape was pipe type, the construction. found to describe the sorption isotherm
height of compressive stress block derived by using As compared the concrete (Xi et al.
3 parameters 1994). shape,
of hollow However,8 in th
strain compatibility method. paper the
types of hollow spheres semi-empirical
which were dividedexpression
into 3 pro
The strain compatibility method is the rigorous Norlingvaried
classes. The first group Mjornell
with the(1997)
typicalissphere
adopted b

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

 D (The
J = −2.
Table h, T )properties
∇h of hollow spheres. (1) explicitly accounts for the evolution of hydration
Rectreaction andDonuts
Donuts Rect SF content.
Round RectThis sorption
Round Rect isotherm
Solid Sphere Mushroom Ellipse Square
The proportionality coefficient D(h,T) is called reads
(D=50mm) (D=30mm) (R=70mm) (R=50mm)
moisture permeability and it is a nonlinear function
of the relative- humidity h and temperature T (Bažant
Shape ⎡ ⎤
& Najjar 31972). The moisture mass balance requires w (h α α ) = G (α α ) −
⎢ ⎥
+ 1

that the variation

Volume(cm ) - in 1436
time of the5625
water mass per unit 7380
6300 e c s
7650 c s
7785 ⎢
(g α
8910 ∞
,
10125⎥
− α c )h ⎥
,
1
, 1

27 ⎣ e 27 c
10

volume of concrete
Diameter(cm) - (water
14 content
27 w) be equal
27 to the 27 27
⎢
27 ⎦ (4) 1

divergence of- the moisture

Height(cm) 14 flux14J 14 14 14 14 ⎡ 14
(g α
∞ − α )h ⎤
14
K (α34.6%
10
c c 45%
c α s )⎢e − ⎥
Weight reduc- ⎢ 1
0% 20% 25% 28% 32.8% 34% ,
39.6% 1
tion∂w(%) ⎥
(2)
1
− = ∇•J ⎣ ⎦
∂ t

shapes such as Sphere, Mushroom, Ellipse and mesh

wheresize
thewas 30mm
term into
firstequations
(gel27.1kN. Thisrepresents
isotherm) is theoretical
the
The water
Round content w can
Rect(R=70mm). Thebesecond
expressed as thevaried
group sum value by using (5), (6).
physically bound (adsorbed) water and the second
of the evaporable water
with the corner radius such w e (capillary water,
as Square, Round water term (capillary isotherm) represents the capillary
vapor, and adsorbed
Rect(R=50mm) water) and
and Round the non-evaporable
Rect(R=70mm). And water.
M n = (This
As f yexpression
− A f )( d is
− valid
) + Aonly
' ' for low
f ( d − d ) content
' ' a '
(chemically
the bound)withwater
last group varied the hole (Mills such
wn diameter 1966,
as of SF. The coefficient G12 represents the amount of
s s s s (5)
Pantazopoulo
Round & Mills Rect
Rect(R=70mm), 1995).Donuts(D=50mm)
It is reasonableand to water per unit volume held in the gel pores at 100%
assume that the evaporable water is
Rect Donuts(D=30mm). Further information anda function of relative
2 Mhumidity, and it can be expressed (Norling
relative humidity,
properties h, degree
of the hollow spheresofare
hydration, αc, 2.
shown Table and Pn =
Mjornell
n
1997) as (6)
degree of silica fume reaction, αs, i.e. we=we(h,αc,αs) l
= age-dependent sorption/desorption isotherm
4(Norling
3D FEM Mjonell 1997). OF
ANALYSIS Under this assumption
HOLLOW SLAB and G In c the hollow
this study, s slab model was generated
(α α ) = k α c + k α s (5)
by substituting Equation 1 into Equation 2 one 1 c s vg c vg s
,

obtains
4.1 Modeling
where kcvg and ksvg are material parameters. From the
To derive the optimal hollow shape, numerical simu-
∂w ∂h ∂w ∂w maximum amount of water per unit volume that can
− e e α&Element
c +
e α&Methods
s + wn
&program fill all pores (both capillary pores and gel pores), one
lations using nonlinear Finite
+ ∇ • ( D ∇h ) =
were
∂h ∂t by the h nonlinear (3)
performed ∂α
c finite∂α element
s can calculate K1 as one obtains
LUSAS. To perform numerical simulations of the
slabs by using finite element method, two dimen-
where model
sional ∂we/∂hwas is theoften
slopeusedof the
when sorption/desorption
slabs were uni-
⎡
⎢ 10⎜
⎛
g α c∞ − α c ⎞⎟h ⎤⎥
isotherm (also called moisture
form in transverse axis. However, in the capacity). The
case of biax-
w − 0.188 α s + 0.22α s G
c s − ⎢1 − e ⎝ 1 ⎠
⎥

governing equation (Equation 3) must be completed (6)

0 1
⎢ ⎥
ial hollow slabs, it was impossible to use two- K1(α c ,α7.s ) Finite ⎣ ⎦
by appropriate
dimensional boundary
model because andofinitial conditions.section
the unequally
Figure =
element mesh
10⎜ g α
⎛
modeling
∞
− α ⎟h
⎞
(In the view of the
⎝ 1 c c⎠
cross-section).
The relation
geometry along thebetween the amount
longitudinal of evaporable
and transverse axis e −1

by extraordinary hollow sphere shape. ‘‘adsorption

water and relative humidity is called So, it was
isotherm”using if measured with increasing relativity The material parameters kcvg and ksvg and g1 can
σ

modeled three-dimensional model which could σy

humidity and ‘‘desorption isotherm”

be considered concrete web parts between hollow in the opposite be calibrated by fitting experimental data relevant to
case. Neglecting
spheres like Figure their
7. difference
And to generate (Xi et al. 1994),ele-
a finite in free (evaporable) water content in concrete at
the following, ‘‘sorption isotherm”
ment mesh of the hollow slab which has extraordi- will be used with various ages (Di Luzio & Cusatis 2009b).
εy εu ε

reference
nary shapes to of
both sorption
hollow and inside,
sphere desorption conditions.
tetrahedral ele-
By the way, if the hysteresis
ments which have four nodes were used because of the moisture 2.2 Temperature evolution
isotherm
other mesh would be taken
elements such into account, two
as Pentahedral different
or Hexahe- a) Bi-linear model b) LUSAS Concrete Model 94
dral were not able to generate the geometry ofmust
relation, evaporable water vs relative humidity, the Note that,
Figure at earlymaterial
8. Nonlinear age, since
model. the chemical reactions
be usedslab.
hollow according to the sign of the variation of the associated with cement hydration and SF reaction
relativity humidity.
Two material models Thewereshapeused to of perform
the sorption
nonli- are exothermic, the temperature field is not uniform
isotherm
finitefor HPC ismethod
influenced by many
The parameters, for non-adiabatic systems even if the environmental
40

near element analysis. one, which 39

especially
was those that influence
used reinforced steel bar, extent and rate
is bi-linear modelof the
of temperature is constant. Heat conduction can be
38
37

chemical reactions and, in turn, determine

steel. It is assumed that steel behavior will be totally pore described in concrete, at least for temperature not
36
35

structure and inpore

elasto-plastic size and
tensile distribution
compressive (water-to-cement
loading con- exceeding 100°C (Bažant & Kaplan 1996), by
34
Load (kN)

ratio, cement
(See Fig.chemical
8-a) Andcomposition, SF content, Fourier’s law, which reads
33

dition. another material model,

32

curing time and method, temperature,

which was used concrete, is ‘LUSAS concrete mix additives,
model
31
30

etc.).ItIncan
94’. the consider
literaturemulti-cracks
various formulations
of concrete canand
be q = − λ ∇T 29
28
(7)
found to describe the sorption
strength softening of concrete.(Fig. 8-b) isotherm of normal Pn 27

concrete (Xi et al. 1994). studyHowever, in the present

26

where q is the heat flux, T is the absolute

25
The convergence was performed about
paper the semi-empirical expression proposed by
160 140 120 100 80 60 40 20

mesh size ranged 20mm~150mm to verify the nu- temperature, and λ is the heat conductivity; in this Mesh Size (mm)

merical simulation result. The result converged whenit

Norling Mjornell (1997) is adopted because Figure 9. Nonlinear material model.

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

a finite element mesh with 30mm-sized. Because the = − D (h, T )were
The conditions ofJ support ∇h fixed. Because, In
use of small mesh size less than 30mm did not change FEM Analysis, simple supported condition might
the results significantly, but led to an increased com- The asproportionality
bring about error such coefficient
stress concentration. The D(h,T)
putation time. The relationship between mesh size moisture permeability and it is a nonlinea
reason of using distributed load was the same. Loads
and convergent tendency in this hollow slab model is were imposed onof slabs
the relative
in order humidity h and temperature
likes self-weight,
shown Figure 9. dead load and live&load
Najjar 1972). The moisture mass balanc
until the slab was destroyed.
that the and
Further information variation in timeofof hollow
properties the water mas
4.2 Hollow slab analysis model volume of concrete (water content w) be eq
sphere is shown Table 3.
divergence of the moisture flux J
To perform 3D nonlinear FEM analysis, the proper-
ties of hollow slab model was idealized. It had 8.9m 5 RESULTS AND∂ANALYSIS
length, 300mm width. The reasons of performing
w
− = ∇•J
∂
t
idealize process are like this 5.1 Results of numerical simulations
It was possible to estimate the whole capacity of The 3D hollow slabThe water applied
models, content above
w can be expressed a
hollow
slab with the capacity of one module slab model be- of the evaporable water
spheres (Table 2), were analyzed by numerical w (capillary wa
e simu-
cause hollow spheres were located with uniform gap vapor, Finite
lations using nonlinear and adsorbed
Element water)
Method andpro-the non-e
toward width direction of slab. gram. (chemically bound) water wn (Mil
So, to compare the impact of hollow shapes on Pantazopoulo
Table 4 is shown & Mills
that results 1995). It is reas
of computation
flexural capacity of slabs, it was enough to perform about 8 cases ofassume
hollow that
slabsthe evaporable
which water is a fu
can be distin-
FEM analysis with one module model of slab. And guished by hollowrelative
sphere humidity,
shapes andh,1 degree
case of of hydration
solid
this way was time-efficient than using whole slab slab. degree of silica fume reaction, αs, i.e. we=w
model. In the results = age-dependent
of computation, 8 casessorption/desorption
of hollow
slabs were judged(Norling Mjonellload.
safe in design 1997).AndUnder
somethis of assum
Table 2. Properties of hollow slab model.
hollow slabs showbya good
substituting Equation
load bearing 1 into
capacity and Equati
Width 300 mm stiffness, comparedobtains
with solid slab,
Height 250 mm The results, analyzed more concretely, are as fol-
Length 8900 mm lows: ∂w ∂h ∂w ∂w
Upper D10 x 2 − e
As compared difference ∇ • (D ∇
+ design h
load e ultimate
) = and α &c +
e α& + w
Reinforced bar ∂h ∂t h ∂α ∂α s
Lower D13 x 2 load, hollow slab applied 'Rect Donuts(D=50mm)' c s
Boundary condition Fixed end was shown the largest difference, 26.24kN. It means
Load Self-weight 330~600kg/m2 where ∂w
additional load bearing e/∂h is of
capacity theslab
slopeafter
of the
deadsorption/
Dead load 185kg/m2
load. Because a isotherm
dead load (also
varied called moisture
with hollow vol- capac
Live load 200kg/m2 ~
governing
ume, it was reasonable equation real
to compare (Equation
capacity 3) of
must be
hollow slab. by appropriate boundary and initial conditi
fck 24 MPa The deflections inThe relation
design loadbetween
were variedthe amount
with of e
Fy 400 MPa weight reductionwater and relative
of slabs. humidity is called ‘‘
Round Rect(R=50mm)
isotherm” if measured with increasing
Table 4. Numerical simulation result of three-dimensional hollow slab model.
humidity and ‘‘desorption isotherm” in th
Rect Rect
case. Neglecting their difference (Xi et al.
Solid Sphere Mushroom Ellipse Donuts Donuts theRound
following,Round
‘‘sorptionSquare
isotherm” will be
(D=50) (D=30) reference to both
Rect(R=70) sorption and desorption c
Rect(R=50)
Self-weight By3.8the way,3.5if the hysteresis of the
(kN/m2)
5.9 4.7 4.4 4.2 3.9 3.9
isotherm would be taken into account, two
3.2
Dead Load
9.7 8.5 8.2 8 7.7 7.7
relation, evaporable water vs relative humi
(kN/m2) be7.6 7.3
used according to the7 sign of the varia
Ultimate Load
34.36 33.80 33.35 33.97 33.94 33.50
relativity
33.12
humidity.
32.08
The
28.50
shape of the
(kN/m2) isotherm for HPC is influenced by many p
Dead − Load
28.2% 25.1% 24.6% 23.6% 22.7% 23.0%
especially
22.9%
those that influence
22.8%
extent and
Ultimate −Load chemical reactions and,24.6%
in turn, determ
Deflection at
2.16 1.86 1.72 1.68 1.57 1.58
structure
1.61
and 1.56
pore size distribution
1.61
(water-
D.L (mm) ratio, cement chemical composition, SF
Deflection at
22.42 21.32 22.60 23.71 23.33 24.02 curing
24.23time and method,22.30
24.68 temperature, mix
U.L (mm) etc.). In the literature various formulatio
Failure mode
F F F F F F found
F to describe
F the sorption
F isotherm
at Design Load concrete (Xi et al. 1994). However, in th
Failure mode
F F F F F F paper
F the semi-empirical
F F+S expression pro
at U.L
* F : Flexural Crack Occur, S : Shear Crack Occur
Norling Mjornell (1997) is adopted b

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

 − D (h,Donuts(D=30mm)
J = Rect
and T ) ∇h were the smallest deflec- (1) explicitlyofaccounts
amount influenceforbythe evolution
corner radiusofwashydration
about
tion at dead load about 72% of solid slab. And the reactionofand
4~17% solidSF
slab,content.
with theThis
aspectsorption isotherm
of strength.
The proportionality
deflections in ultimate load coefficient
were varied with isits called
D(h,T) load- reads
moisture permeability and it is a nonlinear function
bearing capacity and stiffness. 5.2.3 Impact of hole diameter
of Looking
the relative humidity
at the and temperature
failurehmode, all slabs wereT (Bažant
shown To compare impact of hole ⎡ diameter, 4-types ⎤ hollow
& Najjar 1972). The moisture mass balance requires
flexural crack in design load. When load reaches ul- spheres having different⎢ hole diameter
we (h, α 10-c) G1 (αshown were⎥ used.
c , α s ) =was 1−
c , α s )⎢the +
1

that the variation in time of the water mass per unit

timate load, the failure mode was shown flexural fail- Figure ∞
load-deflection
10(g α
⎥curve of
− α c )has
volume
ure of concrete
behavior except (water
Squarecontent w) beSquare
shape. The equal shape
to the the having different hole ⎢ e 1 c such
⎣ diameter ⎥
⎦ Solid,
(4)
divergence
had flexuralofandthe moisture flux Jbehavior by judging
shear failure Round Rect(R=70mm), Rect Donuts(D=50 mm) and
α ∞ − α as
with strain of reinforced bar. Rect Donuts(D=30mm).⎡⎢ 10 It (gshowed
1 c c )h following.
⎤

When hole diameter K1 (α c , α was

s )⎢e increased, slab − 1⎥strength
− ∂ = ∇•J
w
(2) ⎣
⎥
⎦
5.2∂ Impact of hollow sphere shape
t
35

To The
find water
out the impactw of
content canhollow
be weresphere as
expressed shape,
the the
sum where the Solid
first term (gel isotherm) represents the
results of each
of the evaporable group which divided
(capillary
water 3wewere water,into
water3 physically
30
bound
Sphere (adsorbed) water and the second
classes
vapor, ofin above
andeach chapter
adsorbed water) analyzed. And
and the non-evaporable the term25 (capillary
Mushroom
Ellipse isotherm) represents the capillary
impact
(chemically parameterwater
bound) is as following.
w (Mills 1966, water. ThisRound
expression is valid only for low content

Load(kN/mm )
Rect

2
Pantazopoulo & Mills hollow
1995).shape
n
It is reasonable to of SF.
20 coefficient G1 represents the amount of
The(r=70mm) 10

5.2.1
assume Impact
that of typical
the evaporable water is shape,
a function of water per unit volume held in the gel pores at 100%
To compare
relativespheres impact
humidity, of typical hollow
h, degree of hydration, αc, and 5-types relative
15
humidity, and it can be expressed (Norling

Load(kN/mm )
2
hollow
degree of silica having
fume difference
reaction, α shape and volume
s, i.e. we=we(h,αc,αs)
Mjornell
10
1997) as 5

were used because a typical

= age-dependent hollow shape correlates
sorption/desorption isotherma Dead Load(Solid)
with a
(Norling hollow
Mjonellvolume.
1997). The
Underrelationship
this betweenand
assumption c α c+ ks α s
G (α5c α s ) = k vg (5)
c vg s
0
,
typical hollow shape and hollow volume was not lin-
0

by but
ear
substituting
staircase
Equation
type.
1 into Equation 2 one 1
0
Deflection(mm)

obtains 0 5 10 15 20 25
Figure 10-a) was shown the load-deflection curve where kcvg and ksvg areDeflection(mm)
material parameters. From the
of the typical sphere shapes such as Solid, Sphere,
∂w ∂h maximum a amount of water
)Typical shape per unit volume
: Load-Deflection Curve that can
− e
Mushroom, Ellipse and∂wRound we
e α& + ∂Rect(R=70mm).
α&s + dead
w&n load. fill all poresSolid
(both capillary pores and gel pores), one
(3)It
35
+ ∇ • ( D ∇h ) =
∂h ∂t differenth behaviors
showed ∂α α at
of ∂slab
c
K1 as
can 30calculateRound one obtains
c s
When volume of hollow sphere was increased, slab
Rect(r=70mm)
Round Rect(r=50mm)
stiffness and deflection tended to be decreased. Square
/∂h is the of
where ∂wereduction slope of thewassorption/desorption 25 ⎡ ⎛
g α c∞ − α c ⎞⎟h ⎤⎥
Load(kN/mm )

However, stiffness not according to ⎢ 10⎜

2

changing of hollow shape. As seeingcapacity).

isotherm (also called moisture The w α s + 0.22α s G
− 0.188
c s − ⎢1 − e ⎝ 1 ⎠
⎥
the crack pat-
governing equation (Equation 3) must be completed (6)
20 0 1
⎢ ⎥
tern, it was according to reduction of section by hol- K (α c α s ) = ⎣ ⎦
by appropriate
volume. boundary and initial conditions.
10
,

low 151
g αc − αc h
⎛ ∞ ⎞

The relationwhen between

cornerthe amount
and ofholeevaporable
10⎜ ⎟
e ⎝ 1 ⎠
Load(kN/mm )
2

−
Therefore, radius diameter 1

water and relative humidity is

were same and hollow volume was increasing, called ‘‘adsorption
stiff-
10
Dead Load(Solid)

isotherm” if measured with increasing

ness and deflection of slab tended to decrease at dead relativity The materialDead
parameters
(Round
Load
kcvg and ksvg and g1 can
Rect(r=50mm)
humidity
load. and ‘‘desorption
At ultimate load, an isotherm”
amount ofin influence
the opposite by be calibrated by(Round
5
fitting experimental data relevant to
Rect(r=70mm)
5
1.6 1.8 2.0 2.2 2.4 2.6 2.8
Deflection(mm)
3.0 3.2 3.4 3.6 3.8 4.0

case. Neglecting
typical hollow shape theirwas
difference
about (Xi et al. 1994),
1.6~3.6% of solidin free 0 (evaporable) water content in concrete at
(Square)

the following, ‘‘sorption isotherm”

slab, with the aspect of strength. will be used with various0
ages (Di5 Luzio & 10
Cusatis152009b). 20 25

reference to both sorption and desorption conditions. Deflection(mm)

By theImpact
5.2.2 way,ofifcorner
the radius
hysteresis of the moisture b ) Corner radius : Load-Deflection Curve
2.2 35Temperature evolution
isotherm
To compare would be taken
impact into radius,
of corner account,4-types
two different
hollow Solid
relation, evaporable water vs relative
spheres having different corner radius were used.humidity, must Note30 that, atRound
early age, since the chemical reactions
Rect(r=70mm)

Figure 10-b) was shown the load-deflection curvethe

be used according to the sign of the variation of associated with cement hydration and SF reaction
Rect Donut(D=30mm)
of
relativity humidity. The shape of
the having different corner radius such as Solid, the sorption are 25exothermic, the temperature field is not uniform
Rect Donut(D=50mm)
Load(kN/mm )

isothermRect(R=70mm),
for HPC is influenced
Round by many parameters, for 20non-adiabatic systems even if the environmental
2

Round Rect(R=50mm)
Square. It showed as following. Like and
especially those that influence extent rate ofand
a tendency the
of temperature is constant. Heat conduction can be
10.0

chemical reactions and, in turn, determine

2D FEM analysis, when corner radius was increased, pore described
15 in concrete, at least for temperature not 9.5

structure and pore size distribution

tended (water-to-cement exceeding 100°C (Bažant & Kaplan 1996), by
L o ad (k N /m m 2 )

9.0

slab stiffness and strength to be increased. Dead Load(Solid)

ratio, cement
However, chemical
as seeing composition,
the deflection at deadSFload,
content,
stiff- Fourier’s law, which
10 reads 8.5

curing time and method, temperature,

ness was not only according to changing of corner mix additives, Dead Load
(Round Rect(r=70mm)
8.0

etc.). In
radius butthealsoliterature
changingvarious
of hollow formulations
volume by cancornerbe q = 5− λ∇T (Rect Donut(D=30mm)
7.5
2.2 2.4 2.6 2.8 3.0

(7)
found to describe the sorption isotherm of normal
Deflection(mm)
(Rect Donut(D=50mm)
radius. 0
concrete
Therefore,(Xi when
et al. corner
1994). radius
However, in the present
was increasing and where0 q is 5the heat10 flux, 15 T is the20 absolute
25

the other conditions were same, stiffnessproposed

paper the semi-empirical expression and strength by temperature, and λ is Deflection(mm)
the heat conductivity; in this
of slab tended to increasing. At ultimate load, anit
Norling Mjornell (1997) is adopted because c ) Hole diameter : Load-Deflection Curve
Figure 9. 3D FEM Load-Deflection Curve.

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

 J = − D ( h , T ) ∇h

The proportionality coefficient D(h,T)

moisture permeability and it is a nonlinea
of the relative humidity h and temperature
& Najjar 1972). The moisture mass balanc
that the variation in time of the water mas
volume of concrete (water content w) be eq
divergence of the moisture flux J

− ∂ = ∇•J
w

a )Typical shape : Sphere b ) Typical shape : Elliptical Sphere

∂ c ) Typical shape : Cuboid
t

Figure 10. Parametric analysis.

The water content w can be expressed a
tended to be increased a little, 1.3~4% of solid slab. tion (at same loadingthe
of evaporable
condition) werewater we (capillary wa
considered.
However, as seeing the deflection and stiffness, 1) Safety vapor, and adsorbed water) and the non-e
hollow slabs, applied Rect Donuts(D=50mm) and (chemically
Judged from safety, bound)
all hollow watersafewinn (Mil
slabs were
Rect Donuts(D=30mm), showed bigger stiffness and Pantazopoulo & Mills
design load. However, failure mode of square 1995).type
It is reas
smaller deflection than hollow slab, applied Round hollow slab was shear in ultimate load. The shear is a fu
assume that the evaporable water
Rect(R=70mm) which had no hole. In addition, vol- relative humidity,
failure mode is dangerous because ith,is degree of hydration
very sudden
ume of Round Rect(R=70mm) was more than vol- degree of silica fume reaction,
and brittle. Therefore square shape hollow sphere αs, isi.e. we=w
ume of Rect Donuts(D=50mm) and Rect Donuts = age-dependent
not suitable for the hollow slab system. sorption/desorption
(D=30mm). Therefore, the hole of hollow sphere 2) Strength (Norling Mjonell 1997). Under this assum
functions as preventing a deflection of slab. As Compared by substituting
to solid Equation
slab, merely 3 cases1ofinto
hol- Equati
obtains
low sphere shapes, 'Rect Donuts(D=50mm)', 'Rect
Donuts(D=30mm)' and 'Round Rect(R=70mm)' were
5.3 Parametric analysis w ∂h without decreasing
good load-bearing ∂capacity ∂w we
− e (See +∇ • ( D11-a)
∇h ) = e α& + ∂of α&s + w
Based on the results of impact of hollow shape weight reduction ratio.
∂h ∂t Fig. h ∂α c
∂α
(chapter 5.2), parametric analysis was performed. 3) Deflection c s
The unknown results, which were not performed The hollow slab stiffness tended to decrease be-
numerical simulations, were assumed by linear inter- cause of reductionwhere is the
∂we/∂harea.
of section So,slope of the sorption/
as comparing
polation. isotherm (also called moisture capac
A parametric analysis was performed as following governing equation (Equation 3) must be
conditions. The typical shape parameters were fix by appropriate boundary and initial conditi
and the other parameters such as corner radius and The relation between the amount of e
hole diameter were changing. As compared the ulti- water and relative humidity is called ‘‘
mate load-bearing capacity, the effects of corner ra- isotherm” if measured with increasing
dius and hole diameter were evaluated in combina- humidity and ‘‘desorption isotherm” in th
tion. Figure 10 was shown the tendency of the each case. Neglecting their difference (Xi et al.
parameter in combination as following. the following, ‘‘sorption isotherm” will be
reference to both sorption and desorption c
1) When the typical hollow shape is same, and the By the way, if the hysteresis of the
corner radius or the hole diameter is increasing, the isotherm would be taken into account, two
capacities of hollow slab is increasing. relation,
a ) Ultimate evaporable
Load-Dead water vs relative humi
Load capacity
2) When the typical hollow shape is cuboid type, the be used according to the sign of the varia
increase by corner radius or hole diameter will be relativity humidity. The shape of the
larger than the other typical hollow shape. isotherm for HPC is influenced by many p
especially those that influence extent and
chemical reactions and, in turn, determ
6 DEVELOPMENT OF THE OPTIMAL structure and pore size distribution (water-
HOLLOW SPHEAR SHAPE ratio, cement chemical composition, SF
curing time and method, temperature, mix
Based on the results of numerical simulation, the op- etc.). In the literature various formulatio
timal hollow sphere shape was derived. To derive the found to describe the sorption isotherm
optimal hollow sphere shape, three aspects of criteri- concrete (Xi et al. 1994). However, in th
ons which were safety, strength (additional load paper the semi-empirical expression pro
b )Norling
Deflection Mjornell
at same load(1997) is adopted b
bearing capacity of slab after dead load) and deflec- Figure 11. Derivation of the optimal hollow sphere shape.

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

 D (h, T )∇h at same load, the optimal hollow
J = −deflection
the (1) explicitly accounts for the
ACKNOWLEDGMENTS evolution of hydration
sphere shape was derived. (See Fig. 11-b) reaction and SF content. This sorption isotherm
The proportionality coefficient D(h,T) is called readswork was supported by Samsung Construction
This
moisture permeability
After all and it'Rect
the analysis, is a Donuts(D=50mm)'
nonlinear function and Trade Inc, 2009
of the which
shape relativehad
humidity h andwas
hole inside temperature (Bažant
judged asTthe opti- ⎡ ⎤
& Najjar 1972). The moisture mass balance requires
mal hollow sphere shape.
we (h, α c , α s ) = G1 (α c , α s )⎢⎢1 − + 1 ⎥
that the variation in time of the water mass per unit REFERENCES (g α
∞ − α c )h ⎥⎥
volume of concrete (water content w) be equal to the e ⎢
⎣ c 10

(4)1
⎦
7divergence
CONCLUSIONof the moisture flux J Aldejohann. M. & Schnellenbach-Held. 2003. Investigations
on the Shear Capacity of ⎡Biaxial ∞ − α Slabs
(g αHollow )h -⎤ Test Re-
K (α c Darmstadt
sults and Evaluation. c c
α ) e Concrete V. 18− ⎥
⎢
10
1
In ∂this study, several hollow slabs which had different s Building
⎢
,

⎥
1

(2)
1
ACI Committe 318. 2005. Code Requirement for
− w = ∇sphere
hollow • J shapes were analyzed to compare with ⎣ ⎦
∂t Structural Concrete. American Concrete Institute.
structure capacity and failure mode by using finite J.H. Chung, N.K. Ahn, H.K. Choi. & C.S. Chang. 2009. An
element method
The water program.
content w canBased on the as
be expressed results
the sumof where the study
analytical first ofterm (gelhollow
optimal isotherm) represents
sphere shapes the
in hollow
numerical
of the simulation,
evaporable the impacts of hollow
we (capillary water, water
water sphere sphere physically
slab. Journalbound of the(adsorbed) water
korea institute for and the second
structural mainte-
and the
vapor,There optimal
and adsorbed hollow and shape
water) about the could be de-
non-evaporable term (capillary
nance. 159-162 isotherm) represents the capillary
rived.
(chemically are conclusions
bound) water this study.
wn (Mills 1966, water. This expression is valid only for low content
J.H. Chung, H.K. CHOI, S.C. LEE, J.K. Oh. & C.S. CHOI.
2009. An Analytical Study of the Impact of Hollow Sphere
1) Capacity &
Pantazopoulo of hollow
Mills slab and
1995). It failure mode was
is reasonable to of onSF.Biaxial
The Hollow
coefficient G1 represents
slab. Proceeding theconference
of annual amount of of
related
assume with
that hollow
the sphere
evaporable shape.
water Especially,
is a function theseof water per unit volume
the architectural instituteheld in the
of korea. gel pores at 100%
475-478
are proved
relative that corner
humidity, radius, hole
h, degree diameter and
of hydration, , vol-
and
αcthem. relative
J.H. Chung,humidity,
H.K. CHOI, andS.C.
it can
LEEbe& expressed
C.S.CHOI. (Norling
2009. An
ume
degree of hollow
of silica sphere
fume were closely
reaction, α related with
s, i.e. we=we(h,αc,αs)
Mjornell
analytical1997)
studyas of optimal hollow sphere shapes in hollow
Slabs. Computational Design in Engineering.
= 2)age-dependent
As corner radiussorption/desorption
of hollow sphere wasisotherm smaller,
LUSAS Co. 1985. LUSAS : Modeller Reference Manual. LU-
crack caused
(NorlingThe Mjonellby concentrated
1997). Under stress was
thishaving developed
assumption and G1SAS
(α , α
c
)= k α c + k α s
Corporation. s (5)
earlier.
by substituting hollow sphere
Equation shapes
1 appropriate
into Equation corner2 onera- c s
Schnellenbach-Held & KarstensPfeffer. 2002. Punching behav-
vg c vg
dius
obtains more than 50mm was in 250mm ior of biaxial hollow slabs. Cement & Concrete Compos-
thick hollow slab. where kcvg24.and
ites. V.
vg are material parameters. From the
I. 6.ks551-556
Z. Lounis. & M.Z. Cohn. 1993. Optimization of precast
3) The hole of hollow sphere functions as prevent-
∂w ∂h we as hole ∂w maximum amount of water per unit volume that can
− e
of slab. ∂And & e diameter
α&s +ofw&nslab
prestressed concrete bridge girder systems. PCI J. 123 (3)
c +
ing a deflection
+ ∇ • ( D ∇h ) = α of hol-
(3) fill60–
all pores
77. (both capillary pores and gel pores), one
low∂hsphere
∂t was hsmaller, ∂α the deflection
∂α was can calculate K1 as one obtains
bigger.
c s
4) The hollow slab having optimal hollow sphere,
where
Rect ∂we/∂h is the slopethat
donuts(D=50mm), of the sorption/desorption
demonstrated in this
⎡
⎢ 10⎜
⎛
g α c∞ − α c ⎞⎟h ⎤⎥
study showed more than 99% of load capacity).
isotherm (also called moisture The w − 0.188 α s + 0.22α s G
c s − ⎢1 − e ⎝ 1 ⎠
⎥
resisting capac-
governing equation (Equation 3) must be completed (6)
0 1
⎢ ⎥
ity and less than 72% deflection at design load which K (α c α s ) = ⎣ ⎦
by appropriate
compared normalboundary and initial itconditions.
slab. Therefore,
1
,
⎛
g αc − αc h
∞ ⎞

the The
Rectrelation amount ishollow
between theis optimal ofproved
evaporablethat 10⎜ ⎟

donuts(D=50mm) sphere e ⎝ 1
− ⎠ 1

water inand
shape 250mmrelative
thickhumidity
hollow slab. is called ‘‘adsorption
isotherm” if measured with increasing relativity
humidity and ‘‘desorption isotherm” in the opposite
case. Neglecting their difference (Xi et al. 1994), in
the following, ‘‘sorption isotherm” will be used with
reference to both sorption and desorption conditions.
By the way, if the hysteresis of the moisture
isotherm would be taken into account, two different
relation, evaporable water vs relative humidity, must
be used according to the sign of the variation of the
relativity humidity. The shape of the sorption
isotherm for HPC is influenced by many parameters,
especially those that influence extent and rate of the
chemical reactions and, in turn, determine pore
structure and pore size distribution (water-to-cement
ratio, cement chemical composition, SF content,
curing time and method, temperature, mix additives,
etc.). In the literature various formulations can be
found to describe the sorption isotherm of normal
concrete (Xi et al. 1994). However, in the present
paper the semi-empirical expression proposed by
Norling Mjornell (1997) is adopted because it
The material parameters kcvg and ksvg and g1 can
be calibrated by fitting experimental data relevant to
free (evaporable) water content in concrete at
various ages (Di Luzio & Cusatis 2009b).
2.2 Temperature evolution
Note that, at early age, since the chemical reactions
associated with cement hydration and SF reaction
are exothermic, the temperature field is not uniform
for non-adiabatic systems even if the environmental
temperature is constant. Heat conduction can be
described in concrete, at least for temperature not
exceeding 100°C (Bažant & Kaplan 1996), by
Fourier’s law, which reads
q = − λ ∇T (7)
where q is the heat flux, T is the absolute
temperature, and λ is the heat conductivity; in this

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