Construction

and Building

Construction and Building Materials 21 (2007) 150–156

MATERIALS
www.elsevier.com/locate/conbuildmat

Hybrid fiber reinforced self-compacting concrete

with a high-volume coarse fly ash
Mustafa Sahmaran, I. Ozgur Yaman *

Department of Civil Engineering, Middle East Technical University, 06531 Ankara, Turkey

Received 25 November 2004; received in revised form 16 June 2005; accepted 30 June 2005
Available online 19 August 2005

Abstract

This paper presents a study on the fresh and mechanical properties of a fiber reinforced self-compacting concrete incorporating
high-volume fly ash that does not meet the fineness requirements of ASTM C 618. A polycarboxylic-based superplasticizer was used
in combination with a viscosity modifying admixture. In mixtures containing fly ash, 50% of cement by weight was replaced with fly
ash. Two different types of steel fibers were used in combination, keeping the total fiber content constant at 60 kg/m3. Slump flow
time and diameter, V-funnel, and air content were performed to assess the fresh properties of the concrete. Compressive strength,
splitting tensile strength, and ultrasonic pulse velocity of the concrete were determined for the hardened properties. The results indi-
cated that high-volume coarse fly ash can be used to produce fiber reinforced self-compacting concrete, even though there is some
reduction in the concrete strength because of the use of high-volume coarse fly ash.
Ó 2005 Elsevier Ltd. All rights reserved.

Keywords: Self-compacting concrete; Fiber reinforcement; High-volume coarse fly ash

1. Introduction
The term fiber reinforced concrete (FRC) is defined
by ACI 116R, Cement and Concrete Terminology, as
a concrete containing dispersed randomly oriented
fibers. Inherently concrete is brittle under tensile loading
and mechanical properties of concrete may be improved
by randomly oriented short discrete fibers which prevent
or control initiation, propagation, or coalescence of
cracks [1]. The character and performance of FRC
changes depending on the properties of concrete and
the fibers. The properties of fibers that are usually of
interest are fiber concentration, fiber geometry, fiber
orientation, and fiber distribution. Using a single type
of fiber may improve the properties of FRC to a limited
level. However, the concept of hybridization, adding
*
Corresponding author. Tel.: +90 312 210 5473; fax: +90 312 210
1262.
E-mail address: ioyaman@metu.edu.tr (I.O. Yaman).
two or more types of fiber into concrete, can offer more
attractive engineering properties as the presence of one
fiber enables more efficient utilization of the potential
properties of other fiber [2,3].
On the other hand, use of self-compacting concrete
(SCC) in the construction industry has grown signifi-
cantly due to its technical advantages [4]. Generally,
SCC is achieved using new generation superplasticizers
to reduce the water–binder ratio. In addition, supple-
mentary cementitious or inert materials such as lime-
stone powder, natural pozzolans, and fly ash is also
used to increase the viscosity and reduce the cost of
SCC. Among these materials fly ash, a by-product of
thermal power plants, has been reported to improve
the mechanical properties and durability of concrete
when used as a cement replacement material [5]. Con-
cretes having large amounts of fly ash are termed as
high-volume fly ash (HVFA) concrete. HVFA concrete
was initially developed for mass concrete applications
to reduce the heat of hydration. Later on, it was used

0950-0618/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.conbuildmat.2005.06.032
 M. Sahmaran, I.O. Yaman / Construction and Building Materials 21 (2007) 150–156 151

in structural and pavement applications, because of its Table 1

durability and good mechanical properties [6]. Fly ash Chemical composition of the Portland cement and mineral admixtures
is usually separated at the power plants and high quality Chemical Portland Fly ash Limestone
(fine) fly ash meeting the fineness requirement of ASTM analyses (%) cement powder
C 618 can be used in producing blended cements or CaO 61.94 11.31 54.97
added as a separate ingredient at the ready mixed SiO2 18.08 49.55 0.01
Al2O3 5.58 13.34 0.17
concrete batching plants. In addition to this fine fly Fe2O3 2.43 8.51 0.05
ash, there are vast amounts of substandard (coarse) fly MgO 2.43 4.10 0.64
ash that can be utilized in the concrete industry. A suc- SO3 2.54 1.70 0.00
cessful application of the coarse fly ash in producing K2O 0.99 1.99 0.00
blended Portland cements was published by the Na2O 0.18 3.08 0.00
LOI 4.40 2.74 43.66
researchers at CANMET [7]. Fly ash has also been
increasingly used in the Turkish concrete industry.
Recently, to increase the use of fly ash, investigations 2.2. Limestone powder
on HVFA in producing SCC are being performed [8,9].
In this paper, an application of a coarse fly ash will Limestone powder (LP) was used as a mineral viscos-
be presented on SCCs incorporating hybrid fiber rein- ity enhancing admixture. LP was a by product of marble
forcement. Previous investigations showed that the extraction with a CaCO3 content of 98% and a specific
use of steel fibers in SCC is feasible [10,11]. In these gravity of 2.70. The chemical composition of the lime-
mixtures, steel fibers can decrease workability of SCC stone powder is also presented in Table 1.
as the fiber amount and slenderness ratio (length/diam-
eter) increase. However, in case of well-proportioned 2.3. Fly ash
SCC the workability is not influenced by the steel fibers
[11]. The incorporation of fibers in concrete improves A fly ash (FA) from Çayırhan, Turkey was used in
mechanical properties of concrete such as ductility, this study. Its chemical composition is given in Table 1.
toughness, tensile strength, impact resistance and fati- The FA had a relatively low specific gravity and Blaine
gue. fineness of 2.01 and 2420 cm2/g, respectively. The per-
The objective of this study is to assess the effects of centage of fly ash retained when wet sieved on a 45-
HVFA replacement on the fresh and hardened proper- lm sieve was 46. Therefore, this FA failed to meet the
ties of SCCs incorporating different types of steel fibers. fineness requirements of ASTM C 618. To confirm the
Moreover, the fly ash used in this study was a coarse fly fineness of the FA, the particle size distribution of the
ash that does not meet the fineness requirements of FA was also determined. Fig. 1 shows the particle size
ASTM C 618. Even though, the suitability of using such distribution of the FA, as well as the LP, and PC used
a substandard fly ash needs much detailed investiga- in this study. As can be seen from that plot, FA was
tions, this study covers some fresh and hardened proper- much coarser compared to both PC and LP.
ties of such mixtures. In addition to the fly ash, two
different sizes of steel fibers were used at different 2.4. Fibers
proportions in making the concrete. Total mass of
cementitious materials is 500 kg/m3, in which 50% of Two cylindrical steel fiber types, one with hooked
cement is replaced by the coarse grained fly ash. For ends (SF1) and one straight (SF2) were used (Fig. 2).
comparison, a control SCC mixture without any fly
ash was also produced. The commercially available 100
chemical admixtures used in this study included a visco- 90
sity modifying admixture (VMA) and a polycarboxylic- 80
based superplasticizer (SP). 70
% Above

60

50
2. Materials
40
Fly ash (FA)
2.1. Portland cement 30
Portland cement (PC)
20
Lime stone powder (LP)
The cement used in all mixtures was a commercially 10
available Portland cement (PC), which corresponds to 0
0 1 10 100 1000
ASTM Type I cement. It had a specific gravity of 3.09
Particle Diameter (µm)
and Blaine fineness of 3030 cm2/g. The chemical compo-
sition of the PC is given in Table 1. Fig. 1. Particle size distribution of PC, FA and LP.
152 M. Sahmaran, I.O. Yaman / Construction and Building Materials 21 (2007) 150–156

Fig. 2. Steel fibers used for the production of SCC.

Table 2 seen from the gradation of the aggregates presented in

Aggregate grading Table 2, the maximum aggregate size was 19 mm. Both
Sieve size (mm) % Passing the coarse and fine aggregate had a specific gravity of
Fine Coarse 2.70, and water absorptions of 0.5% and 1.2%, respec-
19 100 100
tively.
12.7 100 58.6
9.5 100 35.8 2.6. Chemical admixtures
4.75 96.9 0
2.36 85.5 – A polycarboxylic type superplasticizer (SP) was used
1.18 68.3 –
0.6 42.3 –
in all concrete mixtures. In addition to the SP, a viscos-
0.3 17.4 – ity modifying admixture (VMA) was also used. The
0.15 3.7 – properties of both admixtures, as provided by their
manufacturers, are shown in Table 3.

SF1 type had a tensile strength and a specific gravity of

1100 MPa and 7.85, and SF2 type had a tensile strength
and a specific gravity of 2000 MPa and 7.17, respec-
tively. The length and aspect ratio of the SF1 was
30 mm and 55, respectively, compared to 6 mm and
37.5 for SF2. The SF2 type was made of high strength
steel with a brass coating providing the fiber a relatively
smooth surface. For all the mixtures, the total fiber
content was kept constant at 60 kg/m3.
2.5. Aggregates
As for the aggregates, crushed limestone and crushed
sand from the same local source were used. As can be
3. Experimental procedures
3.1. Mixture proportions
Five concrete mixtures, with the proportions pro-
vided in Table 4, are prepared. The two control mixtures
did not contain any steel fibers. As a binder one of the
control mixture included PC (Control_PC) and the
other one had FA replacing 50%, by weight, of PC
(Control_FA). All of the remaining mixtures had the
same amount of FA as in Control_FA. These were
named as FA_SF1, FA_SF1&SF2, and FA_SF2 indi-
cating the type of steel fiber incorporated in the mixture.

Table 3
Properties of chemical admixtures
Chemical admixture Specific gravity pH Solid content (%) Main component
SP 1.08 5.7 40 Polycarboxylic ether
VMA 1.00 5.8 20 Aqueous dispersion of microscopic silica
 M. Sahmaran, I.O. Yaman / Construction and Building Materials 21 (2007) 150–156
Table 4
Mixture proportions
Mixture ID Ingredient (kg/m3)
Water PC FA
Control_PC 250 500 0 70
Control_FA 230 250 250
FA_SF1 226 250 250
FA_SF1&SF2 205 250 250
FA_SF2 205 250 250
For all the mixtures, the total amount of binder
(PC + FA), the amount of chemical admixtures, and
the amount of LP were all kept constant. Water was
added to the mixture until the SCC characteristics were
observed; therefore, the water/powder ratio (w/p) was
not kept constant and was observed to change between
0.35 and 0.44.
3.2. Preparation and casting of test specimens
The mixtures were prepared at about 5 min with a
70-liter rotating planetary mixer. The sand, coarse
aggregate and fibers were first dry-mixed followed by
the addition of fine materials and 1/3 of water. Finally,
water and chemical admixtures were pre-mixed and
added to the mixture.
After the mixing procedure was completed, tests were
conducted on the fresh concrete to determine slump flow
time and diameter, V-funnel flow time, and air content.
Segregation and bleeding were visually checked during
the slump flow test and was not observed in any of the
mixtures. From each concrete mixture, six 150-mm
cubes and six cylinders 100 mm in diameter and 200
mm high were cast. All specimens were cast in one layer
without any compaction. The cubes were used for the
compressive strength and ultrasonic pulse velocity tests
and the cylinders were used for the splitting tensile
strength tests. After demolding, all specimens were
stored in a curing room at 21 ± 2 °C, and 95 ± 5%
relative humidity until testing.
3.3. Tests on fresh concrete
Deformability and viscosity of fresh concrete is eva-
luated through the measurement of slump flow time
and diameter, and V-funnel flow time (Fig. 3). The
slump flow is used to assess the horizontal free flow
(deformability) of SCC in the absence of obstructions.
The procedure for the slump flow test and the com-
monly used slump test are almost identical. In the slump
test, the change in height between the cone and the
spread concrete is measured, whereas in the slump flow
test the diameter of the spread concrete is determined as
153
LP Aggregate Steel fiber SP VMA
Fine Coarse SF1 SF2
905 539 0 0 5.5 6.25
70 888 529 0 0 5.5 6.25
70 889 530 60 0 5.5 6.25
70 925 550 30 30 5.5 6.25
70 924 549 0 60 5.5 6.25
the slump flow diameter (D). According to Nagataki
and Fujiwara [12], a slump flow diameter ranging from
500 to 700 mm is considered as the slump required for
a concrete to be classified as SCC. According to Specifi-
cation and Guidelines for SCC prepared by EFNARC
[13] (European Federation of National Trade Associa-
tions), a slump flow diameter ranging from 650 to
800 mm can be accepted for SCC. In the slump flow test
concreteÕs ability to flow and its segregation resistance
can also be measured. To measure these properties,
the time it takes for the concrete to reach a 500-mm
spread circle and any segregation border between the
aggregates and mortar around the edge of spread are
recorded. EFNARC [13] suggests a slump flow time
(t500) of 2–5 s for a satisfactory SCC.
In addition to the slump flow test, V-funnel test is
also performed to assess the flowability and stability of
the SCC. The funnel is filled completely with concrete
and the bottom outlet is opened, allowing the concrete
to flow. The V-funnel flow time is the elapsed time
(tV-f) in seconds between the opening of the bottom
outlet and the time when the light becomes visible from
the bottom, when observed from the top. Good flowable
and stable concrete would consume short time to flow
out. According to Khayat [14], a tV-f which is less than
6 sec. is recommended for a concrete to qualify as a
SCC. According to EFNARC [13], tV-f ranging from 6
to 12 s is considered adequate for a SCC.
3.4. Tests on hardened concrete
Tests performed on cured concrete specimens consist
of the specimen compressive strength, the splitting ten-
sile strength, and the ultrasonic pulse velocity. For each
mixture, cubic specimens were loaded under compres-
sive load to failure (ultimate load) at 28 and 56 d. The
compressive strength was computed from the average
of three specimens. The ultrasonic pulse velocities
(UPV) of all six cubic specimens were measured on the
two smooth sides of the specimen at 7, 14, 28, and 56
d. The UPV test was conducted with direct transducer
arrangement using a pair of narrowband 54 kHz trans-
ducers with a commercially available PUNDIT system.
154 M. Sahmaran, I.O. Yaman / Construction and Building Materials 21 (2007) 150–156

Fig. 3. Workability tests on the HVFA–SCC.

4. Results and discussion
4.1. Fresh concrete properties
The results of fresh concrete tests are shown in
Table 5, which included the w/p of the mixture, slump
flow diameter and time, V-funnel flow time and air
content. As seen in that table, the slump flow diameters
of all mixtures were in the range of 560–700 mm. The
slump flow times were less than 2.9 s, and the V-funnel
flow times (tV-f) were in the range of 2.4–4.3 s. There-
fore, all concrete mixtures were considered as SCC.
In all of the SCC mixtures, there was no segregation
of aggregate near the edges of the spread-out concrete
as observed from the slump flow test.
Also observed in Table 5, is the change in w/p for the
same workability measure, i.e. the same D, t500, and tV-f.
The Control_PC mixture had the highest w/p, but as part
of the PC was replaced by FA the w/p of all mixtures
decreased. This phenomenon was also observed by other
researchers [7,15]. In such studies, even though finer FAs
were used, which is expected to increase the water
requirement of a concrete mixture, the smooth surface
characteristics and spherical shape of the FA improved
the workability characteristics of concrete mixtures,
and similar workability properties was achieved by a
smaller w/p. Therefore, using a coarser FA with higher
volumes naturally decreased the water demand of a
SCC mixture for similar workability measures.
The steel fibers also affected the fresh properties of
the concrete mixtures. The addition of SF1 type steel fi-
bers did not affect the water requirement of the mixture
for the same workability. However, addition of SF2 type
fibers which had smaller diameters and sizes reduced the
amount of water. This could be explained by the geo-
metry of the fibers as well as the surface characteristics
of these fibers. SF2 fibers had smaller dimensions when
compared to SF1 fibers, thus had less potential to
prevent the movement of aggregates. In addition, SF2
fibers were coated with brass and had smooth surface
characteristics as compared to SF1 fibers. Therefore,
these characteristics of SF2 type of fibers reduced the
energy loss during the movement of concrete ingredients.
4.2. Hardened concrete properties
The results of hardened concrete tests are presented in
Table 6, which included the 28 and 56 d compressive and
splitting tensile strength tests in addition to 7, 14, 28, and
56 d ultrasonic pulse velocity tests. To evaluate the effect
of fibers on the hardened properties, the 56-d properties
are normalized with respect to the Control_FA mixture

Table 5
Fresh properties
Mixture ID w/pa Fiber factor Slump flow V-Funnel tV-f (s) Air content (%)
D (mm) t500 (s)
Control_PC 0.44 0 560 2.9 4.3 2.5
Control_FA 0.40 0 690 <2.0 2.4 1.8
FA_SF1 0.40 42 660 <2.0 2.8 4.2
FA_SF1&SF2 0.36 35 630 <2.0 4.0 4.5
FA_SF2 0.36 29 700 <2.0 2.7 3.6
a
p (powder) = PC + FA + LP.
 M. Sahmaran, I.O. Yaman / Construction and Building Materials 21 (2007) 150–156 155

Table 6
Hardened properties
Mixture ID Compressive strength Splitting tensile strength Ultrasonic pulse velocity (m/s)
(MPa) (MPa)
28 da 56 da 28 da 56 da 7 db 14 db 28 db 56 da
Control_PC 40.7 (0.5)c 41.7(0.4) 3.58 (0.3) 3.68 (0.1) 4562 (22) 4570 (45) 4578 (35) 4609 (7)
Control_FA 23.3 (1.0) 28.6 (1.2) 2.82 (0.4) 3.34 (0.5) 4161 (31) 4260 (51) 4436 (74) 4564 (29)
FA_SF1 19.6 (0.1) 24.5 (0.6) 3.10 (0.2) 3.69 (0.6) 3963 (10) 4007 (44) 4157 (45) 4317 (28)
FA_SF1&SF2 22.8 (0.5) 26.1 (2.0) 3.40 (0.0) 3.82 (0.3) 3970 (42) 4100 (41) 4249 (40) 4383 (88)
FA_SF2 22.5 (2.9) 31.8 (0.8) 3.08 (0.1) 3.23 (0.2) 4142 (109) 4224 (109) 4359 (82) 4506 (114)
a
Tests are performed on three specimens.
b
Tests are performed on six specimens.
c
Numbers in parenthesis are the standard deviations.

and the results are presented in Fig. 4. As seen from that 35

figure, the most significant changes were observed on the 56 day
2
30 28 day R = 0.89
compressive strength and later on splitting tensile

Compressive Strength (MPa)

strength. The effects of high-volume FA replacement 25
and fiber inclusion on the ultrasonic pulse velocity 2
R = 0.70
(UPV) seemed to be negligible. 20

Substitution of PC with a coarse FA resulted in lower 15

compressive and splitting tensile strengths both at 28
and 56 d, even though the w/p of the mixture was 10
reduced. The reduction in compressive strength was
5
43% at 28 d and 31% at 56 d. This reduction could be
attributed to the low pozzolanic activity of the coarse 0
FA. Fiber inclusion did not significantly affect the mea- 0.0 0.2 0.4 0.6 0.8 1.0
Vf, SF2 / Vf, SF1+SF2
sured mechanical properties. However, as seen in Fig. 5,
as the volumetric ratio of the SF2 type fibers increased Fig. 5. Effect of steel fibers on the compressive strength.
the compressive strength increased. This was due to
the relatively small dimensions of SF2 type fibers, which
gave these fibers the ability to delay the micro-crack 4.5
formation and to arrest and prevent their propagation 4.0
afterwards up to a certain extent. Another explanation
Splitting Tensile Strength (MPa)

3.5
to the increase in the compressive strength could be
the decrease in w/p which decreased as the amount of 3.0

SF2 type fibers increased. However, when the splitting 2.5

tensile strengths were examined (Fig. 6), it could be seen 2.0
that there was a reduction in the splitting tensile
1.5
strengths as the volume of SF2 type fibers were increa- 56 day
sed or the w/p decreased. The reduction in the splitting 1.0 28 day

tensile strength was explained by the loss of the presence 0.5

0.0
0.0 0.2 0.4 0.6 0.8 1.0
Vf, SF2 / Vf, SF1+SF2

Fig. 6. Effect of steel fibers on the splitting tensile strength.

of longer SF1 fibers that also had hooked ends. There-

Fig. 4. 56-d hardened properties of the HVFA–SCC.
fore, SF1 fibers were mainly responsible for the increase
in tensile strengths.
As presented in Fig. 4, the UPV of the SCC mixtures
seemed to be unaffected by the steel fiber inclusion.
However, the UPV could be used to assess the harden-
ing of the SCC mixtures. It was clear from Fig. 7 that
as hydration continued, the UPVs increased for all the
156 M. Sahmaran, I.O. Yaman / Construction and Building Materials 21 (2007) 150–156
4700
4600
4500 2
R = 0.93
4400
UPV (m/s)
4300
4200
R = 0.98
4100
4000
3900
0 10 20 30
Test Age (days)
Fig. 7. Strength gain of SCC mixtures.
SCC mixtures. However, the slope of that curve was
quite different for the PC and FA mixtures. For the
Control–PC mixture the slope was much smaller as most
of the hydration was complete by 7 d. However, for the
mixtures with FA the hydration reactions continued
after 7 d indicating a higher slope.
5. Conclusion
In this paper, it was observed that, incorporation of
HVFA reduced the water requirement of a SCC mix-
ture. In other words, using high-volumes of a coarse
and substandard FA increased the workability charac-
teristics of SCC mixtures. On the other hand, using
the coarse FA caused significant strength losses to the
SCC mixtures, as they were used in high-volumes. The
compressive strength reduction due to low pozzolanic
activity of the FA was partially off-set by the use of
smaller sized SF2 type steel fibers. As for the splitting
tensile strength, the longer fibers with hooked ends, i.e.
SF1 fibers were more effective in characterizing the
tensile strength of concrete. Therefore, it could be
concluded that the fiber geometry affected the properties
of SCC mixtures not only in the hardened state but also
in the fresh state.
References
2
R = 0.99
[1] Hannant DJ. Fiber cements and fiber concrete. Chiches-
ter: Wiley; 1987.
2
R = 0.97 [2] Bentur A, Mindess S. Fiber reinforced cementitious compos-
ites. London: Elsevier; 1990.
2
R = 0.92 [3] Mobasher B, Li CY. Mechanical properties of hybrid cement-
based composites. ACI Mater J 1996;93(3):284–92.
2 Control_PC
[4] Okamura H, Ouchi M. Self-compacting concrete: development,
Control_FA
present use and future. In: Skarendahl A, Petersson O, editors.
Proceedings of the 1st international RILEM symposium on self-
FA_SF2
compacting concrete; 1999. p. 3–14.
FA_SF1& SF2
[5] Bilodeau A, Sivasundaram V, Painter KE, Malhotra VM.
FA_SF1
Durability of concrete incorporating high volumes of fly ash
40 50 60 from sources in US. ACI Mater J 1994;91:3–12.
[6] Bilodeau A, Malhotra VM. High-volume fly ash system: concrete
solution for sustainable development. ACI Mater J 2000;97:41–8.
[7] Bouzoubaâ N, Zhang MH, Malhotra VM. Mechanical properties
and durability of concrete made with high-volume fly ash blended
cements using a coarse fly ash. Cement Concrete Res
2001;31:1393–402.
[8] Bouzoubaa N, Laclemi M. Self-compacting concrete incorporat-
ing high volumes of class F fly ash preliminary results. Cement
Concrete Res 2001;31(3):413–20.
[9] S
ß ahmaran M, Yaman IO, Tokyay M. Self-compacting concrete
with high volume fly ash content and new generation of
superplasticizer. In: Proceedings of the concrete 2005 national
symposium; 2004. p. 225–33 [in Turkish].
[10] Khayat KH, Roussel Y. Testing and performance of fiber-
reinforced, self-consolidating concrete. In: Skarendahl A, Peters-
son O, editors. Proceedings of the 1st international RILEM
symposium on self-compacting concrete; 1999. p. 509–21.
[11] Grouth P, Nemegeer D. The use of steel fibres in self-compacting
concrete. In: Skarendahl A, Petersson O, editors. Proceedings of
the 1st international RILEM symposium on self-compacting
concrete; 1999. p. 497–507.
[12] Nagataki S, Fujiwara H. In: Malhotra VM, editor. Self-compact-
ing property of highly flowable concrete, vol. SP 154. American
Concrete Institute; 1995. p. 301–14.
[13] EFNARC. Specification and guidelines for self-compacting con-
crete, English edition. European Federation for Specialist Con-
struction Chemicals and Concrete Systems. Norfolk, UK;
February 2002.
[14] Khayat KH, Guizani Z. Use of viscosity-modifying admixture to
enhance stability of fluid concrete. ACI Mater J 1997;94(4):
332–41.
[15] Yahia A, Tanimura M, Shimabukuro A, Shimoyama Y. Effect of
rheological parameters on self-compactability of concrete con-
taining various mineral admixtures. In: Skarendahl A, Petersson,
O editors. Proceedings of the 1st international RILEM sympo-
sium on self-compacting concrete; 1999. p. 523–35.