Materials and Structures (2006) 39:683–693

DOI 10.1617/s11527-006-9102-3

ORIGINAL ARTICLE

About the influence of the content and composition of the

aggregates on the rheological behaviour of self-compacting
concrete
Hans W. Reinhardt · Timo Wüstholz

Received: 7 June 2005 / Accepted: 22 August 2005


C RILEM 2006

Abstract Within the scope of a parameter study the contrôlant la consistance (volume de la pâte pure de ci-
influence of the mixture composition of different self- ment, du mortier et des gros granulats) ont pu être con-
compacting concretes on the fresh concrete proper- vertis en un paramètre épaisseur de pâte. Ce paramètre
ties was investigated. For this purpose the standard permet de décrire les valeurs caractéristiques des tests
test methods as well as the fresh concrete rheome- standards – comme par exemple l’essai d’étalement et
ter “BTRHEOM” were used. The concrete was mod- l’essai d’écoulement – et les paramètres rhéologiques
elled as a two-phase system, consisting of the fluid fondamentaux que sont la limite d’écoulement et la vis-
phase paste and the solid phase aggregates. The consis- cosité plastique.
tency control parameters paste volume, mortar volume
and the coarse aggregate volume could be transferred
into the model parameter thickness of excess paste. By
1. Introduction and object
means of this model parameter the characteristic values
of the standard test methods like slump flow test and
The flow behaviour (rheological behaviour) of self-
V-funnel test as well as the fundamental rheological pa-
compacting concrete is mostly characterized by test
rameters yield stress and the plastic viscosity could be
methods like the slump flow test and the V-funnel test
described. A comparative study showed that the yield
etc. The equipment is inexpensive and the tests are easy
stress and the plastic viscosity of self-compacting con-
in handling and therefore suitable for use on site. But
crete can be estimated based on the characteristic values
they have one disadvantage: they provide no fundamen-
of the slump flow test.
tal physical flow parameters. Fundamental parameters
can only be derived from so-called flow curves which
Résumé Dans le cadre d’une étude paramétrique,
are obtained by means of viscometers or rheometers.
l’influence de la composition sur les propriétés à l’état
A flow curve of a fluid describes the relationship be-
frais des bétons autoplaçants a été analysée. A cet effet,
tween the shear stress τ and the shear rate γ̇ . Many
les essais standardisés et le rheomètre “BTRHEOM”
fluids like water or oil behave like a Newtonian fluid
etaient utilisées. Le béton a été modélisé par un système
where the shear stress is directly proportional to the
bi-phase composé de la phase fluide “pâte pure” et
shear rate. This relationship is expressed by Equation
de la phase solide “granulats”. Ainsi les paramètres
(1). The characteristic value of such a Newtonian fluid
is the viscosity η
H. W. Reinhardt · T. Wüstholz
Construction Materials Institute, University of Stuttgart,
Germany τ = η · γ̇ (1)
684 Materials and Structures (2006) 39:683–693

In contrary to a Newtonian fluid fresh ordinary con-
crete and also self-compacting concrete starts to flow
only if the shear stress exceeds the yield stress τ 0 . In
the simplest way fresh concrete can be compared with a
Bingham fluid (Equation (2)). In such a case two char-
acteristic parameters are necessary to describe the flow
curve. These are the yield stress τ0 and the plastic vis-
cosity η pl
τ = τ0 + η pl · γ̇
In the literature papers are dealing mostly with the
modelling of the rheological parameters τ0 and η pl . In
this article a simple model approach is presented which
is also applied to the standard consistency parameters
of the slump flow test and the V-funnel test. In this
approach the self-compacting concrete is considered
as a two phase system which consists of the fluid phase
paste and the solid phase aggregates.
2. Scope of investigation
2.1. Varied parameters
Three self-compacting concrete mixtures (Table 1)
were the backbone of a test programme in which the
influence of the concrete composition on the rheolog-
ical behaviour was investigated. The main differences
between these three reference concretes were the type
of filler (limestone powder LS respective fly ash
FA), the equivalent water-cement ratio (w/c)eq and the
particle-size distribution of the aggregates (Fig. 1a).
These reference concretes are identified by the filler
type followed by “A” or “B”. Table 2 gives a survey of
the parameter variations. Based on the reference con-
cretes LS (A) and FA (A) the paste volume VPaste , mor-
tar volume V M and also the coarse aggregate volume
(2) V A>8 was changed. The change of the coarse aggregate
volume V A>8 was controlled by the mass ratio m A,4/8 :
m A,8/16 whereas the total volume of aggregates, paste
and mortar were held constant. To investigate how the
particle-size distribution of the aggregates influences
the flow behaviour of SCC a set of different particle-
size distributions was chosen (Fig. 1b) which are lying
within the range of the standardized particle-size dis-
tributions A16 and C16 according to DIN 1045-2 Ap-
pendix L [1]. For this variation the reference concrete
FA (B) has built the base (Table 1). All types of vari-
ations are given in Table 2, whereas the italic printed
numbers indicate the reference concrete mixtures. The
paste volume VPaste consists of the volume parts of wa-
ter, cement, filler and admixtures and includes an air
content of 15 l per m3 concrete. Unless otherwise noted
aggregates consisted of rounded river sand and gravel.
All concretes were produced with a maximum aggre-
gate diameter of 16 mm.

Table 1 Mixture proportions of the reference concretes. Filler types: limestone powder (LS) and fly ash (FA)

FA (A) FA (B)
LS (A) Combination Combination-
Component Powder type type type

Strength class C30/37 C45/55 C45/55

Cement content (CEM II/A-LL 32.5R) mC [kg/m3 ] 239 345 311
Equivalent water-cement ratio (w/c)eq 1 [−] 0.70 0.43 0.45
Superplasticizer m S P /mC [% by mass of cement] 1.00 1.05 1.25
(Woermann FM/BV 375, type: PCE)
Total water content mw [kg/m3 ] 166.1 169.6 158.0
(inclusive water from superplasticizer)
Viscosity agent (Woermann UW Compound) mV A /mC [% by mass of cement] 0 0.10 0.10
Mass of filler (limestone LS or fly ash FA) mF [kg/m3 ] 337 194 175
Aggregates (round, river sand and gravel) mA [kg/m3 ] 1600 1604 1686
Powder content (cement and filler) mP [kg/m3 ] 576 540 486
Water-powder ratio (by volume) Vw /V P [−] 0.82 0.86 0.89
Paste volume (inclusive 15 litres of air) VPaste [l/m3 ] 385 383 352
Mortar volume (paste and aggregates < 4 mm) VM [l/m3 ] 670 709 648
mW
1
(w/c)eq = m C +0.4·m FA
according to DIN 1045-2 [1]
Materials and Structures (2006) 39:683–693 685

Table 2 Parameter variations. Reference concretes are indicated by Italic numbers

Type of
Variation filler Nr. 1 2 3 4 5

Paste volume VPaste [l/m3 ] LS (A) 335 366 385 396 –

FA (A) 340 370 383 392 –
Mortar volume V M [l/m3 ] LS (A) 604 653 670 704 –
FA (A) 638 691 709 730 –
Coarse aggregate volume V A>8 [l/m3 ] LS (A) 294 (0:100) 185 (39:61) 155 (50:50) 17 (100:0) –
(mass ratio) (m A,4/8 :m A,8/16 )
FA (A) 210 (20:80) 160 (41:59) 137 (50:50) 15 (100:0) –
Grading curve acc. to DIN and k [−] FA (B) A 16 4.54 AAB16 4.33 AB16 4.12 B16 3.65 C16 2.81
k value
(VPaste = constant)

100 per sieve, whereas the nine sieves from 0.25 mm up to

LS (A)
90 63 mm are incorporated. The k-value increases with
FA (A)
Passing [% by mass]

80 FA (B) an increasing maximum aggregate diameter and an in-

70 creasing coarse aggregate content from C16 to A16
60
(Table 2).
50
B16
40
30 A16

9
p A,i
k= 1− (3)
20 100
i=1
10
0
0.09 0.13 0.25 0.5 1 2 4 8 16 31.5
Particle size [mm] 2.2. Rheological investigations
(a)

100 Additionally to the standardized tests like slump flow

A16
90
AAB16
and V-funnel (dimensions of devices see Fig. 2a) rhe-
Passing [% by mass]

80 AB 16, reference ological measurements were performed with a con-

70 B16 crete rheometer “BTRHEOM” which was developed
60 C16
at LCPC (Laboratoire Central des Ponts et Chaussées
50
40 Paris, Fig. 2b).
30 The BTRHEOM measures the torque which is nec-
20 essary to shear a concrete sample at a defined number of
10 revolutions. From the measured torque  and the num-
0
ber of revolutions n it is possible to derive a -n-curve
0.09 0.13 0.25 0.5 1 2 4 8 16 31.5
by a regression analysis (Fig. 3). Such a -n-curve can
Particle size [mm]
(b) then transferred into a τ -γ̇ -curve (e.g. Equation (2)). In
the simplest case (Bingham) the measured data can be
Fig. 1 Particle-size distributions of the aggregates: (a) Reference
concretes. A16 and B16 are standing for the standard particle- fitted by a straight line (Equation (4)). From the torque
size distributions according to DIN 1045-2 Appendix L [1] (b)  0,B respective the slope of the straight line A B the
Variations for the concretes including FA (B) yield stress τ 0 respective the plastic viscosity η pl can
be calculated. If this linear approach is applied, often
negative yield stresses – especially for self-compacting
The particle-size distributions of the aggregates concrete – are calculated. But negative yield stresses are
were characterized by the so-called k-value (Equation physicallly not possible. These negative yield stresses
(3)), Table 2. Herein stands p A for the passed aggregates have also been described by de Larrard et al. [2] and
686 Materials and Structures (2006) 39:683–693

By means of the application of the so-called Herschel-

Bulkley approach (Equation (5)) negative yield stresses
can be avoided.

 = 0,HB + AHB · n b (5)

However, it implies the disadvantage of three curve

parameters, whereas only the yield stress can physi-
cally be interpreted. Therefore a combined method was
chosen [2]. This method is called modified Bingham
method. The rheological parameters presented in this
article were obtained by means of this modified Bing-
ham method. The -n- respective the τ -γ̇ -relationships
are given with Equations (6) and (7).

 = 0,HB + A B  · n (6)
τ = τ0,HB + ηpl,HB · γ̇ (7)

3. Results

3.1. Modelling of the rheological behaviour

of SCC

Self-compacting concrete can be considered as a con-

Fig. 2 (a) Dimensions of the slump cone and V-funnel (b)
Photography of the “BTRHEOM”, inner diameter of the con-
tainer D = 240 mm, height of the sheared concrete sample
h = 100 mm
were confirmed by measurements at the Department of
Construction Materials.
 = 0,B + A B · n (4)
centrated supsension in that the solids are dispersed in
the fluid phase water, de Larrard [3] and Ferraris et al.
[4]. Others, Nielsen [5] and Geiker et al. [6], consider
self-compacting concrete as a two phase system which
consists of the Bingham-phase mortar in that the coarse
aggregates are dispersed. The models of these authors
are based on the so-called relative solid concentration
φ/φmax .
Generally, the relative viscosity of a suspension, i.e.
the ratio between the viscosity of a suspension and the
viscosity of the fluid phase, can be traced back to the
so-called relative solid concentration of the particles
φ/φmax . Whereas φ stands for the volume concentration
of the solids in a suspension, φmax represents the maxi-
mum volume concentration of the particles in the state
of the maxium packing and is called maximum packing
fraction [7]. The maximum packing fraction φmax , be-
ing controlled by the type of packing, is very sensitve
to particle-size distribution and particle shape [7]. In
this context the models for concrentrated suspension
of Mooney [8] and Krieger and Dougherty [9] have to
be mentioned which are both based on the well-kown
Materials and Structures (2006) 39:683–693 687

Fig. 3 Examples of 5

Torque [Nm]
different curve fittings
4 AB' (modified Bingham )

3
measured data

2 Bingham

Herschel-Bulkley
1 AB (Bingham) modified Bingham
Γ 0,HB
0
Γ 0,B
-1
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Number of revolutions n [1/s]

Fig. 4 Model of SCC

model of Einstein [10] that is only valid for dilute dis-
persed suspensions [7].
The model presented in this article assumes that SCC
consists of the fluid phase paste and the solid phase ag-
gregates, this stands in contrast to models of [3–6].
The solid-fluid interactions respective the solid-solid
interactions are controlled by the so-called excess paste
thickness tPaste,ex which is equivalent to the half distance
of two neighbouring aggregates if the distance is as-
sumed to be independent of the aggregate size (Fig. 4).
The excess paste thickness tPaste,ex can be determined
if the excess paste volume VPaste,ex , that remains after
the voids between the aggregates VA,void are filled with
paste, is constantly layered on the surface of the aggre-
gates. Another approach with a particle-size dependent
paste thickness is described by Oh et al. [12]. Krell [11]
showed that the consistency of ordinary concrete can
be expressed as a function of the thickness of excess
paste.
The excess paste volume VPaste,ex can be calculated
with Equation (8) where m A respective ρ A stands for
the mass respective the density of the aggregates. Fur-
ther the loose bulk densityρA,bulk of the aggregates is
required.
VPaste,ex
 
mA ρA
= VPaste − VA,void = VPaste − −1
ρ A ρ A,bulk
(8)
In this study the loose bulk density was calculated
using a 10 l container according to DIN EN 1097-3
[13]. For this purpose a mass of about 20 kg of dry
aggregates 0/16 was first homogenized in the concrete
mixer and then filled into the container without any
additional compaction as described in [13]. To calcu-
late the excess paste thickness the surface area of the
aggregates is required. The aggregates were thereby
idealized as spheres. The particle-size distribution of
the aggregates is divided into 9 classes according to
the sieves of DIN 1045-2 Appendix L [1], see also
Table 3. Each class i is represented by the mean par-
ticle diameter di between two subsequent sieves. Ag-
gregate particles with a diameter <0.125 mm are ig-
nored. However, they are not part of the paste volume
688 Materials and Structures (2006) 39:683–693

Table 3 An example of the calculation of the aggregate surface area per m3 concrete, aggregate particles are considered to be spheres

Particle-size distribution Calculation of the surface area S A

Diameter Represantative Number ni of Surface area Part of

(sieve Passed particle diameter particles per m3 S A.i (per m3 the total
opening) agregates Class di Fraction concrete concrete) surface area
[mm] [% by mass] [mm] [mm] [% by mass] [1/m3 ] [cm2 /m3 ] [%]

0.125 0.86 0.09/0.125 0.1075 0.75 – – –

0.25 3.72 0.125/0.25 0.1875 2.86 5.096E+09 5659618.7 18.7
0.5 18.42 0.25/0.5 0.375 14.70 3.274E+09 14544824.3 48.2
1 27.63 0.5/1 0.75 9.21 2.564E+08 4556388.9 15.1
2 33.25 1/2 1.5 5.62 1.956E+07 1390168.6 4.6
4 47.56 2/4 3.0 14.31 6.225E+06 1769867.7 5.9
8 69.95 4/8 6.0 22.39 1.218E+06 1384603.0 4.6
16 97.56 8/16 12.0 27.61 1.877E+05 853704.5 2.8
31.5 100.00 16/31.5 23.75 2.44 2.139E+03 38119.6 0.1
100.00 S A = 30197295.3 100.0

VPaste in Equation (8), since the powder content of the
aggregates is indirectly incorporated in the volume of
voids VA,void that is derived from the loose bulk density
test.
The results in Table 3 show that the particle fraction
0.25/0.5 mm is dominating with about 48% of the total
surface area. This is an evidence for the importance of
the fine sand fraction with regard to the workability of
concrete respective the flow behaviour of SCC.
For a comparison of different particle-size distribu-
tions of aggregates the specific surface area SA,spec can
be calculated as the ratio between the surface area S A
related to the mass of the aggregates m A (Equation (9)):
SA
S A,spec =
mA
If a spherical paste layer on the aggregates is as-
sumed (Fig. 4), the excess paste thickness tPaste,ex can
be calculated using Equation (10), whereas ri is the
radius of a particle of class i, ni are the number of par-
ticles per class and tPaste,ex is the thickness of the paste
layer.
4    distributed. The calculated excess paste thickness is –
VPaste,ex = π n i (ri + tPaste,ex )3 − ri3
3 i
In a simplified way, the excess paste thickness can
directly calculated with Equation (11). This assumes
that the thickness of the paste layer is small compared
to particle size, because Equation (11) is only valid for
a planar paste layer.
VPaste,ex
tPaste,ex,plan = (11)
SA
For small paste thicknesses both methods coincide.
The deviation between both methods is increasing with
an increasing paste thickness (Fig. 5).
It is possible that the calculated excess paste thick-
ness lies below the median particle diameter1 of the
powder which was determined about 10 μm for all in-
vestigated SCC mixtures. Two reasons are responsi-
ble for this: (1) The void volume was determined in
an uncompacted state by filling the aggregates in a
(9) container according to to DIN EN 1097-3 [13]. (2)
The excess paste thickness was assumed to be inde-
pendent of the aggregate particle size (Fig. 4). Based
on a two-dimensional imagination the thickness of the
excess paste should be greater than half of the max-
imum diameter of the particles in the paste. But in
reality a three-dimensional aggregate skeleton exists
with sufficiently sized voids between the aggregates.
Furthermore, also the particles of the paste as well
as the voids between the aggregates are contiuously
(10)
despite of a partially small value – a suitable param-
eter for the evaluation of different aggregate particle-
size distributions, this will be shown later. The excess
1
50% (by volume) of the paste particles are below this diameter
Materials and Structures (2006) 39:683–693 689

0.060
0.055
LS (A), paste volume
0.050 LS (A), mortar volume
Excess paste thickness

0.045 LS (A), coarse aggregate volume

tPaste,ex,spher [mm]

FA (A), paste volume

0.040
FA (A), mortar volume
0.035 SFA (A), coarse aggregate volume
0.910
0.030 y = 0.634x FA (B), particle-size distribution between A16 and C16
2
0.025
R = 0.999

0.020
0.015
0.010
particle-size distribution C16 (VPaste = 352 litres/m)
0.005
0.000
0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040 0.045 0.050 0.055 0.060

Excess paste thickness tPaste,ex,plan [mm]

Fig. 5 Comparison of both methods for calculation of the excess paste thickness after Equation (10) and Equation (11)

800 B16 0.024

35 AAB16
FA (B) AB16
A16 700 0.021
FA (A)
Porosity ε (uncompacted) [%]

34 A16

Excess paste thickness

LS (A) 600 0.018
Slump flow sf [mm]

33

tPaste,ex [mm]
500 0.015
32 C16
400 0.012
31 AAB16 B16
300 0.009
C16
30 Particle-size FA (B)
distributions between 200 0.006
AB16 and B16 Slump flow
29 100 0.003
Excess paste thickness
28 0 0.000
10 14 18 22 26 30 34 38 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Specific surface area SA,spec [cm 2/g] k [-]
(a) (b)
Fig. 6 (a) Porosity versus specific surface area of poured aggregates with different particle-size distributions (b) Slump flow and excess
paste thickness versus the k-value

paste thickness, which is given in the following figures, Figure 6b shows the influence of the aggregate
was calculated after the spherical method of Equation particle-size distribution (expressed by the k-value,
(10). Table 2) on the slump flow and on the excess paste
It can be seen from Equations (8) and (11) that thickness for a constant paste volume of 352 l/m3 .
(for a constant paste volume) both, the void volume The maximum paste thickness was calculated for the
V A,void of the poured aggregates (respective the poros- mixture containing the particle-size distribution AB16.
ity ε = V A,void /(V A + V A,void )) and the specific surface This concrete reached also the maximum slump flow
area S A,spec are controlling the thickness of the paste value. It has to be mentioned that the origin of the sand
layer tPaste,ex . Hence, particle-size distributions of ag- fraction of the aggregates of the SCC mixtures pro-
gregates have to be preferred which possess a small duced with limestone powder LS (A) and fly ash FA
specific surface area and a small porosity. Both crite- (A) differed from those aggregates which were used for
ria are fulfilled by particle-size distributions which are the concretes including fly ash FA (B). This is probably
lying approximately in the range between AB16 and one reason for the differences in the porosity shown in
B16 (Fig. 6a). These particle-size distributions have of Fig. 6a.
a reduced coarse aggregate content, which reduces also Figure 7 shows the fresh concrete parameters yield
the risk of blocking. stress τ 0,HB , slump flow sf, plastic viscosity η pl,HB and
690 Materials and Structures (2006) 39:683–693

0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040 0.045 0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040 0.045
800 LS (A), paste volume 600
-0.67 LS (A), paste volume

Plastic viscosity ηpl,HB [Pa.s]

LS: y = 32.74x LS (A), mortar volume
700 2
R = 0.70 LS (A), coarse aggregate volume
LS (A), mortar volume
500
Yield stress τ0,HB [Pa]

FA (A), paste volume LS (A), coarse aggregate volume

600 FA (A), paste volume
FA (A), mortar volume
FA (A), coarse aggregate volume
400 FA (A), mortar volume
500
FA (A), coarse aggregate volume
400 300
300
200
LS und FA:
200 -1.230
-1.10 y = 1.892x
FA: y = 2.86x 100 2
100 2 R = 0.76
R = 0.55
0 0

800 60
700
50
Slump flow sf B [mm]

V-funnel time t V [s]

600
-1.742
40 LS: y = 0.033x
500 2
R = 0.758
400 2
LS: y = -464391x + 36492x + 24.854
30
2
300 R = 0.93
20
2
200 FA: y = -858516x + 46625x + 106.69 -1.034
2 10 FA: y = 0.317x
R = 0.93
100 2
R = 0.899

0 0
0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040 0.045 0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040 0.045
Excess paste thickness tPaste,ex [mm] Excess paste thickness tPaste,ex [mm]
(a) (b)
Fig. 7 Fresh concrete parameters as function of the excess paste thickness

the V-funnel flow time tV as function of the excess In the slump flow test the flow time t500 is normally
paste thickness tPaste,ex . The different symbols stand recorded to get an indication of the apparent viscosity of
for the variation of the paste volume, mortar volume a SCC mixture. The t500 -value is equivalent to the time
and the coarse aggregate volume (see Table 2). It can that is measured from the start of the upward move-
be seen from Fig. 7 that the excess paste thickness ment of the cone until the concrete spread has reached
is a suitable parameter to describe the consistency of a diameter of 500 mm. The relationship between the
SCC. The differences regarding the filler type and the flow time t500 and the model parameter excess paste
paste compositions (see Table 1) are leading to a differ- thickness (Fig. 8) can be described – independently of
ent flow behaviour. The given functions were derived the filler type – by one function. It will be shown in the
from a regression analysis. The deviations of the yield follwing chapter that the plastic viscosity can directly
stresses are greater than those of the plastic viscosities, be derived from the flow time t500 , even if additional
which can be traced back to the measurement proce- concrete mixtures are taken into account.
dure and the method of calculating the yield stresses by
an extrapolation. The higher yield stresses of the con- 3.2. Relationships between the fundamental
cretes made of limestone powder compared to the fly rheological parameters and the standard
ash concretes are also indicated by the reduced slump consistency parameters
flow values of the concretes containing limestone pow-
der (Fig. 7a). If greater deviations between model and It is interesting to know which of the fundamental rhe-
measured data are accepted it is sufficient to use only ological parameters can be derived from the standard
one function to describe the relationship between the consistency tests like slump flow test and V-funnel
plastic viscosity and the model parameter excess paste test. For this purpose the standard consistency param-
thickness (Fig. 7b). eters are plotted against the fundamental rheological
Materials and Structures (2006) 39:683–693 691

16 1200
LS (A)
LS (A), paste volume 1100 FA (A)
LS (A) + FA (A) CEM II/A-LL 32.5R
14 LS (A), mortar volume
1000 LS (A), Viscosity Agent Type

Slump flow sf [mm]

LS (A), coarse aggregate volume LS (B), crushed aggregates, CEM II/A-LL 42.5R
12 900 Model Eq. (12)
Flow time t 500 [s]

FA (A), paste volume

FA (A), mortar volume 800
10
FA (A), coarse aggregate volume 700
y = -0.44x + 809.56
8 600 R2 = 0.42
500
6
400
4 LS and FA:
-0.986 300
y = 0.143x
2
0 100 200 300 400 500 600 700 800 900
2 R = 0.85
Yield stress τ 0,HB [Pa]

0 (a)
0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040 0.045 80 LS (A)
*) CEM II/A-LL32.5R
Excess paste thickness tPaste,ex [mm] 70 **) CEM II/A-LL42.R FA (A)
*)
LS (A) + FA (A)
60

V-funnel time tV [s]

Fig. 8 Flow time t500 (slump flow test) as a function of the excess LS (A), Viscosity Agent Type

paste thickness tPaste,ex 50 LS (B), crushed aggregates **)

LS: y = 7.047e0.0057x
40
R2 = 0.61
30

20
FA: y = 6.61e0.0034x
10
parameters yield stress and plastic viscosity. Addi- R2 = 0.47

0
tional to the concretes presented in the previous chap- 0 50 100 150 200 250 300 350 400 450 500 550

ter (Table 2) other concrete mixtures were taken into Plastic viscosity η pl,HB [Pa.s]
(b)
account which are taken from a database of self-
compacting concrete mixtures existing at the Depart- 18
*) CEM II/A-LL32.5R
**) CEM II/A-LL42.5R
ment of Construction Materials. The main part of these 16
14
results were obtained in a parameter study in which
Flow time t 500 [s]

12 y = 2.67e0.0034x
the water content, the superplasticizer content and also 10 R2 = 0.58

the amount of viscosity agent were varied. Concretes 8 LS (A)

marked by LS (B) contained crushed aggregates. 6 FA (A)
*)
4 LS (A) + FA (A)
It is generally assumed that the slump flow depends LS (A), Viscosity Agent Type
2
mostly on the yield stress since the concrete deforms 0
LS (B), crushed aggregates **)

itself under its own weight until the shear stress reaches 0 50 100 150 200 250 300 350 400 450 500 550

the yield stress (Bingham material). However, the fric-
tion inside the concrete (viscous friction) and the fric-
tion between the concrete and the base plate leads to
an energy dissipation and has therefore an influence
on the measured slump flow value, see also [4]. The
slump flow is therefore also influenced by the plas-
tic viscosity. Figure 9a shows the slump flow plotted
against the yield stress for different concrete mixtures
without a distinction between different plastic viscosi-
ties. The slump flow tends to decrease with an increas-
ing yield stress. Theoretical considerations in Wüstholz
[14], that are based on formulas presented from Saak
et al. [15], led to Equation (12), in which the slump flow
sf [m] depends on the density of concrete ρ [kg/m3 ],
the acceleration of gravity g [m/s2 ] and the yield stress
τ0 [Pa]. It is assumed that the influence of the plastic
viscosity on the slump flow can be neglected compared
to the influences of the yield stress and density of the
Plastic viscosity η pl,HB [Pa.s]
(c)
Fig. 9 Relationships between the fundamental rheological pa-
rameters and the standard consistency parameters
concrete.

ρg
sf = · (0.1519 [m])3 (12)
τ0
Equation (12) is plotted for a concrete density of ρ =
2320 kg/m3 and g = 9.81 m/s2 in Fig. 9a. In fact there
is a large scatter. In spite of the large scatter and the fact
that Equation (12) was not adjusted to the test data it
corresponds to the lower limit of the slump flow values.
It is assumed and was already mentioned above that
the flow times measured in the standardized tests like
slump flow test and V-funnel test shall provide an in-
dication of the plastic viscosity of a concrete mixture.
692
An important difference between the standardized test
methods and rheometric test methods is that shear ve-
locities in the standardized tests depend on the geome-
try of the test equiupment and the plastic viscosity and
are not constant during the test as they are forced by
the dead weight of the concrete.
In Fig. 9b and c the V-funnel flow time tV and the
flow time t500 in the slump flow test are plotted against
the plastic viscosity η pl,HB . The concretes including
fly ash had approximately a plastic viscosity which
was twice as large compared to the concretes includ-
ing limestone powder despite the V-funnel flow time
was nearly equal. If the flow time t500 is considered, a
higher plastic viscosity corresponds with a longer flow
time, independently of the concrete composition. The
apparent viscosity of the paste of the fly ash mixtures
was clearly higher than those including limestone pow-
der. Probably blocking effects in the vicinity of the V-
funnel orifice were responsible for this behaviour, see
also [16]. It has also to be mentioned that the concrete
mixtures containing limestone powder had a slightly
increased coarse aggregate content compared to the fly
ash mixtures (Table 2). Hence, it is better to use the
flow time t500 to get an indication of the plastic vis-
cosity η pl,HB (Fig. 9c), because the t500 -value is not
influenced by blocking effects. However, the precision
of the measurement of the flow time t500 is – espe-
cially for low- viscous concretes with short flow times –
less precise than the measurement of the V-funnel flow
time tV . The marked concrete of the viscosity agent
type in Fig. 9b and c had relatively long flow times
tV and t500 despite a plastic viscosity of 210 Pa.s. This
behaviour can probably traced back to a small paste
volume (VPaste = 312 dm3 /m3 ) combined with a high
dosage of the viscosity agent.
4. Conclusions
The flow properties of different self-compacting con-
cretes including fly ash respective limestone powder
were investigated by means of standard tests and a fresh
concrete rheometer. It was showed that the influence of
the content and composition of the aggregates on the
flow properties can be described by means of the model
parameter excess paste thickness.
The flow parameters yield stress, slump flow and
V-funnel flow time depended – beside the model pa-
rameter excess paste thickness – also on the paste com-
Materials and Structures (2006) 39:683–693
position which was different for the concretes including
fly ash and limestone powder. The plastic viscosity and
the flow time t500 from the slump flow test could be de-
scribed independently of the filler type only as function
of the excess paste thickness.
Despite of the large scatter the slump flow tends to
decrease with an increasing yield stress. A clear rela-
tionship was found between the flow time t500 and the
plastic viscosity. It is therefore possible to estimate the
fundamental rheological parameters yield stress and
plastic viscosity based on the slump flow test. The V-
funnel flow time tV seems to be less suitable to distin-
guish between different plastic viscosities than the flow
time t500 , as the V-funnel flow time is also influenced
by blocking effects.
Acknowledgements The authors would like to thank the Gips-
Schüle-Foundation, which enabled the rheological investiga-
tions, for the financial support.
References
1. DIN 1045-2 Concrete, reinforced and prestressed concrete
structures. Part 2: Concrete – Specification, properties, pro-
duction and conformity (Application rules for DIN EN 206-
1), 2001-07
2. De Larrard F, Ferraris CF, Sedran T (1998) Fresh Concrete:
A Herschel-Bulkley Material. In: Materials and Structures,
31(211):494–498
3. De Larrard F (1999) Concrete mixture proportioning – A
Scientific approach (E & FN Spon, London)
4. Ferraris CF, de Larrard F (1998) Testing and modelling of
fresh concrete rheology. National Institute of Standards and
Technology, NIST-Report 6094
5. Nielsen LF (2001) Rheology of some fluid extreme compos-
ites. In: Reprint from the Nordic Concrete Research, 2:83–
93
6. Geiker MR, Brandl M, Thrane LN, Nielsen LF (2002) On
the effect of coarse aggregate fraction and shape on the rhe-
ological properties of self-compacting concrete. In: Cement,
Concrete and Aggregates 24(1):3–6
7. Barnes HA, Hutton JF, Walters K (1989) An introduction to
rheology (Elsevier Science Publishers B.V., Amsterdam)
8. Mooney M (1951) Journal of Colloid Science, 6:162–170
9. Krieger IM, Dougherty TJ (1959) Transactions of the Society
of Rheology, III:137–152
10. Einstein A (1911) Eine neue Bestimmung der
Moleküldimension. In: Annalen der Physik 19 (1906);
S. 289–306 und 34; S. 591–592
11. Krell J (1987) Einfluss der Feinststoffe im Beton auf die
Frischbetonkonsistenz. In: Berichte vom Fachkolloquium
“Zementleim, Frischmörtel, Frischbeton“, 1./2. Okt. in Han-
nover. Mitteilungen aus dem Institut für Baustoffkunde und
Materialprüfung der Universität Hannover, Heft 55, S.160–
176
Materials and Structures (2006) 39:683–693
12. Oh SG, Noguchi T, Tomosawa F (1999) Towards mix design
for rheology of self-compacting concrete. In: Skarendahl, Å;
Petersson, Ö. (eds). Proceedings of the First International
RILEM Symposium on Self-Compacting Concrete (Stock-
holm, September 13–14, 1999), Pro 7, S.A.R.L., Cachan, pp
361–372
13. DIN EN 1097-3: (1988-06) Tests for mechanical and physi-
cal properties of aggregates. Part 3: Determination of loose
bulk density and voids; German version, 1988–06
14. Wüstholz T (2005) Experimentelle und theoretische Unter-
suchungen der Frischbetoneigenschaften von Selbstverdich-
693
tendem Beton. Dissertation, Universität Stuttgart, Institut für
Werkstoffe im Bauwesen
15. Saak AW, Jennings HM, Shah SP (2004) A generalized ap-
proach for the determination of yield stress by slump and
slump flow. In: Cement and Concrete Research, 34(3):363–
371
16. Tangtermsirikul S, Khayat K (2001) Fresh concrete proper-
ties. In: Skarendahl, Å; Petersson, Ö. Self-Compacting Con-
crete. State-of-the-Art-Report of RILEM Technical Com-
mittee 174-SCC. RILEM Report 23, S.A.R.L, Cachan,
pp 14–19