Construction and Building Materials 84 (2015) 387–398

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Review

A review on mixture design methods for self-compacting concrete

Caijun Shi ⇑, Zemei Wu, KuiXi Lv, Linmei Wu
College of Civil Engineering, Hunan University, Changsha 410082, China

h i g h l i g h t s

 Five mixture design methods for SCC based on different principles are reviewed.
 Feature and flow chart of mixture design procedure for each method is presented.
 Advantages and disadvantages of each method is compared.
 It provides valuable suggestions for choosing appropriate design method for SCC.

a r t i c l e i n f o a b s t r a c t

Article history: Mixture design is a very important step in production and application of concrete. Many mixture design
Received 1 January 2015 methods have been proposed for self-compacting concrete (SCC). This paper presents a critical review on
Received in revised form 13 March 2015 SCC mixture design methods in publications. Based on principles, those methods can be classified into
Accepted 16 March 2015
five categories including empirical design method, compressive strength method, close aggregate packing
Available online 27 March 2015
method and methods based on statistical factorial model and rheology of paste model. The procedures,
advantages and disadvantages of each method were discussed. The most appropriate method should
Keywords:
be chosen according to actual situations to obtain high quality SCC with satisfactory properties.
Self-compacting concrete
Mixture design method
Ó 2015 Elsevier Ltd. All rights reserved.
Classification
Advantages and disadvantages

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
2. Mixture design methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 389
2.1. Empirical design method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 389
2.2. Compressive strength method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 390
2.3. Close aggregate packing method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 391
2.4. Mixture design method based on statistical factorial model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393
2.5. Mixture design method based on rheology of paste model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395
3. Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 397
Acknowledgement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 397
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 397

1. Introduction segregation or bleeding. The concept of SCC was first proposed by

Okamura in 1986, and the prototype was first developed by Ozawa
Self-compacting concrete (SCC) is a special type of concrete at the University of Tokyo in 1988 [1,2]. SCC has many advantages
which can be placed and consolidated under its own weight with- over conventional concrete, including: (1) eliminating the need for
out any vibration effort due to its excellent deformability, and vibration; (2) decreasing the construction time and labor cost; (3)
which at the same time is cohesive enough to be handled without reducing the noise pollution; (4) improving the filling capacity of
highly congested structural members; (5) improving the interfacial
⇑ Corresponding author. Tel./fax: +86 731 8882 3937. transitional zone between cement paste and aggregate or
E-mail address: [email protected] (C. Shi). reinforcement; (6) decreasing the permeability and improving

http://dx.doi.org/10.1016/j.conbuildmat.2015.03.079
0950-0618/Ó 2015 Elsevier Ltd. All rights reserved.
388 C. Shi et al. / Construction and Building Materials 84 (2015) 387–398

Table 1
Summary of existing mixture design methods for SCC in the literatures.

Classification Authors Year Main features Refs.

Empirical design Okamura, and Ozawa 1995 Fix coarse and fine aggregate first, and then obtain self-compactability by adjusting W/B and [34]
method superplasticizer dosage
Edamatsa, Sugamata, 2003 Use mortar flow and mortar V-funnel testing to select the fine aggregate volume, volumetric water-to- [36]
and Ouchi powder ratio and superplasticizer dosage
Domone 2009 For a given set of required properties, make the best estimation of the mixture proportions, and then [38]
carry out trial mixes to prove
Khaleel and Razak 2014 Conduct in three phases, i.e. paste, mortar and concrete [37]
Compressive Ghazi, and Al Jadiri 2010 Based on the ACI 211.1 method for proportioning conventional concrete and the EFNARC method for [39]
strength proportioning SCC
method Dinakar, Sethy, Sahoo 2013 Use GGBS in SCC based on the strength requirements and consider the efficiency of GGBS [33]
Close aggregate Hwang, and Tsai 2005 Use Densified Mixture Design Algorithm (DMDA), derived from the maximum density theory and [42]
packing method excess paste theory
Petersson, Billberg, and 1996 Mainly based on the void content and the blocking criteria [43]
Van
Su, Hsu, and Chai 2001 Use packing factor (PF) to control the content of fine and coarse aggregate in mixture proportion [44]
Sedran, and De Larrard 1996 Use software to design SCC based on the compressible packing model (CPM) [46]
Shi, and Yang 2005 Use a combination of the excessive paste theory and ACI guidelines to design self-consolidating [3]
lightweight concretes
Sebaibi, Benzerzour, 2013 Based on FN EN 206-1 standard, compressible packing mode (CPM) and packing factor (PF) [47]
Sebaibi, and Abriak
Kanadasan and Razak 2014 Integrate the actual packing level of aggregate and paste volume into the proportioning method based [48]
on the particle packing to obtain the final mixture design
Statistical factorial Khayat, Ghezal, and 1999 Obtain a statistical relationship between five mixture parameters and the properties of concrete [49]
model Hadriche
Ozbay, Oztas, 2009 Design in a L18 orthogonal array with six factors, namely, W/C ratio, water content (W), fine aggregate [52]
Baykasoglu, Ozbebek to total aggregate (S/a) percent, fly ash content (FA), air entraining agent (AE) content, and
superplasticizer content (SP)
Bouziani 2013 Useful to evaluate the effect of three types of sand proportions (river sand, crushed sand and dune [53]
sand), in binary and ternary systems, on fresh and hardened properties of SCC
Rheology of paste Saak, Jennings, and 2001 Avoid segregation of the aggregates as a critical design parameter, then a new segregation-controlled [54]
model Shah design methodology is introduced for SCC
Bui, Akkaya, and Shah 2002 Expand Saak’s concepts to include the effects of aggregate (and paste) volume ratio, particle size [55]
distribution of the aggregates and fine to coarse aggregate ratio, to propose a new paste rheology
model
Ferrara, Park, and Shah 2007 Steel fiber-reinforced self-compacting concrete based on the paste rheology model [57]

Air content: 4-7%

in fresh and hardened states for particular applications. Different
from conventional concrete, a quality SCC should have three key
properties [5]: (1) filling ability – the ability to flow into the form-
Coarse aggregate content: the ratios of the coarse work and completely fill all spaces under its own weight; (2) pass-
ing ability – the ability to flow through and around confined spaces
aggregate volume to solid volume is 0.50
between steel reinforcing bars without segregation or blocking; (3)
segregation resistance – the ability to remain homogeneous both
Fine aggregate content: during transporting, placing and after placing. In addition to good
self-compactability, designed SCC also should meet the require-
V funnel testing using coarse aggregate
ments for strength, volume stability and durability of the hardened
concrete at the same time [6]. Due to those obvious advantages,
VW/VP: mortar flow testing
SCC has been a research focus for many years. Five North
American conferences [7–9], seven RILEM conferences [10–12]
SP dosage: mortar V-funnel testing
and three symposiums on design, performance and use of SCC
[13–15] have been held so far.
It has reported that factors including composition of raw mate-
rials, incorporation of chemical and mineral admixtures, aggregate,
NO Measured properties >
packing density, water to cement ratio (W/C) and design methods
required ones˛ has significant effects on properties in terms of rheology, strength,
shrinkage and durability of SCC [16–19]. Hu and Wang [20] showed
that graded aggregate could considerably reduce yield stress and
YES
viscosity of concrete. The increased paste volume could enhance
SCC the rheological properties of SCC [21,22]. SCC designed using modi-
fied Brouwers’ method exhibited satisfied workability with recom-
Fig. 1. Mixture design procedure proposed by Edamatsa. mended dosage of high range water reducer [19]. With the world
moving toward to sustainable development, waste materials such
the durability of concrete, and (7) facilitating constructability and as fly ash (FA), rice husk ash (RHA), crushed limestone powder
ensuring good structural performance [3,4]. [23], waste glass powder [24,25], recycled and tire rubber aggre-
Concrete mixture design is a selection of raw materials in gates have been used in SCC [26–28]. It is reported that the strength
optimum proportions to give concrete with required properties of SCC improved with the increasing content of superplasticizer
 C. Shi et al. / Construction and Building Materials 84 (2015) 387–398 389

Flowability test of cement paste: Flowability test of cement mortar:

Determine water demand; Determine optimum sand content.
Determine optimum SP dosage.

Control mortar Metakaolin mortars

Absorption of Compressive strength of mortar Flowability & filling ability of

mortar mixtures
mixtures: Determine the mortar mixtures: Determine
optimum replacement level of optimum SP dosage and replace
powder mortars different level of pozzolan
Setting time test

Control Concrete Optimum MK Concrete

Fresh tests of concrete mixes: Accepted results according to

Slump flow cone, V-funnel box, typical acceptance criterion
L-box, and segregation sieve. for SCC

Fig. 2. Flow chart of mixture design procedure of the approach proposed by Khaleel (modified based on Ref. [37]).

of uniform criterion, specific design parameters or factors to evalu-

Specify concrete properties: Filling ability, ate the SCC design process, which makes it difficult to compare the
passing ability and Segregation resistance. effectiveness of different design methods and properties of SCC. This
paper classified the mixture design methods of SCC into five cate-
gories based on their design principles. The procedures, advantages
and drawbacks of each method were presented and compared. It is
Materials information
the purpose to review the progresses and to provide valuable scien-
tific bases for selection of suitable mixture design methods of SCC.
Recommend
Coarse aggregate content Vca
values
2. Mixture design methods

Fine aggregate content: There are many mixture design methods for SCC. Domone [38]
Vfa (%)=0.45(100-Vca) and Petersson [43] presented a model respectively in 1996. In
1999, the Laboratory Central des Ponts et Chausses (LCPC) [46]
developed an approach on the basis of the BTRHEOM rheometer
Paste volume: Vpa (%)=100-Vca-Vfa and RENE-LCPC software. Su et al. [44] introduced a coefficient
called packing factor (PF) to adjust the relative content of aggre-
gate and paste. Hwang [42] et al. proposed a densified mixture
W/P and SP dosage: design algorithm, which was derived from the maximum density
the spread and V-funnel tests theory and excess paste theory. Saak et al. [54] used rheology of
paste model for the design of fiber-reinforced SCC. Ghazi et al.
[39] developed a new method capable of proportioning SCC mix-
Trial concrete mixtures tures with specified compressive strength. Recently, Sebaibi et al.
[51] proposed a new mixture design method based on the
Fig. 3. Mixture design procedure of UCL method. European standard (EN206-1), the Chinese method and the
optimization of the granular packing. Moreover, there are some
(SP) when 10% RHA was incorporated [29]. Economical SCC could be modified mixture design methods based on those existed methods
successfully developed with 28-day compressive strengths from 26 [31–33]. The existing mixture design methods for SCC in the litera-
to 48 MPa with incorporation of 40–60% FA [30]. In addition, Long tures are summarized in Table 1. Based on the design principles,
et al. [28] indicated that the incorporation of rubber aggregates sig- those methods can be classified into five categories: empirical
nificantly influenced yield stress of fresh SCC specimen and the design method, compressive strength method, close aggregate
compressive strength at 28 days, depending on the size distribution packing method, methods based on statistical factorial model
and volume percentage of the rubber aggregate. and rheology of paste model. The following sections discuss these
As a vital step to the production of concrete, many researchers methods in details.
from all over the world have done a lot of researches on mixture
design of SCC, and proposed a variety of mixture design methods 2.1. Empirical design method
based on different principles or control parameters. Mixture design
methods or guidelines for SCC have been promulgated with wide Empirical design method is based on empirical data involving
applications in many countries and regions. However, there is a lack coarse and fine aggregates content, water and cementitious
390 C. Shi et al. / Construction and Building Materials 84 (2015) 387–398

1. Required 2. Maximum weight of 3. W/C, water and

4. Gravel content
compressive water and air content cement contents
strength of
NO
SCC
Measured
properties>required 6. Fine aggregate content 5. Powder content

ones ?

YES

SCC

Fig. 4. Mixture design procedure of the method proposed by Ghazi.

Table 2
SCC compressive strength versus W/C (Table 3 in Ref. [39]).

fc (MPa) 15 20 25 30 35 40 45 50 55 60 65 70 75
W/C 0.8 0.7 0.62 0.55 0.48 0.43 0.38 0.35 0.34 0.33 0.32 0.31 0.29

material contents and superplasticizer dosage to determine the ini- with coarse aggregates of different properties. The mixture design
tial mixture proportions. The best estimates of the mixture propor- procedure is shown in Fig. 2. Experiments were conducted on
tions for required properties are carried out through several trial paste, mortar and concrete to facilitate the mixture design process.
mixes and adjustment. It is indicated that this method was good in production of SCC with
Okamura et al. [1,34] proposed a mixture design method for coarse aggregate of different properties. The use of metakaolin in
SCC based on experiences. The design procedure included the fol- concrete can not only a good choice for utilization of wastes but
lowing aspects: (1) coarse aggregate content in the concrete was also enhance properties of SCC.
fixed at 50% of the solid volume; (2) fine aggregate content was Domone et al. [38] also proposed a method based on experience
fixed at 40% of the mortar volume; (3) water-to-powder ratio and understanding of the behavior of SCC named UCL method. The
was assumed between 0.9 and 1.0 by volume, depending on the method estimated the mixture proportions for a given set of
properties of the powder; (4) superplasticizer dosage and the final required properties, then adjusted it by trial mixes. The mortar
water-to-powder ratio were determined so as to ensure self- fraction of concrete was tested using spread and V-funnel tests
compactability. to determine the water-to-powder ratio and superplasticizer
This approach is very easy to follow, but there were no parame- dosage. Fig. 3 shows the procedure of this method. In this method,
ters describing the properties of aggregate. In order to obtain only standard tests for fresh concrete are needed and other compli-
higher workability and moderate viscosity, higher dosage of super- cated tests such as rheology behavior of mortar or concrete are
plasticizer must be used, which could result in retarding of con- avoided.
crete and increases the cost of SCC as well. Although this method A significant advantage for the empirical design method is its
is based on experiences, it is a simple approach for designing SCC. simplicity. However, intensive laboratory testing is needed to
Edamatsa [35,36] improved the method by fixing fine aggregate obtain compatible behavior for available constituents and satisfac-
ratio, volumetric water-to-powder ratio and superplasticizer tory mixture proportions. Besides that, changes in raw materials
dosage. Fig. 1 shows the mixture design procedure. Compared with will need intensive re-testing and adjustments.
Okamura’s approach, this method can be applicable to powder
materials and aggregates of various qualities. However, further 2.2. Compressive strength method
work is required to characterize the properties of raw materials,
including the compactability between powder materials and This type of method determines cement, mineral admixtures,
superplasticizers. water and aggregate contents based on required compressive
Khaleel et al. [37] proposed a design method, which was similar strength. Ghazi et al. [39] proposed a straightforward method for
to Edamatsa’s approach, for self-compacting metakaolin concrete SCC mixture design based on ACI 211.1 [40] method for

Select components Fix the total powder Fix the GGBS percentage Determine water
or cementitious and calculate efficiency content of
Re-design mixture of GGBS at 28 days mixture
content

NO
Check with Determine sand/total
Determine
EFNARC guidelines aggregate ratio using
YES superplasticizer dosage
standard gradation curves

Go for the
development of Trial mixtures and tests Determine final mixture
SCC on SCC properties composition

Fig. 5. Outline of the mixture design method for SCC containing GGBS (modified based on Ref. [33]).
 C. Shi et al. / Construction and Building Materials 84 (2015) 387–398 391

Using the proposed method and established efficiency values

for GGBS, SCC with strengths range from 30 to 100 MPa at GGBS
replacement levels from 20 to 80% could be developed. This
method considered the efficiency of pozzolanic materials and pre-
sented a way for using high volume replacements up to 80% for
30 MPa.
The compressive strength method presents a clear and precise
procedure to obtain specific quantities of ingredients and mini-
mizes the need for trial mixtures. In addition, the proposed method
takes into consideration the gradation of fine and coarse aggre-
gates or the contributions of pozzolanic materials to the properties
of concrete. However, one of its weekness is that it requires adjust-
ments to all ingredients like sand, coarse aggregate, superplasticiz-
ers and water, to achieve an optimal mixture proportion.

2.3. Close aggregate packing method

Fig. 6. The procedure of aggregate packing (modified based on Ref. [42]). This type of mixture design method determines mixture pro-
portions by obtaining ‘‘the least void’’ between aggregates based
proportioning conventional concrete and EFNARC [41] method for on packing model first, then applying pastes to fill the void
proportioning SCC. In this method, the coarse aggregate content between aggregates.
depended on the maximum aggregate size (MAS) and fineness Hwang et al. [42] proposed a method based on the Densified
modulus of the fine aggregate. The water content was determined Mixture Design Algorithm (DMDA). The effects of three types of
based on both the maximum aggregate size and concrete strength. aggregate packing (primitive, dense, gap gradation) on void within
The W/C and the water-to-powder volume ratios were determined aggregates and the property of produced concrete were investi-
by the compressive strength of concrete. Its brief flow chart is gated [42]. The primitive packing type used sand to fill the void
shown in Fig. 4. between coarse aggregate, and then used fly ash to fill the void
The original ACI 211.1 method covers the design of compressive between aggregates as shown in Fig. 6. Dense packing type used
strength from 15 to 40 MPa. However, this method expanded com- the standard sieves of 3/8 in, Nos. 4, 8, 16, 30 and 50 to separate
pressive strength range from 15 to 75 MPa for SCC, with maximum aggregates into different sizes, and the remained fine particle
W/C as shown in Table 2. This method also needs to use some rele- was omitted. Then followed the similar packing procedure of the
vant tables in reference [39]. primitive packing type as shown in Fig. 6 by iterative filling the
Dinakar et al. [33] proposed a method for SCC containing coarse particle with finer one from 3/8 in to No. 50 and finally filled
granulated blast-furnace slag (GGBS) using efficiency factor. The with fly ash to wholly pack the aggregates. Results indicated that
method consisted of five steps as shown in Fig. 5. The total powder the dense-graded curves were quite close to the Fuller’s curve, as
content was fixed in the first step, the percentage of slag was fixed shown in Fig. 7.
based on the strength required. The efficiency factor (k) was deter- DMDA was derived from the maximum density theory and
mined for the same percentage with the equation proposed in the excess paste theory, and was the durability design concept to
second step. In the third step the water content required for SCC achieve minimum water and cement content by applying fly ash
was determined and the coarse and fine aggregates were then
determined using appropriate combined aggregate gradation
curves of DIN standards. Finally the self-compactability of the fresh
Select proper material source; Obtain the maximum density by
concrete was evaluated through the slump flow measurement
and flowability through V-funnel testing, and passing ability Obtain material information iterative packing of aggregate
through L-box testing.
Assign volume of paste
Calculate the least void VV
amount VP=nVV

Calculate the volume of aggregate Vagg

Determine the SP and water content

Measured
NO
properties>required
ones ?

YES

SCC

Fig. 7. The gradation curves of three packing types (modified based on Ref. [42]). Fig. 8. Mixture design procedure of the method proposed by Hwang.
392 C. Shi et al. / Construction and Building Materials 84 (2015) 387–398

This method considered concrete as a solid aggregate phase in a

Construction Criteria Void Content Blacking Criteria liquid paste phase formed by powder, water and admixtures. The
paste fills the void in the aggregate matrix and provides a lubricat-
ing layer around each particle. In this method, the risk of blocking
Paste Volume was calculated using the following equation.
X
Risk of blocking ¼ ðV ai =V abi Þ 6 1 ð1Þ
Mortar volume
where Vai is the volume of aggregate group i and Vabi is the blocking
volume of aggregate group i. By using Eq. (1) together with the
blocking criteria, the minimum paste volume for different gravel
Coarse aggregate content,
to total aggregate ratios can be calculated. The procedure of this
SP dosage method is shown in Fig. 9.
This method is notable for its importance but is not that easy to
apply. It enables to design mixtures for a specific bar spacing with
sufficient lubrication between aggregates. However, there are no
Measured adequate methods to justify uniformity of the mixture.
NO
properties>required Su et al. [44,45] proposed a mixture design method for SCC
ones ? using a packing factor (PF). The principal consideration of the
method was to fill the paste of binders into voids of loosely piled
YES aggregate framework. The packing factor (PF) of aggregate is
defined as the mass ratio of tightly packed aggregate to that of
Wanted SCC loosely packed aggregate. Thus the content of fine and coarse
aggregates can be calculated as follows:
Fig. 9. Mixture design procedure of the method proposed by Petersson (modified
based on Ref. [43]). W r ¼ PF  W rL  ð1  S=aÞ ð2Þ

W s ¼ PF  W sL  S=a ð3Þ
to fill the void between aggregates and cement paste to attain ‘‘the
least void’’. The procedure of this method is shown in Fig. 8. The 3
where Wr is the content of coarse aggregates in SCC (kg/m ); Ws is
SCC designed by the DMDA is high flowable, cost-effective and dur- the content of fine aggregates in SCC (kg/m3); WrL is the unit vol-
able. It overcomes concrete problems due to shape, particles dis- ume mass of loosely piled saturated surface-dry coarse aggregates
tribution, gap gradation of aggregates and large amount of in air (kg/m3); WsL is the unit volume mass of loosely piled satu-
cement paste. However, there is very little information concerning rated surface-dry fine aggregates in air (kg/m3); S/a is the volume
the passing ability through reinforcement and segregation ratio of fine aggregates to total aggregates, which ranges from 50
resistance. to 57%. The procedure of this method is shown in Fig. 10 [45].
Petersson et al. [43] proposed a mixture design method for SCC This method is simple and uses a smaller amount of binders. PF
based on a relationship between the blocking volume ratio and determines the aggregate content and influences the strength,
clear reinforcement spacing to fraction particle diameter ratio. flowability and self-compacting ability. However, how to

Required workability Required strength

Packing factor PF
Water to cement
Cement content C
ratio Wc/C

Fine aggregate Coarse aggregate

content Af content Ac
Water
content Wc

Pozzolanic paste volume Vpp

Fly ash content F GGBS content S SP dosage

Water content Wf Water content Ws

Total water
content W

Fig. 10. Mixture design procedure of method proposed by Su et al (Fig. 1 in Ref. [45]).
 C. Shi et al. / Construction and Building Materials 84 (2015) 387–398 393

and blockage while pumping. Besides, it is difficult for others to use

this method without purchasing the software.
Half saturation amount of SP Initial combination of binders Shi et al. [3] proposed a method for self-consolidating light-
weight concretes (SCLCs), using a combination of the excessive
paste theory and ACI guidelines for the design of conventional
Measure the water demand structural lightweight concrete. Glass powders and ASTM Class F
fly ash were added to produce excessive paste to increase the
flowability and segregation resistance of the concrete. The proce-
Run RENE-LCPC to optimize the mixture proportion dure of this method is shown in Fig. 12. The designed SCLC mix-
tures exhibited good flowability and segregation resistance.
Sebaibi et al. [47] proposed a method based on the compressible
Adjust water content to gain the target viscosity
packing model [46], the method proposed by Su [44] and the EN
206-1 standard. In this method, RENE-LCPC software was used to
optimize the composition of SCC. The Eqs. (2) and (3) were used
Adjust SP dosage to gain suitable slump flow;
to calculate the content of coarse and fine aggregates respectively.
Check with general criteria
The paste amount of pozzolanic materials was calculated using the
NF EN 206-1. The procedure of this method is shown in Fig. 13. The
W/C was selected accoring to Fig. 14.
Measured The SCC designed with the method contains more aggregate but
NO less binder. The ratio of fine aggregate to mortar volume was 60%,
properties>required
which was higher than the value of 40% proposed by Okamura.
ones ?
Then a concrete mixture designed by the proposed method
requires a smaller quantity of binder, rather higher ratio of fine
YES
aggregate to mortar volume.
Check rheological behavior Kanadasan et al. [48] used the particle packing concept to
ensure the fresh and hardened properties of SCC incorporating
waste product of palm oil clinker aggregate. The actual packing
Fig. 11. Mixture design procedure of the method proposed by Sedran (modified level of aggregate and paste volume were integrated into the
based on Ref. [46]). method. The flow chart for the mixture design procedure is shown
in Fig. 15.
The results indicated that the mixture design could be
employed not only for palm oil clinker but also for a variety of
determine the optimum sand to aggregate ratio or the packing fac-
combinations of aggregate. It not only helps to conserve the natu-
tor is not explained. These two values are assumed empirically to
ral resources but also promotes sustainability in preserving the
carry out the mixture design.
environment.
Sedran et al. [46] proposed a method based on the compressible
packing model (CPM), which is the third generation of packing
models developed at LCPC. CPM first calculated virtual packing 2.4. Mixture design method based on statistical factorial model
density of solid particles with different particle size distributions
according to the packing structure; then through the compaction This method is based on the effects of different key parameters
index K, the relationship between virtual packing density and such as the contents of cement and mineral admixtures, water-to-
actual packing density was established in different packing pro- powder ratio, volume of coarse aggregate, and dosage of SP etc. on
cess. Finally, a nonlinear equation was solved to get the actual workability and compressive strength of fresh and hardened SCC.
packing density. In this method, a BTRHEOM rheometer and a Reasonable ranges for each parameter are determined, and mixture
RENE-LCPC software were needed to be used together for SCC proportion is calculated according to mixture design of conven-
design. The procedure of this method is shown in Fig. 11 [46]. tional concrete.
The method focuses on optimizing the granular skeleton of con- Khayat et al. [49,50] proposed a statistical factorial model by
crete from the viewpoint of packing density. Sometimes, it could selecting five key mixture parameters to design SCC. The five key
result in very low paste content, causing a rapid loss of slump flow parameters were the cementitious material content (CM), the ratio

Determine the void volume in the Determine optimum Determine cement content and
dry binary aggregate mixtures combination of coarse W/C according to strength
according to ASTM C29 and fine aggregates requirement and ACI 211.2,

NO

Measured
Determine mineral Determine volume of excess
properties>required
admixtures content paste through experiment
ones ?

YES

SCLCs

Fig. 12. Mixture design procedure of the method proposed by Shi.

394 C. Shi et al. / Construction and Building Materials 84 (2015) 387–398

Use Rene-LCPC to calculate Calculate fine and coarse

Calculate cement
the experimental packing aggregate content,
content: C=fc’/0.14
density of the binary mixture according to (2) and (3)

NO

Select W/C according

Measured
YES to Fig. 14
SCC properties>required
ones ?

Calculate silica fume:

Use marsh cone to obtain the SF/(SF+C)=0.10
optimum dosage of SP And (W/b)max=0.45

Fig. 13. Mixture design procedure of Sebaibi’s method.

Age 28d design was carried out to mathematically reflect the influence of
five key parameters on filling and passing abilities, segregation
Compressive strength/MPa

60 and compressive strength, which are important for the successful

50 development of medium strength SCC incorporating pulverised
40 fuel ash (PFA). The parameters were the contents of cement and
PFA, water-to-powder (cement + PFA) ratio (W/P) and dosage of
30 SP. The responses of the derived statistical models are slump flow,
20 fluidity loss, Orimet time, V-funnel time, L-box, J-Ring combined to
10 the Orimet, J-Ring combined to cone, rheological parameters,
segregation and compressive strength at 7, 28 and 90 days.
0 Twenty-one mixes were prepared to derive the statistical models,
0.2 0.3 0.4 0.5 0.6 0.7
and five were used for the verification and the accuracy of the
W/C
developed models. The results showed that medium strength SCC
Fig. 14. Relationship between compressive strength and water-to-cement ratio. with 28-day compressive strengths of 30 to 35 MPa could be
achieved by using up to 210 kg/m3 of PFA.
Ozbay et al. [52] analyzed mixture proportion parameters of
of water to cementitious materials (W/CM), the concentrations of high strength self-compacting concrete (HSSCC) by using the
high-range water reducer (HRWR), viscosity-enhancing agent Taguchi’s experiment design method for optimum design.
(VEA) and the volume of coarse aggregate (Vca). Statistical factorial Mixtures were designed using L18 considering six factors including
design models were used to derive design charts which correlate W/C, water content (W), fine aggregate to total aggregate percent
input mix-design variables to output material properties, mainly (S/a), fly ash content (FA), air entraining agent (AE) content and
consisting of the measurements of fresh state properties as well superplasticizer content (SP). One of the advantages of the
as the compressive strength. The resulting understanding of the Taguchi method is that it minimizes the variability around the tar-
interaction between the key parameters can be used for both mix get when bringing the performance value to the target ones.
optimization and quality control. Another advantage is that the optimum working conditions deter-
Sonebi [51] used statistical factorial model to design medium mined from the laboratory can also be reproduced in full scale
strength SCC containing fly ash. In his experiment, a factorial production.

Select materials Physical characterization tests Determine of aggregate

substitution ratio

Determine aggregate and Select of correction Measure particle packing:

cement content lubrication factor Void volume; Particle packing

NO
Determine paste volume

Check with YES

POC SCC Design
Determine water and EFNARC guidelines
additional powder content

Verification test - Trial mix Excess paste effect

Fig. 15. Flowchart of achieving and verifying the mixture design for SCC using POC aggregate (modified based on Ref. [48]).
 C. Shi et al. / Construction and Building Materials 84 (2015) 387–398 395

ðq þ m  1Þ!
C¼ ð4Þ
m!ðq  1Þ!

where q is the number of factors and m is the number of levels.

When three factors and five levels are considered, the number of
combinations to be treated is 21.
A mathematical model describing the effects of three sands and
their combinations on given property can be established using this
approach. A second-degree model was used with three non-
independent variables (proportions of RS, CS and DS) and five
levels, as expressed as follows:

Y ¼ b1  RS þ b2  CS þ b3  DS þ b4  ðRS  CSÞ þ b5  ðRS  DSÞ

þ b6  ðCS  DSÞ ð5Þ
The model’s coefficients (bi) represent the contribution of the
associate variables on the response Y, which were determined by
a standard least-square fitting using statistical software. Although
this method is accurate and avoids extensive repeated experiments,
it refers to specialized statistics knowledge, which makes it difficult
for people to follow without this basic knowledge.
Fig. 16. Illustration of the simplex-lattice design with three factors (RS, CS and DS) The factorial design approach is valid for a wide range of mix-
and five levels (Fig. 1 in Ref. [53]).
ture proportion and provides an efficient means to determine the
influence of key variables on SCC properties. Such understanding
can facilitate the test protocol required to optimize SCC, hence
Bouziani et al. [53] developed a mixture design method to reduce the effort necessary to optimize specified concrete to secure
evaluate the effects of three types of sand including river crushed balance between various variables affecting flowability, deforma-
and dune sand, in binary and ternary combinations, on properties bility, stability and strength. However, establishment of statistical
of fresh and hardened SCC. A simplex-lattice mixture design with relationships needs intensive laboratory testing on available raw
three factors and five levels was carried out. All other SCC compo- materials.
nents (coarse aggregate, cement, addition, superplasticizer and
water) were kept constant. The simplex-lattice design is a space 2.5. Mixture design method based on rheology of paste model
filling design that creates a triangular grid of combinations, as
shown in Fig. 16, where the number of combinations (C) is Saak et al. [54] developed a ‘‘rheology of paste model’’ to design
expressed by the following equation: SCC. The method proposed that the rheology of the cement paste

Fresh SCC

Solid phase (fine and Design and Liquid phase (cement, air
coarse aggregates) Construction criterion and admixtures)

Criteria for solid phase Criteria for liquid phase

(aggregate blocking model) (paste model)

Water to binder ratio,

Minimum paste volume,
Coarse - total aggregate ratio

Adjust W/B
Adjusted
or paste Superplasticizer
Paste rheology
volume
Unsatisfactory Adjusted
Concrete trial
If no OK

Final mixture proportion

(High performance and economic efficiency)

Fig. 17. Flow chart for mixture design procedure using rheology models (Fig. 13 in Ref. [55]).
396 C. Shi et al. / Construction and Building Materials 84 (2015) 387–398

Select raw materials for Select fine and coarse

cement paste aggregate, and fibers

Model for rheological behavior Optimal grading of solid skeleton:

of cement paste: Average diameter of particle: dav
Mini-cone flow test: rheometer test Measure void ratio: Vvoid

Assess paste volume ratio Vp Assess solid volume ratio Vsolid

Average spacing of
solid particles dss

Assess correlation between cement paste

rheology, solid skeleton gradation and
paste/solid volume ratio

Identify allowable values of dss for self-compactability

Select paste/solid volume ratio

Identify optimum Optimally graded

rheological properties Mix-design of solid skeleton for the
of cement paste and SCSFRC given paste/solid ratio
select its composition

Fig. 18. Flow chart for mixture design of SCSFRC (modified based on Ref. [57]).

"sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi #
matrix largely dictated the segregation resistance and workability 3 V paste  V void
of fresh concrete, given a specified particle size distribution and dss ¼ dav 1þ 1 ð7Þ
V concrete  V paste
volume fraction of aggregate. The applicability of the method is
tested by measuring the flow properties of fresh concrete. This can be hence regarded as an indicator of the degree of sus-
Additionally, it is proposed that a minimum paste yield stress pension of the given solid skeleton. The rheological properties of
and viscosity must be exceeded to avoid segregation under both the paste (yield stress and viscosity) have to be optimized with
static (rest) and dynamic (flow) conditions, respectively. respect to the average aggregate diameter and as a function of
Bui et al. [55] extended Saak’s concepts to include the effects of the aggregate spacing. The procedure of this method is shown in
aggregate (and paste) volume ratio, particle size distribution of the Fig. 17.
aggregates and fine to coarse aggregate ratio. These factors, The paste rheology model and criteria related to aggregate spac-
together with the aggregate shape, influence the void content ing and average aggregate diameter can be applied for different
and the average diameter of the solid skeleton particles. The aver- coarse-to-total aggregate ratios, cement contents, and water-to-
age diameter of the solid skeleton particles is defined as: binder ratios as well as different contents and types of fly ash.
P The paste rheology model can reduce the extent of laboratory work
di mi
dav ¼ Pi ð6Þ and materials used, and provide the basis for quality control and
i mi
further development of new mineral and chemical admixtures.
where di is the average diameter of aggregate fraction i and mi is the Farrara et al. [57] proposed a method for steel fiber-reinforced
mass of that fraction. SCC based on the paste rheology model. The applicable fibers are
A minimum volume of cementitious paste is needed to fill the treated as an ‘‘equivalent spherical particle’’ fraction, with 100%
voids between the aggregate particles and create a layer envelop- passing fraction at an equivalent diameter, deq-fibers, defined
ing the particles, thick enough to ensure the required deformability through the specific surface area equivalence:
and segregation resistance of concrete. Hence, the average aggre-
3Lf cfiber
gate spacing dss [56], defined as twice the thickness of the excess deq-fibers ¼ ð8Þ
paste layer enveloping the aggregates: 1 þ 2 dLf caggregate
f
 C. Shi et al. / Construction and Building Materials 84 (2015) 387–398 397

where Lf and df are the length and diameter of the fibers, respec- References
tively, cfiber is the specific weight of fibers and caggregate is the
weighted average specific weight of all the aggregates. [1] Okamura H. Self-compacting high-performance concrete. Concr Int Design
Constr 1997;19(7):50–4.
For the fiber-reinforced skeleton, the ‘‘average equivalent [2] Ozawa K. Development of high performance concrete based on the durability
diameter of solid particles’’ can be expressed as: design of concrete structures. EASEC-2 1989;1:445–50.
[3] Shi C et al. Design and application of self-compacting lightweight concrete. In:
P SCC’2005-China: 1st international symposium on design, performance and use
i di mi þ deq-fibers mfibers
dav ¼ P ð9Þ of self-consolidating concrete. RILEM Publications SARL; 2005.
i mi þ mfibers [4] Shi C, Wu Y. Mixture proportioning and properties of self-consolidating
lightweight concrete containing glass powder. ACI Mater J 2005;102(5).
[5] Goodier CI, Development of self-compacting concrete. 2003.
where di, mi and deq-fibers are defined as above; mfibers is the mass of [6] Liu YH, Xie YJ, Long GC. Development of self-compacting concrete. J Chin
the fibers. Ceram Soc 2007;35(5):671–8.
Optimization of rheological properties of cement paste and [7] Attiogbe EK, See HT, Daczko JA. Engineering properties of self-consolidating
concrete. In: Proceedings of the First North American conference on the design
choice of its volume ratio stand as further keys of the method. and use of self-consolidating concrete, 2002, November.
The model proved to be an efficient tool for designing fiber-rein- [8] Billberg P. Development of SCC static yield stress at rest and its effect on the
forced SCC mixtures with selected fresh state properties, employ- lateral form pressure. In: SCC 2005, combining the second North American
conference on the design and use of self-consolidating concrete and the fourth
ing different ratios and types of steel fiber reinforcement. The international RILEM symposium on self-compacting concrete, 2005, p. 583–
procedure of this method is shown in Fig. 18. 589.
[9] Martys N, Ferraris CF. Simulation of SCC flow. In: Proc. 1st North American
Conf. on the design and use of Self-Consolidating Concrete, Chicago, IL, 2002, p.
27–30.
3. Conclusions [10] Klug Y, Holschemacher K. Comparison of the hardened properties of self-
compacting and normal vibrated concrete. In: 3rd RILEM symposium on self
Based on the extensive review on different SCC mixture design compacting concrete, Reykjavik, 2003, p. 596–605.
[11] Fava C, Bergol L, Fornasier G, Giangrasso F, Rocco C. Fracture behaviour of self-
methods in the literatures, the following conclusions can be compacting concrete. In: Wallevik O, Nielsson I, editors. Proceedings of the 3rd
drawn: international RILEM symposium on self-compacting concrete. RILEM
Publications SARL; 2003, p. 628–636.
[12] Wallevik OH, Kubens S, Müller F. Influence of cement-admixture interaction
(1) The empirical design method is easy to follow. However, on the stability of production properties of SCC. In: De Schutter G, Boel V,
intensive laboratory testing on available raw materials are editors. 5th International RILEM symposium on self-compacting concrete.
needed to obtain satisfactory mixture proportions. RILEM Publications SARL; 2007, p. 211–216.
[13] Yu ZW, Shi C, Khayat KH. PRO 42: 1st international symposium on design
(2) The compressive strength method presents a clear and pre- performance and use of self-Consolidating concrete. RILEM Publications; 2005.
cise procedure to obtain specific quantities of ingredients [14] Shi C, Yu ZW, Khayat KH. PRO 65: 2st international symposium on design
and minimizes the need for trial mixtures. Besides that, performance and use of self-Consolidating concrete. RILEM Publications; 2009.
[15] Shi C, Ou ZH, Khayat KH. PRO 93: 3st international symposium on design
the gradation of fine and coarse aggregates, and con-
performance and use of self-Consolidating concrete. RILEM Publications; 2014.
tributions of pozzolanic materials to the properties of SCC [16] Han J, Fang H, Wang K. Design and control shrinkage behavior of high-strength
is taken into consideration. However, this method requires self-consolidating concrete using shrinkage-reducing admixture and super-
absorbent polymer. J Sustain Cem Based Mater 2014:1–9.
adjustments to all the ingredients to achieve an optimal
[17] Esmaeilkhanian B, Khayat KH, Yahia A, Feys D. Effects of mix design
mixture proportion. parameters and rheological properties on dynamic stability of self-
(3) The close aggregate packing method mainly considers the consolidating concrete. Cem Concr Compos 2014;54:21–8.
relationships between paste and aggregate. Hence, this [18] Siddique R, Aggarwal P, Aggarwal Y. Mechanical and durability properties of
self-compacting concrete containing fly ash and bottom ash. J Sustain Cem
method is simpler and requires a smaller amount of binders. Based Mater 2012;1(3):67–82.
However, SCC produced based on this method tends to seg- [19] Wang XH, Wang KJ, Taylor P, Morcous G. Assessing particle packing based self-
regate easily, which is a problem for construction. consolidating concrete mix design method. Constr Build Mater
2014;70:439–52.
(4) The method based on statistical factorial model can simplify [20] Hu J, Wang K. Effect of coarse aggregate characteristics on concrete rheology.
the test protocol required to optimize a given mixture by Constr Build Mater 2011;25(3):1196–204.
reducing the number of trial batches to achieve a balance [21] Jalal M, Mansouri E, Sharifipour M, Pouladkhan AR. Mechanical, rheological,
durability and microstructural properties of high performance self-
among mixture variables. However, establishment of sta- compacting concrete containing SiO2 micro and nanoparticles. Mater Des
tistical relationships needs intensive laboratory testing on 2012;34:389–400.
available raw materials. [22] Girish S, Ranganath RV, Vengala J. Influence of powder and paste on flow
properties of SCC. Constr Build Mater 2010;24:2481–8.
(5) The method based on rheology of paste model can reduce
[23] Shi C, Wu Y. and Riefler C. Use of crushed stone dust in production of self-
the laboratory work and materials, and provide the basis consolidating concretes. In: Liu T, Meyer C, editors. Recycling concrete and
for quality control and further development of new mineral other materials for sustainable development. ACI SP-219, American Concrete
Institute, Farmington Hills, MI 48333; May 2004. p. 115–129.
and chemical admixtures.
[24] Shi C, Wu Y, Riefler C. Use of self-consolidating lightweight concrete for
insulated concrete form system. Concr Int Feb 2006;28(2):40–3.
Mixture design is a critical step to obtain high quality SCC. A [25] Shi C. Effect of different mineral powders on properties of fresh and hardened
good SCC mixture design method should consider: (1) widely self-consolidating concretes. In: Shi C, Khayat K, editors. Workability of Self-
consolidating concrete – roles of its constituents and measuring techniques,
applicable; (2) strong robustness for variable raw materials; (3) ACI SP-233. American Concrete Institute, Farmington Hills, MI 48333; 2006. p.
technical requirements, (4) sustainability and (5) cost. So far, there 65–76.
is no method fully meet the five requirements. Hence, appropriate [26] Sun Z, Young C. Bleeding of SCC pastes with fly ash and GGBFS replacement. J
Sustain Cem Based Mater 2014.
model should be selected according to specified requirements. [27] Hu J, Wang Z, Kim Y. Feasibility study of using fine recycled concrete aggregate
in producing self-consolidation concrete. J Sustain Cem Based Mater
2013;2(1):20–34.
Acknowledgement [28] Long G, Feys D, Khayat KH, Yahia A. Efficiency of waste tire rubber aggregate
on the rheological properties and compressive strength of cementitious
materials. J Sustain Cem Based Mater 2014:1–11.
Financial supports from National Science Foundation of China [29] Memon SA, Shaikh MA, Akbar H. Utilization of rice husk ash as viscosity
under contract Nos. 51378196 and U1305243 are greatly modifying agent in self compacting concrete. Constr Build Mater
2011;25:1044–8.
appreciated.
398 C. Shi et al. / Construction and Building Materials 84 (2015) 387–398

[30] Bouzuubaa N, Lachemi M. Self compacting concrete incorporating high [44] Su N, Hsu K, Chai H. A simple mix design method for self-compacting concrete.
volumes of class F fly ash preliminary results. Cem Concr Res 2001:413–20. Cem Concr Res 2001;31(12):1799–807.
[31] Wu Q, An XH. Development of a mix design method for SCC based on the [45] Su N, Miao B. A new method for the mix design of medium strength flowing
rheological characteristics of paste. Constr Build Mater 2014;53:642–51. concrete with low cement content. Cem Concr Compos 2003;25(2):215–22.
[32] Choi YW, Kim YJ, Shin HC, Moon HY. An experimental research on the fluidity [46] Sedran T, De Larrard F. Optimization of self-compacting concrete thanks to
and mechanical properties of high-strength lightweight self-compacting packing model. In: Proceedings 1st SCC Symp, CBI Sweden, RILEM PRO7; 1999.
concrete. Cem Concr Res 2006;36:1595–602. [47] Sebaibi N et al. Composition of self - compacting concrete (SCC) using the
[33] Dinakar P, Sethy KP, Sahoo UC. Design of self-compacting concrete with compressible packing model, the Chinese method and the European standard.
ground granulated blast furnace slag. Mater Des 2013;43:161–9. Constr Build Mater 2013;43:382–8.
[34] Okamura H, Ozawa K. Mix design for self-compacting concrete. Concr Lib JSCE [48] Kanadasan J, Razak HA. Mix design for self-compacting palm oil clinker
1995;25:107–20. concrete based on particle packing. Mater Des 2014;56:9–19.
[35] Edamatsu Y, et al. A mix-design method for self-compacting concrete based on [49] Khayat KH, Ghezal A, Hadriche MS. Factorial design model for proportioning
mortar flow and funnel tests. In: International RILEM symposium on self- self-consolidating concrete. Mater Struct 1999;32(9):679–86.
compacting concrete. RILEM Publications SARL; 2003. [50] Khayat KH, Ghezal A, Hadriche MS. Utility of statistical models in
[36] Edamatsu Y, Nishida N and Ouchi M. A rational mix-design method for self- proportioning self-consolidating concrete. Mater Struct 2000;33(5):338–44.
compacting concrete considering interaction between coarse aggregate and [51] Sonebi M. Medium strength self-compacting concrete containing fly ash:
mortar particles. In: Proceedings of the first international RILEM symposium modelling using factorial experimental plans. Cem Concr Res
on self-compacting concrete, Stockholm. Sweden; 1999. 2004;34(7):1199–208.
[37] Khaleel OR, Abdul RH. Mix design method for self compacting metakaolin [52] Ozbay E, Oztas A, Baykasoglu A, Ozbebek H. Investigating mix proportions of
concrete with different properties of coarse aggregate. Mater Des high strength self compacting concrete by using Taguchi method. Constr Build
2014;53:691–700. Mater 2009;23(2):694–702.
[38] Domone P. Proportioning of self-compacting concrete – the UCL method. 2009. [53] Bouziani T. Assessment of fresh properties and compressive strength of self-
[39] Ghazi FK, JadiriAl RS. New method for proportioning self-consolidating compacting concrete made with different sand types by mixture design
concrete based on compressive strength requirements. ACI Mater J modelling approach. Constr Build Mater 2013;49:308–14.
2010;107(5). [54] Saak AW, Jennings HM, Shah SP. New methodology for designing self-
[40] Institute AC, Standard practice for selecting proportions for normal, compacting concrete. ACI Mater J 2001;98(6).
heavyweight and mass concrete (ACI 211.1-91), A. Committee, A. Committee [55] Bui VK, Akkaya Y, Shah SP. Rheological model for self-consolidating concrete.
Editors; 1991. p. 33. ACI Mater J 2002;99(6).
[41] Efnarc. Specification and guidelines for self-compacting concrete. 2002. p. 32. [56] Van Khanh B, Montgomery D. Mixture proportioning method for self-
[42] Hwang C, Tsai C. The effect of aggregate packing types on engineering compacting high performance concrete with minimum paste volume. In:
properties of self-consolidating concrete. In: Zhiwu Yu et al., editors. International RILEM symposium on self-compacting concrete. 1999.
SCC’2005-China: 1st international symposium on design, performance and [57] Ferrara L, Park Y, Shah SP. A method for mix-design of fiber-reinforced self-
use of self-consolidating concrete. RILEM Publications SARL, 2005. compacting concrete. Cem Concr Res 2007;37(6):957–71.
[43] Petersson O, Billberg P, Van BK. A model for self-compacting concrete. 1996.