Construction and Building Materials 135 (2017) 343351

Contents lists available at ScienceDirect

Construction and Building Materials

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

A Taguchi approach for investigating the engineering properties of

concretes incorporating barite, colemanite, basaltic pumice and ground
blast furnace slag
Ahmet Hayrullah Sevin a,, Muhammed Yasin Durgun b, Mustafa Eken a
a _
Kahramanmaras St Imam University, Elbistan Vocational School of Higher Education, Construction Program, 46340 Kahramanmaras, Turkey
b
Bartn University, Engineering Faculty, Department of Civil Engineering, Kutlubey-Yazclar Campus, 74110 Bartn, Turkey

h i g h l i g h t s

Taguchi method used for reducing the number of samples.

Effectiveness of the mineral additives determined by Taguchi method.

Barite is the most effective parameter at early ages.

BFS is the most effective parameter at later ages.

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

Article history: Natural and artificial mineral additives have been frequently used for improving the physical, mechanical
Received 28 June 2016 and durability properties of concrete in recent years. In this study four different mineral additives were
Received in revised form 25 December 2016 used which are barite, colemanite, basaltic pumice and ground granulated blast furnace slag (BFS). Each
Accepted 30 December 2016
mineral additive was used one by one and binary, ternary and quaternary combinations. Utilization of
Available online 10 January 2017
these mineral additives as fine aggregates and cement replacement material was investigated and the
effect of these materials on compressive strength, splitting tensile strength, ultrasonic sound velocity,
Keywords:
capillary water absorption and Bhme abrasion resistance were examined. The Taguchi method was also
Barite
Colemanite
used to determine the optimum working conditions for each parameter and to reduce the number of
Pumice experiments. Besides, contribution of each material on test results was also determined by ANOVA.
Blast furnace slag According to the analysis, it was seen that blast furnace slag and colemanite were the two most effective
Concrete parameters on results in general.
Engineering properties 2017 Elsevier Ltd. All rights reserved.
Taguchi method

1. Introduction
Concrete is the most commonly used construction material in
the world. It became more common compared to the other build-
ing materials due to its features of being economic, easy formable,
more durable, consuming less energy during production, being
aesthetic and easily producible everywhere aesthetic. Neverthe-
less, in some cases conventional concrete remains incapable.
Today, not only the strength of the concrete but also its durability
is required. In this case, the durability related concrete design
comes to the forefront. The type of concrete which provides both
Corresponding author.
E-mail addresses: ahsevinc46@gmail.com (A.H. Sevin), mydurgun@gmail.com
(M.Y. Durgun), meken@ksu.edu.tr (M. Eken).
high strength and high durability is called high performance con-
crete [1].
High performance concrete production is generally provided by
using natural and artificial pozzolans such as fly ash, silica fume,
blast furnace slag, metakaolin and various organic ashes [1].
A serious number of recent studies stated that using ground
granulated blast furnace slag both as fine aggregates and cement
substitution admixture had successful results in terms of mechan-
ical, durability and microstructure properties [25]. Also the con-
venience of basaltic pumice as fine aggregate in concrete was
investigated and different engineering properties were tested by
various researchers. The results showed that basaltic pumice incor-
porating mixtures have better workability and abrasion resistance
than conventional concrete. Also pumice aggregate used samples
have a superior shock absorbing property. Research results showed
that basaltic pumice as fine aggregates could be used for produc-

http://dx.doi.org/10.1016/j.conbuildmat.2016.12.209
0950-0618/ 2017 Elsevier Ltd. All rights reserved.
344 A.H. Sevin et al. / Construction and Building Materials 135 (2017) 343351

tion of lightweight concrete [4,69]. It is also known that pumice is Table 2

used for producing other construction materials such as bricks Physical properties of aggregates.

[10]. Property Sand (0 411 mm 1122 mm

On the other hand, it is known that using granulated blast fur- 4 mm) Aggregate Aggregate
nace slag and basaltic pumice as cement substitution material has Specific gravity (g/cm3) 2.65 2.74 2.73
an increasing effect on durability and mechanical properties [11 Compressed unit weight (g/cm3) 1.94 1.60 1.57
13]. Loose unit weight (g/cm3) 1.74 1.59 1.57
Water absorption (%) 3.41 1.21 0.40
Generally the aggregates which are used for heavyweight con-
crete production are either natural aggregates such as barite, limo-
nite, magnetite containing iron ore or artificial aggregates such as
iron and lead particles. The most important difference between
conventional concrete and heavyweight concrete is the unit weight aggregates 411 and 1122 mm sized river aggregates were used.
[1416]. Investigations showed that heavyweight concretes con- The grain distribution and the physical properties of the aggregates
taining barite had better freeze thaw and radiation resistance are shown Table 2 respectively.
[17,18].
In recent years, extensive researches were carried out about
how to use boron, boron minerals (colemanite, kernite, ulexite 2.2. Design of experiments
etc.) or boron wastes in concrete. The economic value of the inter-
mediate products produced from boron minerals is quite high. The design of the experiments was carried out by using the
Researches which used colemanite wastes as concrete admixture Taguchi technique. By this way, it was possible to achieve more
showed that they prevented the strength gaining but they signifi- extensive results with less number of experiments. Therefore the
cantly increased the compressive strength even when they were time and cost saving was ensured. Experimental study can be
used in small amounts [19]. designed easily by obtaining the optimum working conditions of
Turkey has significant amount of boron, barite and pumice the parameters affecting the results with the help of the Taguchi
reserves. These products are considered as high competitiveness method [20]. A systematic approach is needed to apply the Taguchi
products. At the same time as a result of the developing industry method to the solution of problems. A systematic and efficiency
high amounts of industrial by-products and wastes emerge. In this approach for this purpose is given in Fig. 1.
study the compressive strength, splitting tensile strength, ultra- The operations of the flowchart can be grouped into 13 opera-
sonic sound velocity, capillary water absorption and Bhme abra- tional steps for achieving design optimization. These steps are
sion resistance of concrete samples were studied including the followings:
various amounts of pumice, barite, blast furnace slag and
colemanite.
Determining of the problem and organizing the team.

Determining the performance characteristic(s) and the measur-
ing system.
2. Experimental study
Determining the variables (parameters) affecting the perfor-
mance characteristic(s).
2.1. Materials
Conducting the screening design.

Determining the interactions to be examined.
TS EN 197-1 CEM I 42.5 type cement produced in Turkey was
Selecting appropriate OAs and assigning the variables to the
used for the study. Barite was obtained from Barite Mine Turk suitable columns.
Inc. in Osmaniye, colemanite was selected from ETI Mine Works,
Determining the loss function(s) and performance statistics.
pumice was obtained from Adana Cement and blast furnace slag
Conducting experiments and recording the results.
was selected from ISDEMIR (Iskenderun Iron and Steel Works
Analyzing data and selecting the optimum value of the control-
Co.). All materials were obtained as grounded from the suppliers. lable variables.
The chemical and physical properties of the materials used in the
Testing the results.
study are given in Table 1. Aksu river aggregate was used as natural
Establishing tolerance design.
aggregate. As fine aggregate 04 river sand was used and as coarse
Evaluation, implementation and observation [21,22].

Table 1
Chemical and physical analysis results of materials.

Compounds (%) Cement Basaltic Pumice Blast Furnace Slag Colemanite Barite
Chemical analysis results
SiO2 18.63 47.63 43.71 4 4
Al2O3 5.38 15.99 11.14 0.4 0.84
Fe2O3 2.80 11.24 1.21 0.08 0.7
CaO 64.39 9.65 32.28 26 1.7
MgO 3.73 8.05 8.42 3 0.3
SO3 2.63
Na2O + K2O 0.62 0.35 0.8
SrSO 1.5 0.65
B2O3 40
BaSO4 89.56
MnO 0.42
Loss of ignition (%) 0.58 0.73 24.6 1.03
Physical analysis results
Specific Weight (g/cm3) 3.20 2.91 2.86 2.47 3.95
Fineness 3497 cm2/g 3710 cm2/g 5500 cm2/g 75 lm 75 lm
 A.H. Sevin et al. / Construction and Building Materials 135 (2017) 343351 345

Determination of
the problem and 1
organising the team

Determination of the loss

Determination of the
function(s) and the
performance characteristic (s)
performance statistic (s)
and the measuring system

2
Conducting experiments
No Are there
and recording the results
required measurement
system?

Yes 2
Analysing the data and selection
of the optimum value of the
Determination of the
variables affecting controllable variables
performance characteristic(s)

Test the results

Conducting screening
design and determining Yes Are The
the important controllable variables
controllable variables too many ?

No Does the
No searching the causes
performance statistic (s)
and taking the
fall into the confidence necessary measurement
Determination the number
interval?
of levels and values of
the controllable and the Yes
uncontrollable variables

Is the
improvement good
Yes
Determination of the enough ?
interactions which
will be studied
No

Selection of appropriate Tolerance design

orthogonal arrays and
assigning the variables to
the appropriate columns

Is the No
improvement good
Decrease the number of No Is the enough ?
variables/levels and/or experimental cost
acceptable level ?
select smaller design model
Yes
Yes

Evaluation, implementation
1 and observation

Fig. 1. Flowchart of the systematic approach to the application of Taguchi methods [21,22].

The mechanical and physical properties of hardened concrete to understand the effects of mineral admixtures on properties of
depend on many different parameters such as components, manu- concrete, all other conditions were kept constant. For Taguchi
facturing process and curing conditions. In this research, in order design, mineral additives were selected as the factors and the pro-
346 A.H. Sevin et al. / Construction and Building Materials 135 (2017) 343351

Table 3 portions of these additives were selected as the levels, which are
Experimental factors and level values. shown in Table 3. Only 16 experiments were needed for the
Factor Level 1 Level 2 Level 3 Level 4 research according to L16 (44) orthogonal array Taguchi design
A, Pumice (%) 0 2.5 5 10 (Table 4.)
B, Barite (%) 0 2.5 5 10
C, BFS (%) 0 2.5 5 10 2.3. Mix proportions
D, Colemanite (%) 0 0.25 0.5 0.75

The mix proportions are given in Table 5. Barite, basaltic pumice

and blast furnace slag were used instead of fine aggregates and
Table 4
A four-level (L16) orthogonal array. cement in weight ratio of 5%, 10% and 20% while colemanite was
used in ratio of 0.25%, 0.5% and 0.75%. Colemanite proportions
Trial number Factors and their levels
were chosen differently than other admixtures due to its damaging
A B C D E effect on the hydration process of cement when it is used more
1 1 1 1 1 1 than 2% [23]. Some researcher showed that using high amount of
2 1 2 2 2 2 colemanite decreases the mechanical properties of concretes
3 1 3 3 3 3
[24]. Substitution proportion of each material was divided into
4 1 4 4 4 4
5 2 1 2 3 4 two parts and half of the materials were substituted from cement
6 2 2 1 4 3 and the other half of the materials was substituted from fine aggre-
7 2 3 4 1 2 gate. In the sample codes, pumice, barite, blast furnace slag and
8 2 4 3 2 1 colemanite denoted as P, B, S and C, respectively. For exam-
9 3 1 3 4 2
10 3 2 4 3 1
ple, PBC indicates the sample prepared with pumice, barite and
11 3 3 1 2 4 colemanite.
12 3 4 2 1 3
13 4 1 4 2 3 2.4. Production of specimens
14 4 2 3 1 4
15 4 3 2 4 1
16 4 4 1 3 2 The reference mixture was prepared according to TS 802 [25]
and the other mixtures were prepared according to L16 design

Table 5
Mix proportions of samples.

Trial Sample Code A B C D Cement (kg/m3) Water (kg/m3) Aggregate (kg/m3) Sand (kg/m3)
1 R 0 0 0 0 320 160 1151 760
2 BSC1 0 2.5 2.5 0.25 303.2 160 1151 720.1
3 BSC2 0 5 5 0.50 286.4 160 1151 680.2
4 BSC3 0 10 10 0.75 253.6 160 1151 602.3
5 PSC1 2.5 0 2.5 0.50 302.4 160 1151 718.2
6 PBC1 2.5 2.5 0 0.75 301.6 160 1151 716.3
7 PBS1 2.5 5 10 0 264 160 1151 627
8 PBSC1 2.5 10 5 0.25 263.2 160 1151 625.1
9 PSC2 5 0 5 0.75 285.6 160 1151 678.3
10 PBSC2 5 2.5 10 0.50 262.4 160 1151 623.2
11 PBC2 5 5 0 0.25 287.2 160 1151 682.1
12 PBS2 5 10 2.5 0 264 160 1151 627
13 PSC3 10 0 10 0.25 255.2 160 1151 606.1
14 PBS3 10 2.5 5 0 264 160 1151 627
15 PBSC3 10 5 2.5 0.75 261.6 160 1151 621.3
16 PBC3 10 10 0 0.50 254.4 160 1151 604.2

Table 6
Ultrasonic sound velocity, compressive strength, splitting tensile strength, slump and unit weight results of samples.

Ultrasonic Sound Compressive Strength Splitting Tensile Slump (cm) Unit weight (kg/m3)
Velocity (m/s) (MPa) Strength (MPa)
7 days 28 days 7 days 28 days 7 days 28 days
R 5342 5430 23.61 33.61 7.96 9.06 5.4 2385.3
BSC1 5397 5557 27.34 37.15 8.47 9.49 5.1 2391.7
BSC2 5138 5575 21.85 38.03 6.98 10.04 4.6 2397.1
BSC3 5013 5592 18.78 41.87 5.96 11.07 3.6 2401.3
PSC1 5358 5568 26.89 39.18 8.15 10.18 5.2 2389.4
PBC1 5349 5557 24.15 37.27 8.00 9.71 5.6 2393.3
PBS1 5343 5560 23.63 37.69 7.96 9.89 3.3 2337.6
PBSC1 5330 5466 23.18 33.99 7.73 9.25 4.8 2419.7
PSC2 5214 5581 22.23 40.78 7.31 10.63 4.2 2342.6
PBSC2 5305 5583 22.87 41.33 7.44 10.68 4.1 2414.8
PBC2 5136 5393 21.14 32.72 6.95 8.73 4.5 2341.9
PBS2 5078 5376 20.86 31.45 6.53 8.04 4.6 2354.3
PSC3 5279 5569 22.55 40.07 7.41 10.21 3.1 2344.2
PBS3 5274 5498 23.96 35.25 8.01 9.66 4.2 2348.7
PBSC3 4923 5461 18.10 38.06 5.63 10.08 4.4 2408.2
PBC3 4727 5284 16.78 28.85 4.87 7.33 4.6 2394.8
 A.H. Sevin et al. / Construction and Building Materials 135 (2017) 343351 347

given in Table 5. Cement, aggregates, sand and admixtures were ity and capillary water absorption tests. For Bhme abrasion
mixed dry in a 50 dm3 laboratory type concrete mixture, then resistance test 100 mm cube samples were produced and they
water was added to mixture and it was mixed for 7.5 min including were cut into 71 mm cubes before the test according to DIN
two stops for 90 s each. 52108. [26].
The specimens were cast in cubes of 100 mm length for com- All samples were cured in a temperature controlled water tank
pressive strength, splitting tensile strength, ultrasonic sound veloc- at 2123 C until the 7th and 28th test day.

0.6

0.5148
0.4881
0.4766
0.4414
0.5
0.4225

0.4176
Capillary Water Absorption (cm/s1/2)

0.3857

0.3788
0.3739

0.3729
0.3698

0.3705
0.3613

0.3545
0.4

0.3269
0.3
0.2246

0.2

0.1

0.0
R BSC1 BSC2 BSC3 PSC1 PBC1 PBS1 PBSC1 PSC2 PBSC2 PBC2 PBS2 PSC3 PBS3 PBSC3 PBC3

Mixtures

Fig. 2. Capillary water absorption results (After 24 h).

6.98
6.73
6.56
6.47

7
6.28

6.17
6.03

6.01

5.88

5.87
5.81

5.75

5.58

6
5.31

5.23
5.04
Bhme Abrasion (cm3/50cm2)

0
R BSC1 BSC2 BSC3 PSC1 PBC1 PBS1 PBSC1 PSC2 PBSC2 PBC2 PBS2 PSC3 PBS3 PBSC3 PBC3

Mixtures
Fig. 3. Mass loss results due to Bhme abrasion.
348 A.H. Sevin et al. / Construction and Building Materials 135 (2017) 343351

2.5. Testing procedure 3. Results and discussion

The slump test [27] and the specific bulk density test [28] were
performed on fresh concrete. Ultrasonic sound velocity tests [29]
were performed on hardened concrete samples. Before testing
ultrasonic sound velocity, cured samples were dried in a laboratory
oven for minimum 18 h, at 105 C to a constant weight. After the
ultrasonic sound velocity test, compressive strength [30] and split-
ting tensile strength [31] tests were performed to the same sam-
ples. It is known that the ultrasonic sound velocity is affected by
the same properties affecting other mechanical properties. It
depends on the density and the other elastic properties of the con-
crete. Therefore, there is a logic correlation between the mechani-
cal properties and the ultrasonic sound velocity [32]. Also this test
provides information about the porosity of sample.
The capillary water absorption tests [33] were performed on the
samples which had coated sides with water insulating materials.
Bhme abrasion resistance test were performed on the 71 mm
length cube samples [26].
3.1. Ultrasonic sound, compressive and splitting tensile strength results
As fresh state tests slump and bulk density test results and as
hardened state tests results of ultrasonic sound velocity, compres-
sive strength and splitting tensile strength at 7th and 28th days
were given in Table 6.
3.2. Capillary water absorption and wear resistance test results
The capillary water absorption test results and Bhme abrasion
results were illustrated in Figs. 2 and 3 respectively.
3.3. Statistical results
The collected data was analyzed by using the Taguchi method
and average effects calculated at each level of factors are given in
Table 7. In the Taguchi method signal and noise for output

Table 7
Mean S/N ratios for each level of each factor.

Factors Mean S/N Ratios (dB)

Level 1 Level 2 Level 3 Level 4
Compressive Strength (7 days)
A (Pumice)b 27.12 27.51* 26.53 27.23
B (Barite)a 27.76 27.79* 27.22 27.40
C (BFS)d 26.75 26.48 27.15* 26.77
D (Colemanite)c 26.08 25.91 26.80* 26.31
Compressive Strength (28 days)
A (Pumice)d 31.49 31.66* 30.36 30.74
B (Barite)c 31.36 31.52* 31.21 31.09
C (BFS)a 31.20 31.26 31.35* 31.25
D (Colemanite)b 30.95 30.56 32.09* 31.92
Splitting Tensile Strength (7 days)
A (Pumice)b 17.24 17.73* 16.67 17.60
B (Barite)a 18.02 18.03* 17.02 17.64
C (BFS)d 16.96 16.69 17.50* 16.57
D (Colemanite)c 16.06 15.83 17.09* 16.46
Splitting Tensile Strength (28 days)
A (Pumice)d 19.90 20.00* 18.75 19.21
B (Barite)b 19.78 19.89* 19.47 19.47
C (BFS)a 19.51 19.71 19.90* 19.52
D (Colemanite)c 19.31 18.90 20.38* 20.31
Ultrasonic Sound Velocity (7 days)
A (Pumice)b 74.35 74.48* 74.21 74.42
B (Barite)a 74.56* 74.54 74.30 74.46
C (BFS)d 74.29 74.21 74.38* 74.20
D (Colemanite)c 74.06 74.04 74.38* 74.19
Ultrasonic Sound Velocity (28 days)
A (Pumice)c 74.87* 74.86 74.67 74.75
B (Barite)b 74.87 74.88* 74.79 74.80
C (BFS)a 74.78 74.80 74.85* 74.81
D (Colemanite)d 74.73 74.69 74.93* 74.88
Capillary Water Absorption
A (Pumice)c 9.461* 8.425 6.991 7.432
B (Barite)d 8.113 8.529* 8.010 7.663
C (BFS)a 7.847 8.084 8.135 8.241*
D (Colemanite)b 7.636 8.018 9.921* 9.720
Bhme Abrasion Resistance
A (Pumice)d 15.29 15.21* 16.25 15.99
B (Barite)c 15.53 15.34* 15.68 15.71
C (BFS)a 15.44 15.60 15.39* 15.43
D (Colemanite)b 15.75 15.86 14.68* 14.88
*
Optimum level.
a
Most effective factor.
b
2nd effective factor.
c
3rd effective factor.
d
4th effective factor.
 A.H. Sevin et al. / Construction and Building Materials 135 (2017) 343351 349

Table 8
Results of ANOVA for compressive strength.

Factor df SS VA F Contribution (%)

Compressive Strength/7 days
A, Pumice (%) 3 5.9078 1.9693 38.87 30.32
B, Barite (%) 3 9.2775 3.0925 61.04 47.62
C, BFS (%) 3 1.2425 0.4142 8.18 6.38
D, Colemanite (%) 3 2.9036 0.9679 19.10 14.90
Error 3 0.1520 0.0507 0.78
Total 15 19.4834 100.00
Compressive Strength/28 days
A, Pumice (%) 3 0.6479 0.2160 3.49 5.11
B, Barite (%) 3 2.9170 0.9723 15.73 23.01
C, BFS (%) 3 5.9808 1.9936 32.26 47.17
D, Colemanite (%) 3 2.9480 0.9827 15.90 23.25
Error 3 0.1854 0.0618 1.46
Total 15 12.6792 100.00

df: Degree of freedom SS: Adj. Sum of square VA: Adj. Mean Square F: F value.

Table 9
Results of ANOVA for splitting tensile strength.

Factor df SS VA F Contribution (%)

Splitting Tensile Strength/7 days
A, Pumice (%) 3 7.8477 2.6159 43.98 29.63
B, Barite (%) 3 12.1764 4.0588 68.23 45.97
C, BFS (%) 3 1.3906 0.4635 7.79 5.25
D, Colemanite (%) 3 4.8956 1.6319 27.43 18.48
Error 3 0.1785 0.0595 0.67
Total 15 26.4887 100.00
Splitting Tensile Strength/28 days
A, Pumice (%) 3 0.8491 0.2830 1.08 6.51
B, Barite (%) 3 2.9609 0.9870 3.75 22.72
C, BFS (%) 3 5.7484 1.9161 7.28 44.10
D, Colemanite (%) 3 2.6864 0.8955 3.40 20.65
Error 3 0.7898 0.2633 6.02
Total 15 13.0346 100.00

df: Degree of freedom SS: Adj. Sum of square VA: Adj. Mean Square F: F value.

Table 10
Results of ANOVA for ultrasonic sound velocity.

Factor df SS VA F Contribution (%)

Ultrasonic Sound Velocity/7 days
A, Pumice (%) 3 0.51151 0.17050 23.53 33.45
B, Barite (%) 3 0.66652 0.22217 30.67 43.58
C, BFS (%) 3 0.08294 0.02765 3.82 5.42
D, Colemanite (%) 3 0.24665 0.08222 11.35 16.13
Error 3 0.02173 0.00724 1.42
Total 15 1.52934 100.00
Ultrasonic Sound Velocity/28 days
A, Pumice (%) 3 0.054500 0.018167 7.37 16.82
B, Barite (%) 3 0.088553 0.029518 11.98 27.33
C, BFS (%) 3 0.139392 0.046464 18.86 43.02
D, Colemanite (%) 3 0.034191 0.011397 4.63 10.55
Error 3 0.007392 0.002464 2.28
Total 15 0.324028 100.00

df: Degree of freedom SS: Adj. Sum of square VA: Adj. Mean Square F: F value.

characteristics and the signal/noise ratio (S/N) represent the desir-
able signal value and the undesirable noise value, respectively. The
ratio is used to transform the quality characteristics [34].
According to the test results based on Taguchi method, it can be
seen that for compressive strength the optimum conditions were
P2.5B10S10C0.75 for the 7th day and P2.5B2.5S10C0.75 for the 28th day
in which pumice was denoted as P, barite as B, blast furnace
slag as S and colemanite as C. Analysis showed that for the
splitting tensile strength the optimum working conditions were
P5B10S5C0.5 at the 7th day and P10B0S5C0.75 at the 28th day. For
ultrasonic sound velocity tests the optimum conditions were
P2.5B10S5C0 at the 7th day and P0B10S5C0 at the 28th day. For the
capillary water absorption and for the Bhme abrasion test the
optimum working conditions were P10B2.5S0C0.5 and P2.5B2.5S0C0.75
respectively.
The optimum working conditions and the effectiveness level of
each factor were determined by the Taguchi analysis. ANOVA was
performed to determine the contribution percentage of the param-
eters to test results. Tables 812 show the ANOVA results for com-
pressive strength, splitting tensile strength, ultrasonic sound
350 A.H. Sevin et al. / Construction and Building Materials 135 (2017) 343351

Table 11
Results of ANOVA for capillary water absorption.

Factor df SS VA F Contribution (%)

Capillary Water Absorption
A, Pumice (%) 3 8.0941 2.6980 3.35 19.39
B, Barite (%) 3 0.7565 0.2522 0.31 1.81
C, BFS (%) 3 17.7917 5.9306 7.36 42.62
D, Colemanite (%) 3 12.6872 4.2291 5.25 30.39
Error 3 2.4167 0.8056 5.79
Total 15 41.7461 100.00

df: Degree of freedom SS: Adj. Sum of square VA: Adj. Mean Square F: F value.

Table 12
Results of ANOVA for Bhme abrasion resistance.

Factor df SS VA F Contribution (%)

Bhme Abrasion Resistance
A, Pumice (%) 3 0.43859 0.14620 2.71 4.65
B, Barite (%) 3 0.99401 0.33134 6.14 10.54
C, BFS (%) 3 5.10963 1.70321 31.55 54.18
D, Colemanite (%) 3 2.72722 0.90907 16.84 28.92
Error 3 0.16195 0.05398 1.71
Total 15 9.43140 100.00

df: Degree of freedom SS: Adj. Sum of square VA: Adj. Mean Square F: F value.

velocity, capillary water absorption and Bhme abrasion test
respectively.
It is known that pumice, blast furnace slag and colemanite are
used as cement replacement materials and barite is used as filler
material [35]. It could be seen that the compressive strength of
many samples were decreased compared to the reference sample
due to the reduction of cement. Barite was the most effective
parameter on the compressive strength at the 7th day. Barite is
not usually used as cement substitution material but generally it
is used as aggregate for radiation shielding [36]. The literature
did not include any information about barite having an interaction
with cement at the same level as pumice, blast furnace slag and
colemanite. Therefore barite was the most effective parameter on
7-day compressive strength due to its high compressive strength
loss. Second effective parameter on the 7-day compressive
strength was pumice. Pumice was also effective on compressive
strength but it was not as active as blast furnace slag and poz-
zolanic material. The effect of colemanite on the results can be
related to the retarding effect of colemanite on the setting time
[37].
Blast furnace slag was the most effective parameter on the 28-
day compressive strength. It had high contribution to the compres-
sive strength due to the chemical composition similar to cement
and due to the high pozzolanic activity at later ages. Besides blast
furnace slag was the finest material among all. Besides using it as
cement replacement material, using it as fine aggregate showed fil-
ler effect and provided additional contribution to compressive
strength results. It is known that colemanite addition has positive
effect on compressive strength as advanced age resistance. [38].
Barite was negatively affected almost at the same level as coleman-
ite. This result can be related to the inactive structure of barite.
As it can be seen from the splitting tensile test results, similar
results can be seen from the compressive strength results both at
7th day and 28th day. It can be also seen in Table 6, that the com-
pressive strength and splitting tensile strength results were
similar.
The velocity of the ultrasonic wave passing through concrete
and the concretes strength do not have a direct relationship. How-
ever, there is a certain relationship between the density of the con-
crete and the ultrasonic sound velocity. Concrete with high amount
of pores has lower density; therefore the reaching time of ultra-
sonic wave from one surface to other will be longer [39]. According
to the results, it is seen that effects of mineral additives on the
7th and 28th day ultrasonic sound velocity results were similar
to the results of the compressive and splitting tensile strength.
This result can be explained with the indirect relationship between
the mechanical properties of concrete and ultrasonic sound
velocity.
From Table 11, it can be seen that the blast furnace slag and
colemanite had the most effective parameters on capillary water
absorption results. It is known that the blast furnace slag showed
pozzolanic activity at advanced ages and contributed to the devel-
opment of the microstructure [40]. Previous studies showed that
[38] cements incorporating boron had active and stable a or a0
C2S crystal phase and hydration rate was higher compared to the
ordinary Portland cement which contained b C2S phase. This
leads to a continuous hydration rate, in this way more compact
and dense microstructure occurs. This microstructure develop-
ment leads to the disappearing of the capillary pores.
Bhme abrasion results showed that blast furnace slag and cole-
manite were the most effective parameters on abrasion resistance.
This can be related to the contribution of these two mineral addi-
tives to the mechanical properties of the concrete as mentioned
above. Contribution on mechanical properties also affected the
abrasion resistance of the samples similarly to the capillary water
absorption results.
4. Conclusions
Compressive strength, tensile strength, ultrasonic sound veloc-
ity, capillary water absorption and Bhme abrasion resistance tests
were performed on pumice, barite, blast furnace slag and coleman-
ite used as cement replacement and fine aggregates at different
amounts. Taguchi and ANOVA method were carried out in the
study and the followings can be concluded.

Barite was the most effective parameter on early age compres-
sive strength. This result is related to the addition of barite
instead of cement which is an inactive mineral and leads to
compressive strength loss. On the other hand, at later ages blast
furnace slag has shown pozzolanic activity and became the
most effective parameter.
 A.H. Sevin et al. / Construction and Building Materials 135 (2017) 343351

Splitting tensile strength results has shown similarity with the
compressive strength results both at 7th and 28th days. Also
Taguchi results have shown effectiveness of factors are very
similar and ANOVA results have shown that contribution per-
centages are close.

The 7th and 28th day ultrasonic sound velocity result have
shown similarity with the compressive and the splitting tensile
strength results. It can be explained by the indirect relationship
between mechanical properties of the concrete and ultrasonic
sound velocity.

Blast furnace slag and colemanite were the most effective
parameters on capillary water absorption. This result is related
to the contribution of these two additives on microstructure
development. This development causes reduction in capillary
pores.

Bhme abrasion results showed that the most effective param-
eters were blast furnace slag and colemanite. It can be related to
the contribution of these two mineral additives to the mechan-
ical properties.
References
[1] H. Binici, The properties of PC-GBFS-Pumice triple blends PhD Thesis, ukurova
University Civil Engineering Department, Adana, 2002 (in Turkish).
[2] E. zbay, M. Erdemir, H.I. _ Durmus, Utilization and efficiency of ground
granulated blast furnace slag on concrete properties a review, Constr. Build.
Mater. 105 (2016) 423434.
_ Yksel, T. Bilir, . zkan, Durability of concrete incorporating non-ground
[3] I.
blast furnace slag and bottom ash as fine aggregate, Build. Environ. 42 (7)
(2007) 26512659.
[4] H. Binici, M.Y. Durgun, T. Rzaoglu, M. Koluolak, Investigation of durability
properties of concrete pipes incorporating blast furnace slag and ground
basaltic pumice as fine aggregates, Sci. Iranica 19 (3) (2012) 366372.
[5] M. Valcuende, F. Benito, C. Parra, I. Minano, Shrinkage of self compacting
concrete made with blast furnace slag as fine aggregate, Constr. Build. Mater.
76 (2015) 19.
[6] H. Binici, Effect of crushed ceramic and basaltic pumice as fine aggregates on
mortars properties, Constr. Build. Mater. 21 (2007) 11911197.
[7] K. Onoue, H. Tamai, H. Suseno, Shock-absorbing capability of lightweight
concrete utilizing volcanic pumice aggregate, Constr. Build. Mater. 83 (2015)
261274.
[8] L. Gndz, I. _ Ugur, The effects of different fine and coarse pumice aggregate/
cement ratios on the structural concrete properties without using any
admixtures, Cem. Concr. Res. 35 (2005) 18591864.
[9] N. Kabay, F. Akz, Effect of prewetting methods on some fresh and hardened
properties of concrete with pumice aggregate, Cem. Concr. Compos. 34 (2012)
503507.
[10] O. Gencel, Characteristics of fired clay bricks with pumice addition, Energy
Build. 102 (2015) 217224.
[11] N. Kabay, M.M. Tufekci, A.B. Kizilkanat, D. Oktay, Properties of concrete with
pumice powder and fly ash as cement replacement materials, Constr. Build.
Mater. 85 (2015) 18.
[12] W. Deboucha, M.N. Oudjit, A. Bouzid, L. Belagraa, Effect of incorporating blast
furnace slag and natural pozzolana on compressive strength and capillary
water absorption of concrete, Proc. Eng. 108 (2015) 254261.
[13] T. Bilir, I. Yksel, I.B. Topcu, O. Gencel, Effects of bottom ash and granulated
blast furnace slag as fine aggregates on abrasion resistance of concrete, Sci.
Eng. Mater. AOP (2015) 19.
[14] A. Durmus, Y. Grsoy, Comparative study of heavyweight concrete produced
using one of the natural heavyweight aggregates present in the Eastern
Blacksea Region with traditional concrete, in: Proceeding of the Fourth
International Conference on Civil Engineering, Tehran, 1997.
351
[15] A. Durmus, Y. Grsoy, Dogu Karadeniz Blgesi dogal agregalarndan biriyle
retilen agr betonun baslca zelikleri, Turkish Ready Mixed Concrete
Association, J. Ready Mixed Concr. 39 (2000) (in Turkish).
[16] Y. Akgn, A. Durmus, A.Y. Durmus, Barit Agregasyla retilen Agr Bir Betonun
_
zelikleri, in: 7th National Concrete Congress Book, Istanbul, 2007, pp. 465
474 (in Turkish).
[17] I. Akkurt, R. Altindag, C. Basyigit, S. Kilincarslan, The effect of barite rate on the
physical and mechanical properties of concretes under F-T cycle, Mater. Des.
29 (2008) 17931795.
[18] A. Mesbahi, G. Alizadeh, G. Seyed-Oskoee, A.A. Azapeyvand, A new barite-
colematine concrete with lower neutron production in radiation therapy
bunkers, Ann. Nucl. Energy 51 (2013) 107111.
[19] R. Boncukcuoglu, M.T. Yilmaz, M.M. Kocakerim, V. Tosunoglu, Utilization of
trommel sieve waste as an additive in Portland cement production, Cem.
Concr. Res. 32 (2002) 3539.
[20] P.J. Ross, Taguchi Tecniques for Quality Engineering, second ed., McGraw-Hill,
1996.
_ Trkmen, R. Gl, C. elik, A Taguchi approach for investigation of some
[21] I.
physical properties of concrete produced from mineral admixtures, Build.
Environ. 43 (2008) 11271137.
[22] H. Tanyildizi, Statistical analysis for mechanical properties of polypropylene
fiber reinforced lightweight concrete containing silica fume exposed to high
temperature, Mater. Des. 30 (2009) 32523258.
[23] S. Yaln, The inhibitive effect of colemanite added cement on the corrosion of
reinforcing steels M.Sc. Thesis, Gazi University, Institute of Science and
Technology, Ankara, 1996 (in Turkish).
[24] O. Gencel, W. Brostow, C. Ozel, M. Filiz, An investigation on the concrete
properties containing colemanite, Int. J. Phys. Sci. 5 (3) (2010) 216225.
[25] TS 802: Desing of Concrete Mixes, Turkish Standards Institution, Ankara, 2016.
[26] DIN 52108: Testing of Inorganic Non-metallic Materials Wear Test Using the
Grinding Wheel According to Bhme Grinding Wheel Method, Mnchen,
2010.
[27] TS EN 12350-2: Testing Fresh Concrete Part 2: Slump Test, Turkish Standards
Institution, Ankara, 2010.
[28] TS EN 12350-6: Testing Fresh Concrete Part 6: Density, Turkish Standards
Institution, Ankara, 2010.
[29] ASTM C 597: Standard Test Method for Pulse Velocity through Concrete,
American Society for Testing Materials, Pennsylvania, USA, 2016.
[30] TS EN 12390-3: Testing Hardened Concrete Part 3: Compressive Strength of
Test Specimens, Turkish Standards Institution, Ankara, 2010.
[31] TS EN 12390-6: Testing Hardened Concrete Part 6: Testing Splitting Strength
of Test Specimens, Turkish Standards Institution, Ankara, 2010.
[32] M. Benaicha, O. Jalbaud, A.H. Alaoui, Y. Burtschell, Correlation between the
mechanical behaviour and the ultrasonic velocity of fiber-reinforced concrete,
Constr. Build. Mater. 101 (1) (2015) 702709.
[33] TS EN 772-11: Method of Test for Masonry Units Part 11: Determination of
Water Absorption of Aggregate Concrete, Autoclaved Aerated Concrete,
Manufactured Stone and Natural Stone Masonry Units due to Capillary
Action and Initial Rate of Water Absorption of Clay Masonry Units, Turkish
Standards Institution, Ankara, 2011.
[34] A. Kumar, B. Bhunia, D. Dasgupta, T. Mandal, A. Dey, S. Datta, P. Bhattacharya,
Optimization of culture condition for growth and phenol degradation by
Alcaligenes faecalis JF339228 using Taguchi Methodology, Desalin. Water
Treat. 51 (2013) 31533163.
[35] A.H. Sevinc, Engineering properties of barite, basaltic pumice, colemanite and
blast furnace slag added mortars and concretes MSc thesis, Kahramanmaras
Sutcu Imam University Civil Engineering Department, Kahramanmaras, 2011
(in Turkish).
[36] I. Akkurt, H. Akyildirim, B. Mavi, S. Kilicarslan, C. Basyigit, Gamma-ray
shielding properties of concrete including barite at different energies, Prog.
Nucl. Energy 52 (7) (2010) 620623.
[37] E. Erdogmus, Y. Erdogan, O. Gencel, S. Targan, U. Avciata, Influence of
colemanite admixture on Portland cement durability, Adv. Cem. Res. 24 (3)
(2012) 155164.
[38] Turkish General Directorate of State Hydraulic Works Technical Bulletin Issue
105, Ankara, 2009, pp. 123 (in Turkish).
[39] T. Erdogan, Concrete, fourth ed., METU Press, 2013 (in Turkish).
[40] T. Kefeng, P. Xincheng, Strengthening effect of finely ground fly ash, granulated
blast furnace granulated blast furnace slag (GBFS), and their combination,
Cem. Concr. Res. 28 (1998) 18191825.