Construction and Building Materials 261 (2020) 119952

Contents lists available at ScienceDirect

Construction and Building Materials

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

Performance optimization of hybrid fiber concrete according to

mechanical properties
Muzeyyen Balcikanli Bankir ⇑, Umur Korkut Sevim
Iskenderun Technıcal Unıversıty, Department of Cıvıl Engıneerıng, Turkey

h i g h l i g h t s

HFC were produced and analyzed using CCDM and RSM.

FA and EAFS were replaced by cement and NCS, respectively.

Variance analysis was performed on experimental findings for mechanical properties of HFC.

Regression models were established for each feature and optimum production conditions were suggested.

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

Article history: In this research, interactional effects of four different fiber types, (steel, glass, synthetic and polypropy-
Received 16 January 2020 lene) fly ash replacement ratio with cement, electrical arc furnace slag replacement ratio with natural
Received in revised form 2 June 2020 aggregate and binder dosage on the mechanical properties (flexural and compressive strength, pull out
Accepted 11 June 2020
capacity, resonance frequency and abrasion resistance) of Hybrid Fiber Concretes were investigated.
Available online 2 July 2020
The HFC samples were exposed to temperatures of 250 °C, 500 °C and 750 °C. The changes in the HFC
samples after exposed to elevated temperature were determined by measuring the compressive strength
Keywords:
and resonance frequency. 30 different experimental programs were conducted by using the central com-
Hybrid fiber concrete
Elevated temperature
posite design method. By using experimental test results analysis of variance was performed and regres-
Resonance frequency sion models obtained. The optimal combination was proposed by optimizing dependent variables and
Pull out capacity maximization of flexural and compressive strength, pull out capacity and resonance frequency while
Optimization minimization of abrasion resistance.
Response surface Ó 2020 Elsevier Ltd. All rights reserved.

1. Introduction The most important feature of concrete is the compressive

strength which is an increasing function of time and it performs
Concrete is widely used material in construction sector and its duty perfectly against concrete compressive loads. However;
needs desired engineering features that will make a great contribu- it is a material which is weak against tensile loads due to cracks
tion to the economy. Although different concretes can be produced caused by various reasons significantly affects the durability of
by making purposeful designs, the main task of the concrete is to concrete. Due to its low tensile strain capacity, concrete has a brit-
carry the loads on the long term period. Nowadays, in all concretes tle behavior, and cracks are almost inevitable in concrete struc-
produced and used in buildings, this desired feature, in other tures [1,2]. The principle of crack formation is generally that the
words, carrying capacity can be easily achieved. According to data tensile forces occurring at any point exceed the strain capacity in
from European Ready Mixed Concrete Association (2017), the the tensile state. This is due to the internal effects (shrinkage and
annual production has reached 109 million m3 in Turkey. In such settlement), expansion of components in concrete or reinforced
a huge sector, it is thought that reducing cost of concrete compo- concrete elements (volume increase due to corrosion of the rein-
nents will make a significant contribution to the country’s econ- forcement, volume increase due to gel formation as a result of
omy like using factory by-products and recycling to reduce alkali silica reactions) and external effects during the service life
natural resource consumption. (overload capacity, chemical effects) [3]. It is tried to prevent the
formation of these cracks with various additives or auxiliary com-
⇑ Corresponding author. ponents by researchers as embedding steel reinforcement in con-
E-mail addresses: muzeyyen.balcikanli@iste.edu.tr (M. Balcikanli Bankir), ukor- crete to compensate the loads by this reinforcement. This
kut.sevim@iste.edu.tr (U.K. Sevim).

https://doi.org/10.1016/j.conbuildmat.2020.119952
0950-0618/Ó 2020 Elsevier Ltd. All rights reserved.
2 M. Balcikanli Bankir, U.K. Sevim / Construction and Building Materials 261 (2020) 119952

reinforcement system does not always find area of usage in struc- ity of the HFC specimens was examined, it was found that steel
tures because of inappropriate geometry or cross-section that fibers with high modulus of elasticity and assumed the role of
requires close reinforcement, as well as over-weighting the struc- bridge in load transfer and delayed crack formation. It has been
ture. In such cases, fibers are added to the concrete mixture, which stated that steel fibers help to prevent the growth of macro cracks,
will prevent the concrete from collapsing immediately under load while propylene fibers prevent micro-cracks due to its lower mod-
and that can perform load transfer. With different fiber types, it is ulus of elasticity [16]. Singh and Rai stated that a high-strength
possible to propose solutions to different problems. In literature fiber increases toughness and low-strength fibers increase the
studies and concrete sector, it is realized that mono type fiber is short-term performance of concrete. They concluded that flexible
used in concrete production by changing the geometries, quanti- and ductile fibers increased toughness and deformation capacity
ties and sort of the fibers used, problem-specific concretes are pro- in the post-cracking region, while strong and hard fibers increased
duced. For example; glass and synthetic fibers are used to prevent the tensile strength. It has been reported that short fibers prevent
fresh state concrete cracks and improve the mechanical properties micro-cracks and tall fibers prevent macro cracks. They concluded
of concrete under both static and dynamic loads. In experimental that the hybridization of the two types of fibers facilitates the con-
studies, the effect of glass and basalt fiber more than 0.25% in con- trol of crack propagation and improves the overall concrete perfor-
crete was investigated. The authors stated that there was almost mance [17]. A team of researchers investigated the flexural
5% increase in compressive strength, while the flexural strength behavior and strength properties of high performance HFC. In fiber
and fracture energy increased in a significant manner [4–7]. Steel reinforced concrete mixtures, silica fume and Granulated Blast Fur-
or carbon fibers are used to increase flexural and tensile strength nace Slag were used as mineral additives replaced with cement
in hardened state of concrete; polypropylene or steel fibers are 10% and 30% by weight. Double hooked end steel fibers, single
used to increase energy absorption capacity and they improve wing end steel fibers, and polyvinyl alcohol (PVA) fibers were
abrasion resistance. In many studies it is indicated that fibers can mixed in different proportions with total volume fractions of 0%,
be incorporated in concrete to increase tensile strength, ductility, 0.6% and 1.2% to develop HFC. Compressive strength, bending
toughness and to improve durability properties [8,9]. Soltanzadeh behavior, rapid chloride permeability coefficients and electrical
et al. stated that contribution of steel fibers to concrete can resistance values were investigated. According to the results of
increase the ductility of concrete and alter the mode of failure of the experiment, it was stated that fiber additive significantly
the concrete structure [10]. Crack propagation mechanism of steel improved the mechanical strength of concrete [18]. To evaluate
fibers is explained as the steel fibers able to bridge the cracks and the multiple effects of fibers and puzzolans Mahapatra and Barahi
transfer the load. Addition of steel fiber limits the propagation of investigated the properties of hybrid fiber reinforced self-
the cracks [11]. compacting concrete with crimped steel fibers and polypropylene
To investigate the influence of mono type steel fiber on the fibers together with F type fly ash and colloidal nano-silica. In their
compressive strength of concrete which contains silica fume, steel study, they proposed an optimal mixture by performing multiple
fiber was used with aspect ratios of 65 and 80 at two fiber contents linear regression analysis. They stated that a good correlation
of 0.5 and 1.0% by volume. The authors stated that the compressive was obtained between the tested and predicted values [19].
strength of concretes was higher that has aspect ratio of 80 than As can be seen in the literature studies, many studies have been
65. Compressive strength of the concretes with the addition of made using different types of fibers. However, in each study, find-
1% steel fiber increased up to 117.6 and 113.8% compared to that ings were presented without generalization. In this study, 30 dif-
of control specimen for the aspect ratios 80 and 65, respectively ferent experimental designs were created by using the Response
[12]. Surface Methodology (RSM). Four different fiber types (steel, glass,
As can be seen from literature studies, partial solutions can be propylene and synthetic) are combined with each other in a total
proposed in concrete by using mono type fiber. Idea of being able volume of not more than 2%. In order to reduce the increasing cost
to solve more than one problem by using different fiber types or with fiber contribution to concrete, fly ash replaced with cement
length, different studies has been done. In the literature, this type and to reduce consumption of natural resources, EAFS aggregates
of concrete is called Hybrid Fiber Concrete (HFC). Seung et al., pro- were replaced with natural crushed stone aggregates according
duced HFC with the combination of macro and micro steel fibers to the rates determined in the experimental design. According to
(smooth, twisted and hooked) with different lengths and geome- experimental results, regression models were obtained by statisti-
tries. According to the results of the experiment, it is stated that cal analysis. According to the regression models, it was determined
macro fibers were effective on tensile stress, and multiple fracture which independent variable affected which feature more.
behavior occurs with the addition of micro fiber. The authors
reported that the cracking resistance, strength after fracture
increased significantly with the increase in microfiber content 2. Materials and method
[13]. Yaoa et al. produced hybrid fiber concrete by using steel,
propylene and carbon fibers with a ratio of 0.5% by volume of bin- 2.1. Materials
ary combinations. Compressive, tensile and bending strength tests
were performed on concrete samples. They stated that the strength Due to fiber own characteristics like high impact strength,
and toughness properties of HFC were better than mono type fiber structural flexibility, crack control, ductility and chemical resis-
concretes and that the bending strengths of polypropylene fiber tance, four different types of fibers were selected in the production
concretes were lower than those of steel or carbon fiber concretes of HFC. Steel (SF), glass (GF), synthetic STRUX 90/40 (SNF) and
[14]. In a study, fibers with high and low elasticity modulus were propylene (PF) fibers were used. Both propylene and fiber fibers
used to produce hybrid concretes. Polypropylene fiber was used are used, although their properties are close to each other, their
as the low modulus fiber and steel, basalt and polyvinyl alcohol accessibility differs in practice. To include both types of fiber in
fibers were used as the modulus of high elasticity. It has been sta- the research in order to examine the effect of technical use of that
ted that the mechanical properties of concrete, which hybridize fiber type, whichever is easier and cheaper to reach during the
steel and polypropylene fibers, are improved perfectly [15]. To market application. The properties of the fibers are given in Table 1.
investigate the crack formation capacity and flexural capacity of The Electric Arc Furnace Slag aggregate (EAFS) which was obtained
the HFC, Jusoh et al. used steel and propylene fibers in a total of _
from Iskenderun/Turkey was crushed and sieved in two dimen-
1.5% by volume for concrete production. When the flexural capac- sions as 0–4 mm and 4–11 mm. Characteristics of the EAFS are
 M. Balcikanli Bankir, U.K. Sevim / Construction and Building Materials 261 (2020) 119952 3

Table 1
Fiber characteristics.

Properties Steel fiber Glass fiber Propylene fiber synthetic fiber

Fiber length, l (mm) 30 10 40 40
Fiber diameter, d (mm) 0.9 13*10-3 0.75 0.75
Tensile strength (N/mm2) 1150 3400 338 620
Modulus of Elasticity (N/mm2) 200,000 77,000 4800 9500
Density (g/cm3) 7.85 2.6 0.91 0.92
Melting point (°C) 1535 1200 160 160
Aspect ratio (l/d) 33.33 769 53.33 53.33

given in Table 2 and particle size distribution of limestone based 100

natural crushed stone aggregates (NCS) and EAFS are given in 0-4 mm NCS
4-12 mm NCS
_
Fig. 1. NCS with a specific gravity of 2.7 g/cm3 obtained from Isk- 80 0-4 mm EAFS
enderun region, separated 0–4 mm and 4–11 mm in the same 4-12 mm EAFS
dimensions with EAFS to replace with each other. CEM I 42.5 R

Passing (%)
60
Portland cement was used. F Class Fly Ash (FA) was used to replace
with cement. The chemical composition of FA is given in Table 3.
Iskenderun city water was used as mixing water. 40

2.2. Mix design of HFC according to CCDM 20

In this research, Central composite Design Method (CCDM) was 0

used to design the HFC mixtures and Response Surface Methodol- 0 0,25 0,5 1 2 4 8 16 31,5
ogy (RSM) was used to define the effect of several independent Sieve size (mm)
parameters that influence responses. The parameters effect the
result called as independent variables and the characteristics of Fig. 1. Particle size distribution of NCS and EAFS.
the concrete are called as dependent variables. CCDM is composed
of a factorial design, a set of central points, and axial points
Table 3
equidistant to the center point. CCDM is a design technique which
Chemical composition of FA and CEM I 42.5 R.
is used for sequential experimentation and requires limited infor-
mation not involving extraordinary large number of points to Properties FA CEM I 42.5 R
design while testing the lack of fit [20,21]. RSM is a technique SiO2 56.2 21.18
which studies the effect of several independent parameters that Al2O3 20.17 6.28
Fe2O3 6.69 3.15
influence dependent parameters (responses) by varying them
CaO 4.24 61.18
simultaneously and carrying out a limited number of experiments. MgO 1.92 3.10
Namely, RSM is a collection of mathematical and statistical meth- SO3 0.49 2.11
ods to evaluate relationship between independent and one or more Loss of ignition 1.78 0.20
dependent variables. The RSM enables to evaluate parameters that Na2O 0.58 0.47
K2O 1.89 1.13
may or may not have significant effect in the main response [22].
To enable the quantification of the prediction of the responses, a
central composite plan was selected, where the response could be
modeled [23,24]. In the CCDM, each numerical factor was deter- ranging from 0% to 2% of the total blend volume. The limit values
mined to be five different values (+a, a, +1, 1 and center point) for each fiber type were determined to be 0% to 0.25% relative to
and to provide 95% confidence level. Seven numerical and zero cat- the total concrete volume, but experimental groups were formed
egorical factors were defined according to the experimental design. between 0% and 0.5% to represent points as far as ‘‘a”. In this case,
Using Design Expert 7.0.3 commercial program, 30 different mix- the total fiber volume reached 2% by hybridizing four types of
tures were offered with 7 independent variables as four of them fibers. The replacement ratio of FA with cement was determined
are type of fibers; (steel, glass, propylene and synthetic) FA between 10% and 40% and the operating range according to the
replacement ratio with cement, EAFS replacement ratio with CCDM was selected as 0% to 40%. The limit values of EAFS replace-
NCS, binder dosage with an a is 1.62658. All size and type aggre- ment ratio with crushed stone aggregates were selected as 15–85%
gates were replaced with each other, namely both 0–4 mm and and the trial range was proposed in the range of 0–100%. The limit
4–11 mm aggregates replaced with 0–4 mm and 4–11 mm NCS values of the binder dosage in 1 m3 were selected as 300 and
by weight, separately. Among these factors, the total fiber volume 400 kg/m3 and the trial range was proposed and applied between
used in HFC was determined by four groups of fiber combinations, 260–500 kg/m3. Water/binder ratio was kept constant for all mix-

Table 2
Characteristics of NCS and EAFS.

Properties Analysis of EAFS Analysis of NCS

Grain size (mm) 0–4 4–11 0–4 4–11
Density (g/cm3) 3.64 3.46 2.74 2.71
Water absorption capacity (%) 2.9 2.5 3.1 2.8
Cl- content (%) 0.0039 0.0025 0.0043 0.0043
Alkaline silica reactivity (%) 0.022 – – –
Petrographic identification Steel slag aggregate Lime stone based aggregate
4 M. Balcikanli Bankir, U.K. Sevim / Construction and Building Materials 261 (2020) 119952

tures as 0.5. Super plasticizer was used to equalize the Slump value 2.3.3. Compressive strength
of all HFC mixtures in different amounts. The main feature of con- 100  100  100 mm3 HFC samples were produced according to
crete is the behavior against compressive and bending loads. In CCDM and cured 28 days in water. Samples were removed from
addition to excellent compressive strength, it is desired to increase water and kept in laboratory conditions for 28, 180 and 360 days
the tensile strength, which is relatively weak, with fiber additives. and tested in accordance with ASTM C 39 on three samples [27].
In order to examine how the type and amount of fiber used in HFC Since HFC was produced by using fiber types which have different
affect these strength properties and adherence between reinforce- melting points so wanted to be determined how the strength prop-
ment and concrete, it is aimed to investigate the mechanical prop- erties of HFC will change depending on the temperature level. HFC
erties of HFCs. As mechanical properties, flexural and compressive samples were subjected to different elevated temperature curing
strength, pull out capacity, resonance frequency and abrasion and as a result of experiments (strength and RF measurements)
resistance are determined as dependent variables in the experi- strength variations in HFC were determined. In order to observe
mental study. the performance of different types of fibers under different tem-
30 different concrete mixtures are proposed according to peratures, the HFC cube samples were exposed 250 °C, 500 °C
Response Surface Method (RSM) with the independent variables. and 750 °C for 2 h and their compressive strength was determined
According to design parameters the mix design was chosen and after returning to ambient temperature.
given in Table 4 and HFC were produced and tested according to
Table 4.
2.3.4. Pull out capacity
To determine the bond strength of HFC with reinforcement, the
samples were subjected to a pull out test. Reinforcement with
2.3. Testing methods
14 mm diameter was embedded in HFC sample produced in
dimension of /100/200 mm and cured in water for 28 days. At
2.3.1. Slump test
the end of 28 days, the samples were cured in the laboratory until
The Slump value of all the mixtures produced in this study was
the day of the experiment. The claws, load cell, hydraulic pump
determined according to TS EN 12350-2 [25] and kept constant to
and the appropriate computer software to record the load of the
be 160–210 mm (S4 class) by using super plasticizer.
test was used and apparatus are shown in Fig. 2a and schematic
representation is shown in Fig. 2b. The resistance of the reinforce-
ment while stripping from concrete is determined as load (kg).
2.3.2. Flexural strength
Four Point Bending test was conducted on the 28 days old HFC
according to ASTM C1161. After the samples removed from the 2.3.5. Resonance frequency
water and let to reach the ambient humidity, the load was applied Longitudinal Resonance Frequency (RF) measurement test was
on the 100  100  500 mm3 beam samples such that the span performed on three prisms HFC samples in accordance with ASTM
between the supports is divided into 3 equal parts. While the load C 215 [28] and mean values are calculated in Hz. The consistency
is applied, the mid span deflection of the HFC samples was mea- level in frequency readings is expected to be within 2%. Measure-
sured [26]. ments were performed using automatic mode. Automatic mode

Table 4
Mix design of HFC.

Mix No Steel fiber (%) Glass fiber (%) Synthetic fiber (%) Propylene fiber (%) FA (%) EAFS (%) Binder dosage (kg/m3) Total fiber ratio (%)
M1 0.25 0.25 0.25 0.25 25.0 0.0 350.0 1.00
M2 0.00 0.15 0.15 0.15 25.0 50.0 350.0 0.45
M3 0.00 0.50 0.50 0.00 10.0 85.0 400.0 1.00
M4 0.00 0.25 0.25 0.25 10.0 15.0 400.0 0.75
M5 0.25 0.00 0.00 0.75 30.0 85.0 400.0 1.00
M6 0.15 0.15 0.15 0.15 25.0 100.0 350.0 0.60
M7 0.75 0.00 0.25 0.00 10.0 15.0 400.0 1.00
M8 0.15 0.15 0.15 0.15 25.0 50.0 450.0 0.60
M9 0.15 0.55 0.15 0.15 25.0 50.0 350.0 1.00
M10 0.00 0.00 0.00 0.00 0.0 0.0 300.0 0.00
M11 0.75 0.00 0.00 0.00 10.0 85.0 400.0 0.75
M12 0.00 0.00 0.25 0.25 40.0 15.0 500.0 0.50
M13 0.40 0.40 0.40 0.40 25.0 50.0 350.0 1.60
M14 0.15 0.15 0.15 0.00 25.0 50.0 350.0 0.45
M15 0.25 0.25 0.25 0.00 30.0 15.0 300.0 0.75
M16 0.25 0.25 0.25 0.25 25.0 50.0 260.0 1.00
M17 0.00 0.00 0.00 0.00 0.0 0.0 400.0 0.00
M18 0.25 0.00 0.25 0.25 10.0 85.0 400.0 0.75
M19 0.00 0.75 0.00 0.00 10.0 15.0 300.0 0.75
M20 0.00 0.00 0.00 0.00 30.0 85.0 300.0 0.00
M21 0.25 0.25 0.00 0.25 10.0 85.0 300.0 0.75
M22 0.15 0.15 0.15 0.55 25.0 50.0 350.0 1.00
M23 0.15 0.15 0.55 0.15 30.0 50.0 350.0 1.00
M24 0.25 0.00 0.00 0.25 10.0 85.0 300.0 0.50
M25 0.15 0.15 0.15 0.15 0.0 50.0 350.0 0.60
M26 0.00 0.00 0.75 0.25 40.0 85.0 300.0 1.00
M27 0.55 0.15 0.15 0.15 25.0 50.0 350.0 1.00
M28 0.50 0.50 0.50 0.50 25.0 50.0 350.0 2.00
M29 0.25 0.25 0.25 0.00 40.0 85.0 400.0 0.75
M30 0.15 0.35 0.15 0.15 25.0 50.0 350.0 0.80

Note: FA (%) and EAFS (%) represent replacement ratio with cement and crushed stone, respectively.
 M. Balcikanli Bankir, U.K. Sevim / Construction and Building Materials 261 (2020) 119952 5

Fig. 2. Pull out test apparatus (a) and schematic representation (b) of test set up.

is a suitable method for measuring the RF (Fr) and damping coeffi- 3.1. Flexural strength
cient (Q) factor. When measuring in automatic mode, the outputs
fr, fl, fl and Q are measured by the device. Where fr is the resonance The four point flexural strength (FS) test was performed on the
frequency (Hz); fl and fh are low and high frequencies on both 28 days prismatic samples obtained according to the conditions
sides where the vibration amplitude of the fr decreases to 0.707. determined in the experimental design and the test results are
The experimental setup is shown in Fig. 3. The HFC samples were shown in Fig. 4a. Displacement values of the HFC were also mea-
exposed elevated temperatures (250 °C, 500 °C and 750 °C) with sured at the midpoint of the beam sample during the FS test and
a drying oven that temperature rise is 10 °C in a minute. After the results are presented in Fig. 4a together FS. It is seen in
reaching the target temperature, HFC samples were exposed to tar- Fig. 4a that the HFC samples with no fibers (M10, M17 and M20)
get temperature for 2 h and then the HFC samples let to cool to lab- have lower FS and displacement values compared to other HFCs.
oratory temperature. Fiber contribution to a concrete can greatly improve both tough-
ness and flexural strength by the effect of different fiber hybridiza-
2.3.6. Abrasion resistance tion [32,33]. Displacement abilities of concretes are directly
The abrasion resistance of the HFC was determined by using affected from the fiber types and addition ratio Fig. 4b shows the
vertical abrasive device using 100  100  100 mm3 cube samples relationship between displacement-flexural strength and the par-
with the mass losses of the HFC samples according to TS EN 2824 ticipation of different fiber types at varying rates with a total fiber
[29]. The abrasion mass loss of the HFC samples was determined content of 1%. Only steel (M11) and only glass (M19) fibers were
after the abrasion disc rotated 75 cycles. added at a rate of 0.75%, four fiber types were added but one of
them is more than others was added in each mixture (M9, M22,
M23 and M27) and all fiber types were added there equally (M1)
3. Results and discussion
were shown in Fig. 4b. When the graph is examined, it is seen that
HFC, which has the highest toughness (area under the curve), is
According to the results of the experiment, variance analyzes
concrete obtained by low participation of all fiber types of the
(ANOVA) of the findings were also performed with a confidence
mix (M1). In studies, fiber addition in hybrid forms increases the
level of 95%. The results of the experiments are presented both in
engineering properties of concrete as energy absorption capacity,
a simple and comparative manner and the findings of the ANOVA
ductility and durability in comparison with mono fibered concretes
are discussed. According to the CCD, regression models have been
[17,34].
established and which independent variable has an effect on which
When ANOVA of the experimental test results was performed,
dependent variable. Statistically analysis were performed based
the effect of independent variables on the FS was determined.
acceptance on the normally distribution of the residual terms
The 2FI (two factor interaction) regression model that represents
[30,31].
the FS was obtained. More simple equations can be obtained by
eliminating independent variables with an impact rate of <3% (sta-
tistically not significant) in regression models. The regression
model obtained for flexural strength is given in Eq. (1).
FS = 2.8944 + 830.08442*A + 114.72973*B  960.03156*C
þ þ74:872  D þ 1231:3487  A  B
þ 776:44246  A  C  854:54262  A  D 11:86096  A  E
þ 3:92066  A  F  2:57090  A  G  184:52914  B  C
þ 2698:3509  B  D  6:9153  B  E  14:06847  B  F
þ 0:56859  B  G3686:8772  C  D  1:44453  C  E
þ 11:8387  C  F þ 2:51058  C  G
þ 17:68155  D  E1:4399  D  F ð1Þ
Fig. 3. Schematic sketch resonance frequency test set up.
6 M. Balcikanli Bankir, U.K. Sevim / Construction and Building Materials 261 (2020) 119952

25 Flexural Strength (MPa) Displacement (mm)

Flexural Strength and Displacement

20

15

10

0
M1
M2
M3
M4
M5
M6
M7
M8
M9
M10
M11
M12
M13
M14
M15
M16
M17
M18
M19
M20
M21
M22
M23
M24
M25
M26
M27
M28
M29
M30
Mix No a
7 Mix SF GF SNF PF
M1 0,25 0,25 0,25 0,25
M9 0,15 0,55 0,15 0,15
6 M11 0,75 0,00 0,00 0,00
Flexural strength (MPa)

M19 0,00 0,75 0,00 0,00

M22 0,15 0,15 0,15 0,55
5 M23 0,15 0,15 0,55 0,15
M27 0,55 0,15 0,15 0,15

4
M1
M9
3
M11
M19
2
M22
M23
1
M27
0 1 2 3 4 5 6
Deflection (mm) b
Fig. 4. Flexural strength and displacement values of HFC.

Table 5
ANOVA results of HFC for FS.

Dependent parameters Independent parameters Regression model R2 DF SSS MS F value P-value q%

Flexural Strength Constant 2.83944 0.9935 – 36.78 1.360 11.35 0.0841 –
A +830.08442 1 0.39 0.39 3.22 0.2145 3.9
B +114.72973 1 0.44 0.44 3.76 0.1957 4.6
C 960.03156 1 0.44 0.44 3.70 0.1944 4.5
D +74.87200 1 0.50 0.50 4.14 0.1789 5.0
E +0.24042 1 0.016 0.016 0.13 0.7532 0.2
F 0.18247 1 0.024 0.024 0.20 0.7008 0.2
G +0.016501 1 0.032 0.032 0.26 0.6586 0.3
A*B +1231.34870 1 0.46 0.46 3.81 0.1902 4.6
A*C +776.44246 1 0.50 0.50 4.21 0.1768 5.1
A*D 854.54262 1 0.50 0.50 4.15 0.1787 5.0
A*E 11.86096 1 0.44 0.44 3.67 0.1954 4.5
A*F +3.92066 1 0.47 0.47 3.95 0.1851 4.8
A*G 2.57090 1 0.45 0.45 3.72 0.1934 4.5
B*C 184.52914 1 0.44 0.44 3.64 0.1967 4.4
B*D +2698.35090 1 0.41 0.41 3.43 0.2052 4.2
B*E 6.91530 1 0.50 0.50 4.18 0.1776 5.1
B*F 14.06847 1 0.43 0.43 3.58 0.1989 4.3
B*G +0.56859 1 0.54 0.54 4.54 0.1668 5.5
C*D 3686.87720 1 0.43 0.43 3.55 0.2000 4.3
C*E 1.44453 1 0.19 0.19 1.61 0.3325 2.0
C*F +11.83870 1 0.42 0.42 3.49 0.2026 4.2
C*G +2.51058 1 0.41 0.41 3.43 0.2054 4.2
D*E +17.68155 1 0.45 0.45 3.79 0.1910 4.6
D*F 1.43990 1 0.34 0.34 2.81 0.2355 3.4
D*G 0.47812 1 0.17 0.17 1.46 0.3508 1.8
E*F +6.22721E-003 1 0.29 0.29 2.41 0.2604 2.9
E*G 7.03740E-004 1 0.18 0.18 1.50 0.3460 1.8

A: SF, B: GF, C: SNF, D:PF, E:FA, F:EAFS, G:D.

 M. Balcikanli Bankir, U.K. Sevim / Construction and Building Materials 261 (2020) 119952 7

ANOVA results for FS are shown in Table 5. As seen in Table 5, the SF will be used with PF, the SF contribution ratio should be
each independent variable has an effect on the FS. EAFS and D do <0.15% and contribution ratio of PF should be more than 0.1%, as
not have much contribution on FS since the effect ratio (% q) is seen in Fig. 5d. Statistical analysis showed improvement in FS in
around 0.2. In other words, the FS was mostly influenced by the case of hybrid fiber reinforced concrete when compared to individ-
fiber content and type. As can be seen in Table 5, the independent ually steel fiber reinforced concrete [35]. Singh ve Rai stated simi-
variables affect the FS are D, A-C (SF-SNF), A-D (SF-PF), B-E (GF-FA) lar properties of concrete can be accomplished by using smaller
and B-G (GF-D) binary interactions. Binary effects of statistically amounts of hybrid fibers in a concrete mixture as appeared by
significant parameters on the FS are shown in Fig. 5. As seen in standard fiber fortified concrete. Here it has been supported that
Fig. 5a, GF-FA binary interaction affects the FS up to 0.25% GF con- by the method of hybridization fibers, concretes can be produced
tent and 40% FA. According to the content of the contribution ratio with similar properties even at very low fiber contents [17,36].
the FS can be estimated by using the graph. It was monitored that if
the FA content and GF ratio increase together, the FS decreases. It is
3.2. Compressive strength
illustrated in Fig. 5b, while the GF ratio increase, the FS decreases
especially after 0.1% GF. If the SF was used with SNF, the FS getting
A compessive load is applied to determine the compressive
increase for more than 0.15% SNF and 0.05% SF as seen in Fig. 5c. If strength of HFC samples in 28, 180 and 360 days according to

Flexural Strength Flexural Strength

40 400
23.9

16.2 375
30
11.6
G: DOSAGE

2.5
11.6 8.9
E: FA

5.8 3.4 0.4

350
0.4 4.0
5.8
2.5
20 4.03.4
325
16.2

3.4 4.0
10 300
0.00 0.05 0.10 0.15 0.20 0.25 0.00 0.05 0.10 0.15 0.20 0.25

B: Glass Fiber
a B: Glass Fiber
b

Flexural Strength Flexural Strength

0.25 0.25
16.2
16.2
11.6
0.20 0.20 11.6
8.9
D: polypropylene fiber

8.9 0.4
C: Synthetic fiber

6.8
5.8
0.15 5.8 0.15 2.0
4.0
4.03.4 2.5
2.5
0.10 0.4 0.10 2.0
6.8 0.4 2.5
0.05 0.05 4.0
5.8
0.00 0.00
0.00 0.05 0.10 0.15 0.20 0.25 0.00 0.05 0.10 0.15 0.20 0.25

c d
A: Steel fiber A: Steel fiber
Fig. 5. Effect of statistically significant parameters on the flexural strength: (a) GF-FA, (b) GF-D, (c) SF-SNF, (d) SF-PF.
8 M. Balcikanli Bankir, U.K. Sevim / Construction and Building Materials 261 (2020) 119952

50

Compressive Strength (Mpa) 40

30

20

10

0
M4
M5

M9
M1
M2
M3

M6
M7
M8

M10
M11
M12
M13
M14
M15
M16
M17
M18
M19
M20
M21
M22
M23
M24
M25
M26
M27
M28
M29
M30
Mix No 28 180 360
Fig. 6. Compressive strength of HFC in 28, 180 and 360 days.

the mentioned standard above. Experimental results are shown in after a certain temperature value was less than fibreless concrete.
Fig. 6. Since there were no fibers in the mixtures of M10, M17 and A significant part of the losses in strength occurred after exceeding
M20, they were selected as control samples. In these samples the 600 °C and loss in CS was about 70% in fibreless mix, while 60% in
binder dosage, EAFS ratio and FA ratio differ to compare with the fiber-added mixes [40]. The rate of decrease in CS relative to the
other samples. The compressive strength (CS) test is also conducted temperature level is shown in Fig. 9. The highest strength loss was
to samples in 180 and 360 days. The change rate of CS on days 180 observed in M29 sample and the least strength loss was observed
and 360 is calculated on the basis of day 28 and is presented in Fig. 7. in M15 sample. Due to high rate of the fly ash replacement ratio
Here, improvements in time-dependent CS can be observed. When the change of CS increased in M29. This is thought to be due to
the Fig. 7 is examined, it is seen that the CS increase rate of the sam- the fact that fly ash is highly affected by the high temperature,
ples M2, M15 and M28 is highest on both 180 and 360 days (10.51%, and this idea coincides with the literature studies. Rais et al.,
33.5 %and 25.7%, respectively). This is thought to be due to good searched the morphological characteristics of mortar specimens
bonding with the fly ash incorporated matrix and fibers in long term which exposed elevated temperatures up to 500 °C using scanning
period [37,38]. Jiang et al., indicated that strength improvement electron microscopy. It is stated at the end of study mortar contains
with time are significantly dependent on FA content. Incorporation 10% fly ash when exposed up to 500 °C temperatures gives better
of FA can increase or decrease the strength of concretes depending performance. As the fly ash ratio increased, it was found that the
on replacement level and curing time. At the same time the strength loss rate increased [41]. Novak and Kohoutková stated that
polypropylene fiber included hybrid fiber concrete has better CS. hybrid fiber reinforcements have a positive effect on the residual
The reason is explained by Tawfik et al. as monofilament fiber has strength of concretes as the temperature increases [42].
high Young’s modulus and stiffness so they behave more resistant ANOVA results of CS for 28 days HFC samples are shown in
to compressive loads. As it mixes with monofilament type fibers, Table 6. According to ANOVA analysis of the experimental results,
they easily form the network between the cracks and decrease the the effect of independent variables on the CS was determined and
stress concentration at the cracks lines [39]. stated in Table 6. A quadratic model for CS is obtained and given in
HFC samples were exposed elevated temperatures (250, 500 and Eq. (2). To simplify the Eq. (2), impact rate (q %) of<3% are ignored
750 °C) for two hours and after they return to laboratory tempera- in regression model.
ture, the CS of HFC were determined and the test results are given
CS = +36.10 + 18.48*A + 26.57*B  29.25*C  12.97*D
in Fig. 8. The highest point of the ranges indicated in different colors
according to the temperature level shows the final strength at that 10:48  F — 2;55  G25:25  A  C þ 30:99  A  D
temperature. Namely, the highest point of the interval value indi- þ40:64  A  E32:25  A  F þ 34:92  A  G111:97  B  D
cated with yellow represents the strength value at 250 °C. As can þ22:31  B  E þ 155:01  B  F þ 145:18  C
be seen from Fig. 8, the CS of all HFC samples, without exception, 127:81  C  F 35:29  C  G74:08  D  E
decreased depending on the temperature level. It was stated by
þ5:78  D  G 7:49  E  G2:56  A2 ð2Þ
some researchers that the decrease in compressive strength of HFC

40
180 360
Compressive strength change

35
30
ratio by days (%)

25
20
15
10
5
0
M1
M2
M3
M4
M5
M6
M7
M8
M9
M10
M11
M12
M13
M14
M15
M16
M17
M18
M19
M20
M21
M22
M23
M24
M25
M26
M27
M28
M29
M30

-5

Mix No

Fig. 7. Compressive strength change ratio of HFC by days.

 M. Balcikanli Bankir, U.K. Sevim / Construction and Building Materials 261 (2020) 119952 9

50
23 250 500 750

Compressive Strength (MPa) 40

30

20

10

0
M1
M2
M3
M4
M5
M6
M7
M8
M9
M10
M11
M12
M13
M14
M15
M16
M17
M18
M19
M20
M21
M22
M23
M24
M25
M26
M27
M28
M29
M30
Mix No

Fig. 8. Compressive strength of HFC after exposed elevated temperatures.

Mix No
M10
M11
M12
M13
M14
M15
M16
M17
M18
M19
M20
M21
M22
M23
M24
M25
M26
M27
M28
M29
M30
M1
M2
M3
M4
M5
M6
M7
M8
M9
Compressive Strength Change Ratio (%)

0
-10
-20
-30
-40
-50
-60
-70
-80
250 500 750
-90

Fig. 9. Compressive strength change ratio of HFC after exposed elevated temperatures.

Table 6
ANOVA results of HFC for 28 days compressive strength.

Dependent parameters Independent parameters Regression model R2 DF SSS MS Computed value of F distribution P-value q%
Compressive Strength Constant +36.10 0.9976 – 898.09 35.92 67.07 0.0005 –
A +18.48 1 181.41 181.41 338.68 <0.0001 7.5
B +26.57 1 86.48 86.48 161.44 0.0002 3.6
C 29.25 1 129.42 129.42 241.62 0.0001 5.3
D 12.97 1 88.62 88.62 165.44 0.0002 3.6
E +7.01 1 66.66 66.66 124.44 0.0004 2.7
F 10.48 1 120.30 120.30 224.58 0.0001 4.9
G 2.55 1 113.83 113.83 212.51 0.0001 4.7
A*B 42.26 1 21.49 21.49 40.11 0.0032 0.9
A*C 25.25 1 104.53 104.53 195.16 0.0002 4.3
A*D +30.99 1 113.25 113.25 211.44 0.0001 4.7
A*E +40.64 1 101.96 101.96 190.35 0.0002 4.2
A*F 32.25 1 103.20 103.20 192.67 0.0002 4.2
A*G +34.92 1 98.61 98.61 184.10 0.0002 4.1
B*C +4.37 1 70.32 70.32 131.28 0.0003 2.9
B*D 111.97 1 104.00 104.00 194.15 0.0002 4.3
B*E +22.31 1 77.69 77.69 145.05 0.0003 3.2
B*F +155.01 1 108.82 108.82 203.16 0.0001 4.5
C*D +145.18 1 106.01 106.01 197.92 0.0001 4.4
C*F 127.81 1 99.18 99.18 185.15 0.0002 4.1
C*G 35.29 1 102.88 102.88 192.07 0.0002 4.2
D*E 74.08 1 108.22 108.22 202.04 0.0001 4.4
D*F +15.55 1 68.36 68.36 127.63 0.0003 2.8
D*G +5.78 1 78.62 78.62 146.77 0.0003 3.2
E*G 7.49 1 82.93 82.93 154.82 0.0002 3.4
A2 2.56 1 96.53 96.53 180.21 0.0002 4.0

A: SF, B: GF, C: SNF, D:PF, E:FA, F:EAFS, G:D.

10 M. Balcikanli Bankir, U.K. Sevim / Construction and Building Materials 261 (2020) 119952

One Factor
90

28 days Compressive Strength

28 days Compressive Strength

60
90

60

30
30
0.25
0 0.20
0.15 0
400
375 A: Steel
0.10 fiber 0.00 0.25 0.50 0.75 1.00
350 0.05
325
300 0.00 a b
A: Steel fiber
G: DOSAGE
One Factor
90
28 days Compressive Strength

28 days Compressive Strength

90

60 60

30

0 30

0.00
0.05
0.10 0
Synthetic 0.15
fiber 0.05
0.00
0.20 0.10 260 280 300 320 340 360 380 400
0.15
0.20
0.25 0.25
B: Glass Fiber c G: DOSAGE d

One Factor
90
28 days Compressive Strength

28 days Compressive Strength

90

60
60

30
30
0

40
30 300 0
325
20
350 10 20 30 40
E: FA 10 375
0 400 G: DOSAGE f
e E: FA

Fig. 10. Effect of statistically significant parameters on the 28 days compressive strength: (a) SF-D, (b) SF, (c) GF-SNF, (d) D, (e) FA-D, (f) FA.
 M. Balcikanli Bankir, U.K. Sevim / Construction and Building Materials 261 (2020) 119952 11

The single and double interactions of the independent variables fibered matrix has higher supporting capacity so higher pull-out
which have the highest effect ratio (q %) are shown in Fig. 10. resistance is observed. On the other hand, fibers with optimum
Fig. 10a shows the effect of the dual interaction of SF and D on rate help inhibit initiation and growth of micro cracks, thus the for-
CS. Significant increases in CS was found after the steel fiber con- mation of macro cracks delay [47].
tent exceeded 0.15%, binder dosage more than 350 kg/m3. The linear regression model was obtained when the analysis of
Fig. 10b shows the individual effect of the steel fiber on the CS. It variance was performed on the test results obtained from the
is understood that CS decreases after exceeding a certain threshold experimental study (Eq. (3)). The variance analysis of the indepen-
value. Optimum SF contribution rate is around 0.6%. Fig. 10c shows dent variables affecting the pull-out load of the HFCs is shown in
the changes in CS as a result of the dual interaction of SNF and GF. Table 7. When the effect ratios (q%) of the independent variables
It is evident from the Table 6 and Fig. 10c, the usage of SNF and GF on the pull-out load are examined, it is seen that the most effective
together does not have a significant effect rate (2.9%) while the parameter is the binder dosage with the effect rate of 36.1%, fol-
effect rate on the CS is 5.3, while the use of SNF alone. It is known lowed by the glass fiber and steel fiber contribution with the effect
that CS is directly affected by the binder dosage and content rate of 22.6% and 16.5%, respectively. According to the linear
(Fig. 10a, d). The HFC prepared at higher binder dosages gained regression model, it was found that FA and EAFS were not related
higher strength as Fall and Benzaazoua [43] pointed out positive to pull-out load capacity. Since the linear regression model was
effect of increasing binder content is attributed to the formation established, the individual effects of independent variables were
of secondary C–S–H gels as a result of a pozzolanic reaction with examined. Fig. 12 shows how independent variables individually
FA and CH formed during the hydration process [44,45]. As can affecting the pull out capacity. All four types of fibers (SF, GF,
be seen from Fig. 10d and 10f (single effects), the CS strength SNF and PF) increased the pull out load capacity, individually
increases with the addition of fly ash and dosage. It was deter- (Fig. 12a–d). Each of the fibers assumed a separate task and
mined that the CS increased with the increase of both, individually. increased the pull out capacity. For example PF, to bind the
However, when the dual effect of FA-D is examined, it is seen in hydrated particles together to shape into a denser mesh system
Fig. 10e that the amount of fly ash should be increased in HFCs and SF provides additional friction resistance until the fibers are
in low dosage (300 kg/m3) and that the amount of fly ash in HFCs pulled out completelyy [48]. The adherence between reinforce-
in high dosage (400 kg/m3) should be very limited to produce high ment and concrete is also directly affected by the density of the
strength HFC. matrix and the bond between the fibers and concrete matrix. The
bond is affected in mainly three ways as chemical adhesion, fric-
3.3. Pull out capacity tion resistance and mechanical interaction. Straight fibers affect
the bond by only chemical adhesion and friction resistance (GF,
Pull out load is applied to ø14 mm steel bar on the HFC samples SNF and PF), while hooked-end fibers (SF), affect the bond as
and the test results are shown in Fig. 11 and the sample which has mechanical interaction between the matrix and fibers [47]. In this
the same fiber ratio is shown in same color. M10, M17 and M20 case the matrix can be affected from the binder dosage as can be
samples have no fiber so they selected as control samples and indi- seen from Fig. 12e, namely, the increase in binder dosage also
cated with arrows. The sample of the reinforcement debonded increased the pull-out load.
takes place at the lowest load level is M10 and is 226 kg. Fig. 11
also shows the amount of improvement in the pull out capacity Pull out capacity = +427.68 + 39.34*A + 50.90*B + 37.95*C
of the HFC when M10 sample is selected as the reference. The þ 39:73  D þ 95:94  G ð3Þ
HFC named as M28 (2% total fiber ratio) increased the pull out load
75.1% however, the best improvement in pull out load was
observed in M3 sample 73.9% according to reference sample with 3.4. Resonance frequency
a total fiber ratio 1%. Pull out load is improved by the cooperation
of reinforcement-matrix and fibers. It is thought that as a result of Resonance frequency (RF) values of HFC prism samples stayed
the excessive fiber density, the concrete can’t be placed properly in both laboratory conditions (23 °C) and exposed to high temper-
into mold and the bonding surface between the reinforcement atures (250 °C, 500 °C and 750 °C) are shown in Fig. 13 for all HFC
and concrete is reduced. Therefore, it is supported that the bond samples. In Fig. 13, RF of HFC for each temperature level is indi-
between the reinforcement and concrete is weakened [46]. How- cated with a different color and the top level of each color column
ever, in concrete where fiber participation ratio is optimum, the shows the RF at that temperature value. For example, the RF values

1.000 pull out capacity 80

Pull out capacity change ratio (%)

900 pull out capacity change ratio 70

Pull out Capacity (kg)

800
60
700
600 50
500 40
400 30
300
20
200
100 10
0 0
M1
M2
M3
M4
M5
M6
M7
M8
M9
M10
M11
M12
M13
M14
M15
M16
M17
M18
M19
M20
M21
M22
M23
M24
M25
M26
M27
M28
M29
M30

Mix No

Fig. 11. Pull out capacity of HFC.

12 M. Balcikanli Bankir, U.K. Sevim / Construction and Building Materials 261 (2020) 119952

Table 7
ANOVA results of pull out capacity for HFC.

Dependent parameters Independent parameters Regression model R2 DF SSS MS Computed value of F distribution P-value q%
5 5
Pull out capacity Constant +427.68 0.8253 – 8.132*10 1.626*10 22.67 <0.0001 –
A +39.34 1 1.221*105 1.221*105 17.02 0.0004 16.5
B +50.90 1 1.671*105 1.671*105 23.29 <0.0001 22.6
C +37.95 1 87337.45 87337.45 12.17 0.0019 11.8
D +39.73 1 96050.55 96050.55 13.39 0.0012 13.0
G +95.94 1 2.665*105 2.665*105 37.15 <0.0001 36.1

A: SF, B: GF, C: SNF, D:PF, G:D.

of the M1 sample after exposed to 23 °C, 250 °C, 500 °C and 750 °C, loss, caused by not to fill the pores completely with fiber in the
are shown as 3810 Hz, 2845 Hz, 2104 Hz, 1490 Hz, respectively. HFC, increased [51]. Abrasion resistance of concretes with high
Considering the RF values at 23 °C (the upper limit of the columns compressive strength is also high. Namely, abrasion resistance
indicated in red), it is seen that the HFC with the highest RF is M7 increases as compressive strength decreases. It is indicated in stud-
(Fig. 13). When the content of M7 is examined, it can be seen that ies the compressive strength was an important factor which affects
SF’s participation rate is 0.75% (SF contribution ratio is highest the abrasion resistance of concrete [52]. The relationship between
when considering all mixtures). In this case, it can be said that the HFC compressive strength and abrasion rate by mass is shown
RF is directly influenced by the SF ratio, and statistically supported in Fig. 16. As seen in Fig. 16, there is an inverse relationship
by ANOVA with an effect rate 7% (Table 8). These findings are con- between compressive strength and abrasion rate [53].
sistent with the literature. Khalil et al., stated the RF somewhat To obtain the regression model, ANOVA is performed on the
rises with the increase in steel fiber content. The maximum experimental results of abrasion rate of HFC. Linear regression
increase in the RF was noted when the SF content was 0.5%. It is model is obtained from ANOVA and the model is given in Eq. (5).
observed the lowest RF in mixtures where the aggregate was The variance analysis of the independent variables affecting the
replaced with reference aggregate. In M12, EAFS replacement ratio abrasion rate of the HFCs is shown in Table 9. According to the
is 50% and the lowest RF is measured on this specimen [49]. regression model of the parameters affecting the abrasion rate of
After the HFC prism samples were exposed to elevated temper- HFC, the effect rates (q %) of the independent variables were calcu-
atures, the RF was measured again on the same samples and the RF lated and shown in Table 9. According to ANOVA results, FA
change rates are shown in Fig. 14. It was observed that the RF of all replacement ratio (FA) and binder dosage (D) directly affected
HFC samples decreased regardless of the mixture content. Park and the abrasion rate [54]. When the effect rates were taken into con-
Yim have stated the RF of elevated temperature exposed samples sideration, it was seen that the most important independent vari-
decreased considerably with increasing temperature level. Fur- able affecting the abrasion rate was the dosage of the HFC (95.7%),
thermore it is stated there was no effect of mixture properties on while FA effect rate is 4.3%. As can be seen, from Table 9, FA has too
the decreasing trends of the RF [50]. RF values at 23 °C are consid- little effect on abrasion resistance. There are also studies that sup-
ered as reference and the rate of RF change caused by temperature port the abrasion resistance is proportional to the binder dosage
effect is determined. It is understood from Fig. 14 that the RF of the without being affected by the binder content [55].
M1 sample decreases by 25%, 45% and 61% after 250 °C, 500 °C and Abrasion rate = +0.45 + 8.413E-003*E  0.030*G (5)
750 °C, respectively. After exposure to 750 °C, the highest decrease Single effect of the independent variables (FA and D) on the
was observed in M30 (78%), while the least decrease was observed abrasion rate is shown in Fig. 17. As seen in Fig. 17a, abrasion rate
in M16 (51%). increases with increasing FA ratio and abrasion rate decreases with
A quadratic regression model was obtained from the result of increasing D (Fig. 17b). It was determined that only these two
ANOVA of the experimental test results to simplify the eq.4 the parameters (FA and D) were effective on abrasion resistance, so
parameters with an impact rate of <3% were eliminated (Eq. (4)). fiber and aggregate type have no significant effect on abrasion
The variance analysis of the independent variables affecting the resistance.
RF of the HFC is shown in Table 8. When the effect ratios (q %) of
the independent variables and multiple interactions with each
3.6. Multi-objective optimization of HFC production
other on the RF are examined, it is seen that the most effective
parameter is SNF-EAFS dual interaction (9.2%). It is determined
HFC production independent parameters were optimized for
that the SNF-PF and GF-EAFS dual interactions are also have impor-
the maximization of CS, FS, pull out capacity and RF while for the
tant effect on RF (8.9% and 8.6%). It is understood that RF is directly
minimization of abrasion resistance. To optimize the responses
affected by SF and SNF fibers since the high effect rate, individually.
simultaneously, desirability functions (dj) were obtained for each
RF = +3729.02 + 1884.58*A  810.93*C  454.82*G  1045.18
objective or dependent variable with a numerical optimization
*A*B + 491.75*A*D  564.59*A*F  2162.23*B*D + 848.82*B*E + 2
technique by using commercially available statistical software,
367.94*B*F + 2625.90*C*D + 831.69*C*E 1756.46*C*F  370.57*
namely, Design-Expert 7.0.3 [30,56,57]. Multi-objective optimiza-
D*E + 600.96*D*G  904.28*E*G  392.03*B2 (4)
tion studies were performed and for each response lower and
upper values of each independent variable were defined.
3.5. Abrasion resistance Dependent and independent parameters were used in multi-
objective optimization study of current research and different opti-
Abrasion rate of the HFC samples by mass are shown in Fig. 15. mum solutions were proposed. The overall desirability functions of
M12 sample has the lowest (0.36%) abrasion rate by mass while the the proposed optimum solutions varied from 0.748 to 0.504. Opti-
M16 sample has the most (0.5%) abrasion rate by mass. Since the mum solution which has highest desirability value of 0.748 is
binder dosage of the M12 sample is highest and the fiber content selected as optimum proportion for production HFC. The proposed
is lower than M16, the eroded mass is minimal. It is explained that HFC production parameters with highest desirability value and
if the amount of fiber is above a certain level the amount of mass their anticipated test results are given in Table 10.
 M. Balcikanli Bankir, U.K. Sevim / Construction and Building Materials 261 (2020) 119952 13

One Factor One Factor

1000 1000

Pull Out Capacity 800 800

Pull Out Capacity

600 600

400 400

200 200

0.00 0.05 0.10 0.15 0.20 0.25 0.00 0.05 0.10 0.15 0.20 0.25

A: Steel fiber a B: Glass Fiber b

One Factor One Factor
1000 1000

800 800
Pull Out Capacity

Pull Out Capacity

600 600

400 400

200 200

0.00 0.05 0.10 0.15 0.20 0.25 0.00 0.05 0.10 0.15 0.20 0.25

C: Synthetic fiber c D: polypropylene fiber d

One Factor
1000

800
Pull Out Capacity

600

400

200

300 325 350 375 400

G: DOSAGE e
Fig. 12. Effect of statistically significant parameters on the pull out capacity of HFC: (a) SF, (b) GF, (c) SNF, (d) PF, (e) D.
14 M. Balcikanli Bankir, U.K. Sevim / Construction and Building Materials 261 (2020) 119952

6000

Resonance Frequency (Hz)

23 250 500 750
5000

4000

3000

2000

1000

0
M1
M2
M3
M4
M5
M6
M7
M8
M9
M10
M11
M12
M13
M14
M15
M16
M17
M18
M19
M20
M21
M22
M23
M24
M25
M26
M27
M28
M29
M30
Mix No
Fig. 13. Resonance frequencies of HFC due to temperature.

Table 8
ANOVA results of RF for HFC.

Dependent parameters Independent parameters Regression model R2 DF SSS MS Computed value of F distribution P-value q%
Abrasion rate Constant +3729.02 0.9667 – 5.61E+09 2.67E+08 11.05 0.0008 –
A +1884.58 1 2.27E+09 2.27E+09 93.87 <0.0001 7.0
B +10.09 1 2450.79 2450.79 0.10 0.7583 0.0
C 810.93 1 1.54E+09 1.54E+09 63.65 <0.0001 4.8
D 136.29 1 2.57E+08 2.57E+08 10.62 0.0116 0.8
E 37.72 1 3533.02 3533.02 0.15 0.7122 0.0
F +381.07 1 4.43E+08 4.43E+08 18.32 0.0027 1.4
G 454.82 1 1.41E+09 1.41E+09 58.47 <0.0001 4.4
A*B 1045.18 1 1.73E+09 1.73E+09 71.71 <0.0001 5.4
A*D +491.75 1 1.13E+09 1.13E+09 46.56 0.0001 3.5
A*F 564.59 1 1.65E+09 1.65E+09 68.33 <0.0001 5.1
B*C +154.04 1 3.65E+08 3.65E+08 15.10 0.0046 1.1
B*D 2162.23 1 2.73E+09 2.73E+09 112.73 <0.0001 8.4
B*E +848.82 1 2.00E+09 2.00E+09 82.86 <0.0001 6.2
B*F +2367.94 1 2.79E+09 2.79E+09 115.58 <0.0001 8.6
C*D +2625.90 1 2.88E+09 2.88E+09 118.97 <0.0001 8.9
C*E +831.69 1 1.13E+09 1.13E+09 46.86 0.0001 3.5
C*F 1756.46 1 2.97E+09 2.97E+09 122.95 <0.0001 9.2
D* E 370.57 1 2.15E+08 2.15E+08 82.86 0.0176 6.2
D*G +600.96 1 1.05E+09 1.05E+09 43.55 0.0002 3.3
E*G 904.28 1 2.24E+09 2.24E+09 92.63 <0.0001 6.9
B2 392.03 1 1.73E+09 1.73E+09 71.36 <0.0001 5.3

A: SF, B: GF, C: SNF, D:PF, E:FA, F:EAFS, G:D.

Mix No
M10
M11
M12
M13
M14
M15
M16
M17
M18
M19
M20
M21
M22
M23
M24
M25
M26
M27
M28
M29
M30
M1
M2
M3
M4
M5
M6
M7
M8
M9
Resonance Frequency Change Ratio

0
-10
-20
-30
(%)

-40
-50
-60
-70
-80
250 500 750
-90
Fig. 14. Resonance frequency change ratio of HFC due to temperature.
 M. Balcikanli Bankir, U.K. Sevim / Construction and Building Materials 261 (2020) 119952 15

0,6

Abrasion rate by mass (%)

0,5

0,4

0,3

0,2

0,1

0,0
M1
M2
M3
M4
M5
M6
M7
M8
M9
M10
M11
M12
M13
M14
M15
M16
M17
M18
M19
M20
M21
M22
M23
M24
M25
M26
M27
M28
M29
M30
Mix No
Fig. 15. Abrasion rate of HFC by mass.

0,6 50

Compressive Strength (MPa)

Abrasion
Abrasion rate by mass (%)

Comp. Str.
0,5
40

0,4
30
0,3
20
0,2

10
0,1

0,0 0
M1
M2
M3
M4
M5
M6
M7
M8
M9
M10
M11
M12
M13
M14
M15
M16
M17
M18
M19
M20
M21
M22
M23
M24
M25
M26
M27
M28
M29
M30
Mix No
Fig. 16. Relationship between abrasion rate and compressive strength of HFC.

Table 9
ANOVA results of abrasion rate of HFC.

Dependent Independent Regression R2 DF SSS MS Computed value P-value q%

parameters parameters model of F distribution
Abrasion rate Constant +0.45 0.7952 – 0.027 0.014 52.41 <0.0001 –
E +8.413E003 1 1.214 E-003 1.214 E-003 4.68 0.0394 4.3
G 0.030 1 0.027 0.027 103.02 <0.0001 95.7

4. Conclusion 1. It was determined that statistical techniques can be applied in

HFC production and appropriate regression models can be
The influence of fiber type (steel, glass, synthetic and obtained by analysis of variance. Using this technique, it is con-
polypropylene), fly ash replacement ratio, electrical arc furnace cluded that more information can be achieved with less exper-
slag replacement ratio and binder dosage on the flexural and com- imentation, thus saving material, equipment and labor force.
pressive strength, pull out capacity, resonance frequency and abra- 2. Hybridization of fibers in a concrete improved the flexural
sion resistance of hybrid fiber concretes was researched by means strength by combining the properties of fibers. A two factor inter-
of design of experiments. The effects of independent parameters action regression model obtained that represents the FS. Result of
were analyzed with the help of statistically analysis of variance the ANOVA, FS was mostly influenced by the fiber content and
and regression models. To find out the optimum HFC production type. If the steel fiber will be used with polypropylene fiber,
characteristics in the scope of this research for the desired depen- the steel fiber contribution ratio should be <0.15% and
dent parameters mentioned above, a multi-objective optimization polypropylene fiber contribution ratio should be more than 0.1%.
problem was defined and solved based on obtained regression 3. In the ranges of independent variables studied, compressive
models. According to the results, the following conclusions can strength of HFC was in the range of 20.8 to 43.1 MPa. According
be drawn: to the analysis of variance, it was determined that all fiber
16 M. Balcikanli Bankir, U.K. Sevim / Construction and Building Materials 261 (2020) 119952

One Factor One Factor

0.6 0.6

0.5 0.5
Abrasion Rate

Abrasion Rate
0.4 0.4

0.3 0.3

10 20 30 40 300 325 350 375 400

E: FA a G: DOSAGE b
Fig. 17. Effect of statistically significant parameters on the abrasion rate of HFC: (a) FA, (b) D.

Table 10 6. The most effective parameter on the resonance frequency of

Optimum HFC parameters and expected test results for optimum conditions. HFC is SNF-EAFS dual interaction (9.2%). It was observed that
Independent and dependent Optimum values and expected the RF of all HFC samples decreased regardless of the mixture
parameters responses content. After exposure to 750 °C, the highest decrease was
SF (%) 0.02 observed in M30 (78%), while the least decrease was observed
GF (%) 0.23 in M16 (51%).
SNF (%) 0.25 7. Abrasion resistance increases as compressive strength
PF (%) 0.25
decreases. According to ANOVA a linear regression model is
FA (%) 19.95
EAFS (%) 31.64 obtained and as a result it is indicated statistically the most
D (kg/m3) 400 important parameter is binder dosage (D) and secondly FA
Flexural strength (MPa) 6.3 replacement ratio (FA). However, FA replacement with cement
Compressive strength (MPa) 33.96
has too little effect on abrasion resistance.
Pull out capacity (kg) 611
Resonance frequency (Hz) 4763 8. With the multi-objective optimization, the optimum values of
Abrasion rate (%) 0.41 the independent variables were determined as 0.02% SF, 0.23%
GF, 0.25% SNF, 0.25% PF, 19.95% FA, 31.64% EAFS and 400 kg/
m3 binder dosage.
types, the use of EAFS instead of crushed aggregate and the
binding dosage were significant on compressive strength. How-
ever, it was determined that the most effective parameter was
steel fiber content (7.5%). Dual interactions of the independent Declaration of Competing Interest
variables which effects the compressive strength are deter-
mined. CS increase rate of the samples M2, M15 and M28 is The authors declare that they have no known competing finan-
highest on both 180 and 360 days (10.51%, 33.5 %and 25.7%, cial interests or personal relationships that could have appeared
respectively). This is thought to be due to good bonding with to influence the work reported in this paper.
the fly ash incorporated matrix and fibers in long term period.
4. The highest strength loss was observed in M29 sample and the
least strength loss was observed in M15 sample after exposed
Acknowledgements
to 750 °C. Due to high rate of the fly ash replacement ratio,
the change of CS increased in M29. This is thought to be due
Hybrid fiber concrete production was supported by KAAN
to the fact that fly ash is highly affected from elevated
Investment Construction Tourism Trade Inc. and IMFALT Road
temperature.
Building Industry Trade Inc.
5. Maximum total fiber ratio was 2% for the HFC. it was found that
the pulling load in the M28 sample increased the most (75.1%)
When compared to the total fiber content in HFC, the best Author Statement
improvement in pull out load was observed in M3 (73.9%) sam-
ple with a total fiber ratio 1%. It is thought that as a result of the Muzeyyen BALÇIKANLI BANKIR _ performed the experiments,
excessive fiber density, the concrete can’t be placed properly discussed the results and wrote the manuscript.
into mold and the bonding surface between the reinforcement Umur Korkut SEVIM _ discussed the results of the experiments
and concrete is reduced. and helped to write the manuscript.
 M. Balcikanli Bankir, U.K. Sevim / Construction and Building Materials 261 (2020) 119952 17

References [29] TS 2824 EN 1338, 2005. Zemin dösßemesi için beton kaplama blokları - Gerekli
sßartlar ve deney metotlar, Türk Standartları Enstitüsü.
[30] M. Balcikanli, E. Ozbay, Optimum design of alkali activated slag concretes
[1] K. Walter, Strength and deformations of structural concrete subjected to in-
for the low oxygen/chloride ion permeability and thermal conductivity,
plane shear and normal forces, Vol. 234, Birkhäuser, 2013.
Compos. B 91 (2016) 243–256, https://doi.org/10.1016/j.compositesb.2016.
[2] V. Afroughsabet, L. Biolzi, T. Ozbakkaloglu, High-performance fiber-reinforced
01.047.
concrete: a review, J. Mater. Sci. 51 (14) (2016) 6517–6551, https://doi.org/
[31] Design-Expert, Software for design of experiments. USA: Stat-Ease
10.1007/s10853-016-9917-4.
Incorporation, Statistics Made Easy; 2007.
[3] B. Baradan, H. Yazıcı, S.Aydın, Beton, Dokuz Eylül University, Faculty of
[32] A.N. Junior, S.R. Ferreira, F. Toledo, R. D., Fairbairn, J. Dweck, Effect of early age
Engineering publications, 334, 2012.
curing carbonation on the mechanical properties and durability of high initial
[4] A.B. Kizilkanat, N. Kabay, V. Akyüncü, S. Chowdhury, A.H. Akça, Mechanical
strength Portland cement and lime-pozolan composites reinforced with long
properties and fracture behavior of basalt and glass fiber reinforced concrete:
sisal fibres, Composites Part B: Engineering, 163 (2019) 351-362. https://doi.
an experimental study, Constr. Build. Mater. 100 (2015) 218–224, https://doi.
org/10.1016/j.compositesb.2018.11.006.
org/10.1016/j.conbuildmat.2015.10.006.
[33] G. Xue, E. Yilmaz, W. Song, S. Cao, Mechanical, flexural and microstructural
[5] M. Madhkhan, R. Katirai, Effect of pozzolanic materials on mechanical
properties of cement-tailings matrix composites: Effects of fiber type and
properties and aging of glass fiber reinforced concrete, Constr. Build. Mater.
dosage, Compos. B Eng. 172 (2019) 131–142, https://doi.org/10.1016/
225 (2019) 146–158, https://doi.org/10.1016/j.conbuildmat.2019.07.128.
j.compositesb.2019.05.039.
[6] L. Fenu, D. Forni, E. Cadoni, Dynamic behavior of cement mortars reinforced
[34] H.R. Pakravan, M. Latifi, M. Jamshidi, Hybrid short fiber reinforcement system
with glass and basalt fibers, Compos. B 92 (2016) 142–150, https://doi.org/
in concrete: a review, Constr. Build. Mater. 142 (2017) 280–294, https://doi.
10.1016/j.compositesb.2016.02.035.
org/10.1016/j.conbuildmat.2017.03.059.
[7] M.M. Hilles, M.M. Ziara, Mechanical behavior of high strength concrete
[35] P.S. Song, J.C. Wu, S. Hwang, B.C. Sheu, Statistical analysis of impact strength
reinforced with glass fiber reinforced polymer, J. Eng. Res. Technol. 5 (3)
and strength reliability of steel-polypropylene hybrid fiber-reinforced
(2018) 58–67, https://doi.org/10.1016/j.jestch.2019.01.003.
concrete, Constr. Build. Mater. 19 (1) (2005) 1–9, https://doi.org/10.1016/
[8] K.G. Kuder, S.P. Shah, Processing of high-performance fiber-reinforced cement-
j.conbuildmat.2004.05.002.
based composites, Constr. Build. Mater. 24 (2) (2010) 181–186, https://doi.org/
[36] N. Banthia, R. Gupta, Hybrid fiber reinforced concrete (HyFRC): Fiber synergy
10.1016/j.conbuildmat.2007.06.018.
in high strength matrices, Mater. Struct. Mater. Construct. 37 (274) (2004)
[9] A. Lau, M. Anson, Effect of high temperatures on high performance steel fibre
707–716, https://doi.org/10.1617/14095.
reinforced concrete, Cem. Concr. Res. 36 (9) (2006) 1698–1707, https://doi.org/
[37] M. Afroz, S. Venkatesan, I. Patnaikuni, Effects of hybrid fibers on the
10.1016/j.cemconres.2006.03.024.
development of high volume fly ash cement composite, Constr. Build. Mater.
[10] F. Soltanzadeh, J.A.O. Barros, R.F.C. Santos, High performance fiber reinforced
215 (2019) 984–997, https://doi.org/10.1016/j.conbuildmat.2019.04.083.
concrete for the shear reinforcement: experimental and numerical research,
[38] H. Jiang, H. Yi, E. Yilmaz, S. Liu, J. Qiu, Ultrasonic evaluation of strength
Constr. Build. Mater. 77 (2015) 94–109, https://doi.org/10.1016/
properties of cemented paste backfill: effects of mineral admixture and curing
j.conbuildmat.2014.12.003.
temperature, Ultrasonics 100 (2020), https://doi.org/10.1016/j.ultras.2019.
[11] K.H. Mo, K.K.Q. Yap, U.J. Alengaram, M.Z. Jumaat, The effect of steel fibres on
105983 105983.
the enhancement of flexural and compressive toughness and fracture
[39] K. Tawfiq, A. Jamshid, R. Rodolfo, Fatigue cracking of polypropylene fiber
characteristics of oil palm shell concrete, Constr. Build. Mater. 55 (2014) 20–
reinforced concrete, Mater. J. 96 (2) (1999) 226–233.
28, https://doi.org/10.1016/j.conbuildmat.2013.12.103.
[40] A.K. Sahani, A.K. Samanta, D.K. Singharoy, Mechanical behaviour of fire-
[12] F. Koksal, F. Altun, I. Yiğit, Y. Sßahin, Combined effect of silica fume and steel
exposed fibre-reinforced sustainable concrete, Journal of Structural Fire
fiber on the mechanical properties of high strength concretes, Constr. Build.
Engineering 10 (4) (2019) 482–503, https://doi.org/10.1108/JSFE-11-2018-
Mater. 22 (8) (2008) 1874–1880, https://doi.org/10.1016/
0035.
j.conbuildmat.2007.04.017.
[41] M.S. Rais, M. Shariq, A. Masood, A. Umar, M.M. Alam, An experimental and
[13] S.H. Park, D.J. Kim, G.S. Ryu, K.T. Koh, Tensile behavior of ultra high
analytical investigation into age-dependent strength of fly ash mortar at
performance hybrid fiber reinforced concrete, Cem. Concr. Compos. 34 (2)
elevated temperature, Constr. Build. Mater. 222 (2019) 300–311, https://doi.
(2012) 172–184, https://doi.org/10.1016/j.cemconcomp.2011.09.009.
org/10.1016/j.conbuildmat.2019.06.167.
[14] W. Yao, J. Li, K. Wu, Mechanical properties of hybrid fiber-reinforced concrete
[42] J. Novák, A. Kohoutková, Fire response of hybrid fiber reinforced concrete to
at low fiber volume fraction, Cem. Concr. Res. 33 (1) (2003) 27–30, https://doi.
high temperature, Procedia Eng. 172 (2017) 784–790, https://doi.org/10.1016/
org/10.1016/S0008-8846(02)00913-4.
j.proeng.2017.02.123.
[15] D. He, M. Wu, P. Jie, Study on mechanical properties of hybrid fiber reinforced
[43] M. Fall, M. Benzaazoua, Modeling the effect of sulphate on strength
concrete, IOP Conf. Ser. Earth Environ. Sci. 100 (1) (2017).
development of paste backfill and binder mixture optimization, Cem. Concr.
[16] W.A.W. Jusoh, I.S. Ibrahim, A.R.M. Sam, Flexural behaviour of reinforced
Res. 35 (2) (2005) 301–314, https://doi.org/10.1016/j.cemconres.2004.05.020.
concrete beams with discrete steel–polypropylene fibres, MATEC Web Conf.
[44] K.M.A. Hossain, M. Lachemi, Performance of volcanic ash and pumice based
101 (2017) 01020, https://doi.org/10.1051/matecconf/201710101020.
blended cement concrete in mixed sulfate environment, Cem. Concr. Res. 36
[17] N.K. Singh, B. Rai, A review of fiber synergy in hybrid fiber reinforced concrete,
(6) (2006) 1123–1133, https://doi.org/10.1680/adcr.2006.18.2.71.
J. Appl. Eng. Sci. 8 (2) (2018) 41–50, https://doi.org/10.2478/jaes-2018-0017. _ Alp, Cemented paste backfill of
[45] B. Ercikdi, A. Kesimal, F. Cihangir, H. Deveci, I.
[18] M. Pourbaba, E. Asefi, H. Sadaghian, A. Mirmiran, Effect of age on the
sulphide-rich tailings: Importance of binder type and dosage, Cem. Concr.
compressive strength of ultra-high-performance fiber-reinforced concrete,
Compos. 31 (4) (2009) 268–274, https://doi.org/10.1016/j.cemconcomp.2009.
Constr. Build. Mater. 175 (2018) 402–410, https://doi.org/10.1016/
01.008.
j.conbuildmat.2018.04.203.
[46] H.T. Turker, E. Ozbay, M. Balcıkanlı, Pullout capacity development of cast in
[19] C.K. Mahapatra, S.V. Barai, Hybrid fiber reinforced self-compacting concrete
place anchors with embedded studs, Constr. Build. Mater. 102 (2016) 39–43,
with fly ash and colloidal nano silica: a systematic study, Constr. Build. Mater.
https://doi.org/10.1016/j.conbuildmat.2015.10.166.
160 (2018) 828–838, https://doi.org/10.1016/j.conbuildmat.2017.11.131.
[47] F. Deng, X. Ding, Y. Chi, L. Xu, L. Wang, The pull-out behavior of straight and
[20] M. Demirel, B. Kayan, Application of response surface methodology and central
hooked-end steel fiber from hybrid fiber reinforced cementitious composite:
composite design for the optimization of textile dye degradation by wet air
experimental study and analytical modeling, Compos. Struct. 206 (2018) 693–
oxidation, Int. J. Ind. Chem. 3 (1) (2012) 24.
712, https://doi.org/10.1016/j.compstruct.2018.08.066.
[21] A. Özer, G. Gürbüz, A. Çalımlı, B.K. Körbahti, Biosorption of copper (II) ions on
[48] L. Xu, F. Deng, Y. Chi, Nano-mechanical behavior of the interfacial transition
Enteromorpha prolifera: application of response surface methodology (RSM),
zone between steel-polypropylene fiber and cement paste, Constr. Build.
Chem. Eng. J. 146 (2009) 377–387.
Mater. 145 (2017) 619–638, https://doi.org/10.1016/j.conbuildmat.2017.
[22] Parampalli, A., Eskridge, K., Smith, L., Meagher, M., Mowry, M., Subramanian,
04.035.
A., (2007) Development of serum-free media in CHO-DG44 cells using a
[49] W.I. Khalil, W.A. Abbas, I.F. Nasser, Dynamic modulus of elasticity of
central composite statistical design. Cytothecnology, 54, 57–68.
geopolymer lightweight aggregate concrete, IOP Conf. Ser. Mater. Sci. Eng.
[23] M. Sonebi, M.T. Bassuoni, Investigating the effect of mixture design parameters
518 (2) (2019) 022023.
on pervious concrete by statistical modeling, Constr. Build. Mater. 38 (2013)
[50] S.J. Park, H.J. Yim, Evaluation of residual mechanical properties of concrete
147e54, https://doi.org/10.1016/j.conbuildmat.2012.07.044.
after exposure to high temperatures using impact resonance method, Constr.
[24] M. Balcikanli, H.T. Turker, E. Ozbay, O. Karahan, C.D. Atis, Identifying the bond
Build. Mater. 129 (2016) 89–97, https://doi.org/10.1016/j.conbuildmat.2016.
and abrasion behavior of alkali activated concretes by central composite
10.116.
design method, Constr. Build. Mater. 132 (2017) 196–209, https://doi.org/
[51] M. Gupta, M. Kumar, Effect of nano silica and coir fiber on compressive
10.1016/j.conbuildmat.2016.10.034.
strength and abrasion resistance of concrete, Constr. Build. Mater. 226 (2019)
[25] TS EN 12350-2 Testing fresh concrete - Part 2: Slump-test.
44–50, https://doi.org/10.1016/j.conbuildmat.2019.07.232.
[26] ASTM C1161, 2002. Standard Test Method for Flexural Strength of Advanced
[52] A.M. Rashad, H.E.D.H. Seleem, A.F. Shaheen, Effect of silica fume and slag on
Ceramics at Ambient Temperature. ASTM International, West Conshohocken,
compressive strength and abrasion resistance of HVFA concrete, Int. J. Concr.
PA.
Struct. Mater. 8 (1) (2014) 69–81, https://doi.org/10.1007/s40069-013-0051-
[27] ASTM C 39, 1994. Standard test method for compressive strength of cylindrical
2.
concrete specimens. Annual Book of ASTM Standards.
[53] A.M. Rashad, Y. Bai, P.A.M. Basheer, N.B. Milestone, N.C. Collier, Hydration and
[28] ASTM C 215, 2003. Standard Test Method for Fundamental Transverse,
properties of sodium sulfate activated slag, Cem. Concr. Compos. 3 (7) (2013)
Longitudinal, and Torsional Resonant Frequencies of Concrete Specimens,
20–29, https://doi.org/10.1016/j.cemconcomp.2012.12.010.
ASTM International, West Conshohocken, PA.
18 M. Balcikanli Bankir, U.K. Sevim / Construction and Building Materials 261 (2020) 119952

[54] O. Gencel, M.S. Gok, W. Brostow, Effect of metallic aggregate and cement [56] R.H. Myers, D.C. Montgomery, Response surface methodology process and
content on abrasion resistance behaviour of concrete, Mater. Res. Innovat. 15 product optimization using designed experiments, John Willey and Sons, 2002.
(2) (2011) 116–123, https://doi.org/10.1179/143307511X12998222918877. [57] Bankir, Muzeyyen Balcikanli, and Umur Korkut Sevim. ‘‘Performance
[55] K. Turk, M. Karatas, Abrasion resistance and mechanical properties of self- optimization of hybrid fiber concretes against acid and sulfate attack.”
compacting concrete with different dosages of fly ash/silica fume, 18 (1) Journal of Building Engineering (2020): 101443.
(2011) 49-60. http://nopr.niscair.res.in/handle/123456789/11206.