Available online at www.sciencedirect.com

Procedía Engineering

ELSEVIER

Procedía Engineering 54 (2013) 827 - 839

www.elsevier.com/locate/procedia

The 2nd International Conference on Rehabilitation and Maintenance in Civil Engineering

Mixed Concrete Optimization Using Fly Ash, Silica Fume and Iron Slag on the SCC's Compressive Strength Raharjo. Da*, Subakti. Ab, Taviob

aCivil Engineering (structure), Faculty of Civil Engineering and Planning-ITS, Department of Civil Engineering(structure), Faculty of Civil Engineering and Planning-ITS, Surabaya Indonesia

Abstract

Self-Compacting Concrete (SCC) is an innovative concrete that does not require vibration process to its placing. SCC is able to flow under its own weight, enables it to meet or filling formwork and reached its highest density. SCC requires a mineral Admixture such as fly ash, superplaticiser and other compounds such as iron slag waste from steel mill wastes in the form of fine aggregate in order to meet the specified flowability. Some trial mixtures containing fly ash, silica fume, policarboxilate based of superplasticer , and iron slag have been performed that aims to determine the SCC's optimal composition and meet the requirements of filling ability, passing ability , viscosity and segregation. The concrete's filling ability, passing ability, viscosity and segregation were conducted using slump cone, L-box and V-funnel . The cylindrical sample of 10 cm in diameter and 20 cm in heigh of hardenened SCC was also tested at 3.7, 14, 28 and 56 days of concrete age.There were 33 variation of concrete mixture using 495 samples total mixture have been tested. Each composition contained various superplasticizer dosage from 0.5 to 1.8% of cementitious weight. The dosage of silica fume was also varied 0%, 10% and 20% of fly ash weight. The goal that expected from this study is to obtain the optimal material composition of the mixture that produce the maximum compressive strength but cheaper and comptetiteve in price.

© 2013 TheAuthors. Publis hed by Elsevier Ltd.

Selection and peer-review under responsibility of Department of Civil Engineering, Sebelas Maret University

Keywords: Self-compacting concrete; fly ash; iron slag; silica fume; viscocrete; workability; flowability; and compressive strength.

* Corresponding author.

E-mail address: putro_yeah@yahoo.com, subaktiaman@yahoo.com, tavio@its.ac.id 1877-7058 © 2013 The Authors. Published by Elsevier Ltd.

Selection and peer-review under responsibility of Department of Civil Engineering, Sebelas Maret University doi: 10.1016/j.proeng.2013.03.076

1. Introduction

Reinforced concrete is construction material that widely used since it is low cost, its material components such as sand, gravel, and cement are easy to find and this material is easy to be formed depended on the consumer demand

In concrete construction, concrete compaction or vibration is a job that absolutely must be done for resulting better conventional reinforced concrete structures. The purpose of compaction itself is to minimize air trapped in fresh concrete in order to obtain a homogeneous concrete and consequently the cavities in the concrete does not occur. However, improper compaction will affect the concrete quality such as lowering the compressive strength, higher concrete pore that will easily produce rust or corrosion in its steel reinforcement (Handoko, Gideon, 2001).

In Indonesia Research on Self-compacting Concrete (SCC) is rarely done. This motivates authors to carry out this work. There are many materials available for producing SCC in Indonesia, such as fly ash, silica fume, slag and iron. However the best choice which material shoul be taken and how their proportion in concrete mix desing is not clearly known. Therefore the objective of this research is to determine the optimum mix design containg fly ash, silica fume, slag and iron for producing a high compressive strength of SCC with a competitive price.

This works have tested 33 SCC mixtures composition using 495 samples. The statistical computer program of SPSS is applied for data analysis.

2. Work Method

2.1 Work's Flowchart

Figure 1 presents the flowchart for Concrete Mix Design of SCC. The work is a series from the literature study, preparation materials/materials, analysis of materials, mix design per composition of each ingredient, compressive strength testing to statistical analysis with the program.

FINISH

Figure 1. Work flowchart to produce optimum SCC mix design

2.2 Calculation of SCC's material proportion

This study uses a maximum coarse aggregate size of 20 mm with a slump flow 60180 mm Table 1 shows the free water needed for a designed concrete workability and Figure 2 will assist to determine the specific gravity of concrete that is used to estimate the fresh concrete density.

Table 1. Approximation of free water content required _for produce different levels of workability_

Slump (mm) 0-10 10-30 30-60 60-180

Vebe Time (Sec) >12 6-12 3-6 0-3

Size Agrégat Maks Type Agregat W/C (kg/m3)

Bagian A

Concrete with Cement Portland

10 Uncrushed 150 180 205 225

Crushed 180 205 230 250

20 Uncrushed 135 160 180 195

Crushed 170 190 210 225

40 Uncrushed 115 140 160 175

Crushed 155 175 190 205

Bagian B Concrete with PC Fly Ash

Proporsional Fly Ash (cement+fly ash) in % Reduction W/C (kg/m3)

10 5 5 5 10

20 10 10 10 15

30 15 15 20 20

40 20 20 25 25

50 25 25 30 30

(source: Aman Subakti, "Mixed Design for Normal Concrete with DOE & ACI Methode, 1995)

The used portlan cement (PC) is calculated following the equation (1) :

PC = (100-P)x (1)

Content (l00-0.7p)x[W/(C+0.3F)] K }

Where: [W/(C+0.3F)] use 0,3,0,5 and 0.7. while the fly ash is determined as the equation (2):

Fly ashContent= -X (2)

where p = fly ash proportion, w = free water content, and c = cement content

The total aggregate in concrete mixture is determined following Subakti. A. (1995) with following streps: the density of fresh concrete is estimated using Figure 2. Total amount of aggregate reqired then is obtained by reducing the weight of cement, fly ash and water from the density of fresh concrete of Figure 2. This involves multiplying the weight of fly ash added to the calculation of the total compared with the amount of aggregates in saturated surface dry (SSD) state. This can be formulated as on equation

Amount of aggregate (SSD) = D - (C + F) - W (3)

where D is the density of fresh state concrete (kg/m3) of Figure 2.

2400 -

Relatuve Density of combined aggregate for saturated and

for crushed Aggregates

Assu^^d for uncrushed

Aggregates

water content (kg/m )

Figure 2. Estimated spesific gravity of fresh concrete

2.3 Analysis

The following is the scenario analysis that is applied to determine the effect of the response variable (y) = compressive strength of the predictor variable/explanatory/independent for more than 1 variable using multiple linear regression. In this case the compressive strength (y = response variable) is influenced by four variables: water, cement, fly ash and silica fume on the strength of concrete. Assuming the amount of superplasticizer had no effect on the compressive strength of concrete, then regression modeling is as follows:

Yb = ß0 + PÄ + ß2X2 +..........+ ß5X5 + 8

where : Xn = variable (fly ash, silica fume, water, cement and superplasticizer)

p0 = total average of compressive strength/concrete age

P„ = parameter from predictor variable

e = variable error

2.4 Material proportioning of SCC's

The following is the a series steps for materian proportioning of SCC

• Coarse aggregates are limited in number approximately about 45% of the total volume of concrete that can flow and solidified itself without compactor.

• While the comparison with the coarse aggregate fine aggregate is 45: 55.

• Fly ash composition of 0%, 10%, 30% and 50%

• Superplasticizer dose in range 0.5 to 1.8% of the total weight of cement.

• The composition of used silica fume is 0%, 10% and 20% by weight of fly ash pozzolan.

• In this case the fine aggregate sand and iron slag, the comparison only up to 15% replacement of sand, due to the weight of a large iron slag. The use of iron slag is as a filler.

• water content at the beginning of a trial mix using a W/C 0.3 and gradually mix in trial W/C up to 0.5, W/C 0.7 by adding superplaticizer by considering the conditions the SCC, after do visual checking and then added water or reduction of water at a small increment into SCC.

Table 2 lists the mix proportion of SCC

Table 2. Material proportion of each SCC

No. % Fly Ash % Additive W/C Silica Fume Fine Agregat Concrete Test

Compressive strength cylinder 10/ 20

3 day 7 day 14 day 28 day 56 day

1 0% 0,5-1,8% 0,3, 0,5 and 0,7 0%, 10% and 20% N S 3 3 3 3 3

2 10% 0,5-1,8% 0,3, 0,5 and 0,7 0%, 10% and 20% N S 3 3 3 3 3

3 30% 0,5-1,8% 0,3, 0,5 and 0,7 0%, 10% and 20% N S 3 3 3 3 3

4 50% 0,5-1,8% 0,3, 0,5 and 0,7 0%, 10% and 20% N S 3 3 3 3 3

Fine aggregate in the table above can be explained as follows:

> Type N, normal: namely 0% iron slag ratio: 100% Lumajang sand.

> Type S, with slag: the iron slag ratio 15%: 85% Lumajang sand. Where iron slag in SCC is required only as a filler, because its density is large, then the percentage of iron slag was taken to a maximum of 15% of the amount of fine aggregate.

> The applied ratio of fine aggregate to coarse aggregate is 55: 45.

> The applied water cement ratio is 0.3, 0.5 and 0.7 by adding superplasticizer, then do a trial mix in order to get a composition that meets the requirements of SCC.

> The superplasticizer dosage is 0.5 to 1.8% in the table is taken from a producer brochure while the composition of used silica fume is 0%, 10% and 20% by weight of fly ash pozzolan.

Based on Table 2 above then obtained 33 composition, with a total of 495 samples can be seen in Table 3.

In this research is using of superplasticizer ranging from 0.5% s / d 1.8% for all compositions by trial mix prior to qualify SCC. So that for each composition requires 15 test objects (to compressive test at the age of 3,7,14, 28 and 56 days in which each age concrete test object used 3 pieces). Accordingly, the total samples required was 495.

Table 3. Composition and Code Samples

No Code Sample W/C Agregat FA (%) SF (%) Z ages of sample (day)

Type BP L S 3 7 14 28 56

1 P.N.3-FA0SF0 0,3 N V V - 0 0 3 3 3 3 3

2 P.N.5-FA0SF0 0,5 N V V - 0 0 3 3 3 3 3

3 P.N.7-FA0SF0 0,7 N V V - 0 0 3 3 3 3 3

4 P.S.3-FA0SF0 0,3 S V V V 0 0 3 3 3 3 3

5 P.S.5-FA0SF0 0,5 S V V V 0 0 3 3 3 3 3

6 P.S.7-FA0SF0 0,7 S V V V 0 0 3 3 3 3 3

7 P.S.3-FA10SF0 0,3 S V V V 10 0 3 3 3 3 3

8 P.S.5-FA10SF0 0,5 S V V V 10 0 3 3 3 3 3

9 P.S.7-FA10SF0 0,7 S V V V 10 0 3 3 3 3 3

10 P.S.3-FA10SF10 0,3 S V V V 10 10 3 3 3 3 3

11 P.S.5-FA10SF10 0,5 S V V V 10 10 3 3 3 3 3

12 P.S.7-FA10SF10 0,7 S V V V 10 10 3 3 3 3 3

13 P.S.3-FA10SF20 0,3 S V V V 10 20 3 3 3 3 3

14 P.S.5-FA10SF20 0,5 S V V V 10 20 3 3 3 3 3

15 P.S.7-FA10SF20 0,7 S V V V 10 20 3 3 3 3 3

16 P.S.3-FA30SF0 0,3 S V V V 30 0 3 3 3 3 3

17 P.S.5-FA30SF0 0,5 S V V V 30 0 3 3 3 3 3

18 P.S.7-FA30SF0 0,7 S V V V 30 0 3 3 3 3 3

19 P.S.3-FA30SF10 0,3 S V V V 30 10 3 3 3 3 3

20 P.S.5-FA30SF10 0,5 S V V V 30 10 3 3 3 3 3

21 P.S.7-FA30SF10 0,7 S V V V 30 10 3 3 3 3 3

22 P.S.3-FA30SF20 0,3 S V V V 30 20 3 3 3 3 3

23 P.S.5-FA30SF20 0,5 S V V V 30 20 3 3 3 3 3

24 P.S.7-FA30SF20 0,7 S V V V 30 20 3 3 3 3 3

25 P.S.3-FA50SF0 0,3 S V V V 50 0 3 3 3 3 3

26 P.S.5-FA50SF0 0,5 S V V V 50 0 3 3 3 3 3

27 P.S.7-FA50SF0 0,7 S V V V 50 0 3 3 3 3 3

28 P.S.3-FA50SF10 0,3 S V V V 50 10 3 3 3 3 3

29 P.S.5-FA50SF10 0,5 S V V V 50 10 3 3 3 3 3

30 P.S.7-FA50SF10 0,7 S V V V 50 10 3 3 3 3 3

31 P.S.3-FA50SF20 0,3 S V V V 50 20 3 3 3 3 3

32 P.S.5-FA50SF20 0,5 S V V V 50 20 3 3 3 3 3

33 P.S.7-FA50SF20 0,7 S V V V 50 20 3 3 3 3 3

I sample 495

Specification code PN3-FAOSFO composition (proportion of Normal, W / C 0.3, Fly Ash and Silica Fume 0% 0%), where BP = crusher stone, S = slag, L = Lumajang sand, FA = fly ash and SF = silica fume.

2.5 Fresh Scc Concrete Testing

In the fresh state condition, concrete test specimen was analyzed by performing several tests to assess its self compacted behavior , the slump test was conducted in order to find out workabilitas and flowability of the concrete mix. Other test were U-flow test or L-box test which intended to determine the passing ability of self-compacting concrete and VFunnel test for the viscosity of concrete mixture. Figure 3 shows the SCC slump flow test.

Figure 3. Test SCC With Slump test, V funnel and L-Box

3. Results

3.1 Material properties

Table 4 and Table 5 present the properties of iron slag while Table 6 and Table 7 present the properties of used aggregates.

Table 4. Analysis for iron slag materials

Material Test Test Results Unit Standard

Iron Slag Specific weight 3,448 gr/cm3 Standard ASTM C 128 - 78

Volume Weight 1,809 gr/cm3 Standard ASTM C 29/ C29M-91a

Clean iron slag to mud (washing materials) 0,75 % Standard ASTM C 117-95

Table 5. Chemical Analysis of Iron Slag

Parameter Unit Test Results Methode

Silicon dioxide (SiO2) % 35.67 Gravimetri

Aluminium Oxide (Al2O3) % 3.62 Spektrophotometri

Iron Oxide (Fe2O3) % 41.39 AAS

Calsium Oxide (CaO) % 6.12 Titrimetri

Magnesium Oxide (MgO) % 4.27 Titrimetri

Sodium dioxide (Na2O) % 0.81 Flamephotometri

Kalium dioxide (K2O) % 0.43 Flamephotometri

Sulfur trioxide (SO3) % 1.83 Spektrophotometri

Moisture Content (H2O) % 0.82 Gravimetri

Loss on Ignition (LOI) % 4.45 Gravimetri

Result Test in Lab. TAKI Tek. Kimia ITS No. 985/LTAKI/VII/2011 _Table 6. Analysis for Coarse Materials

Test Analysis results Unit Standard

Coarse humidity 0,455 % Standard ASTM C 566 - 89

Specific weight 2,68 gr/cm3 Standard ASTM C 127-88-93

Volume Weight 1,60 gr/cm3 Standard ASTM C 29/ C29M-91a

Table 7. Analysis for Fine Aggregate Materials (Lumajang)

Test Analysis results Unit Standard

Sand humidity 0,62 % Standard ASTM C 566 - 89

Specific weight 2,71 kg/m3 Standard ASTM C 128 - 93

Volume Weight 1,614 gr/cm3 Standard ASTM C 29/ C29M-91a

3.2 SCC's compressive strength

The compressive strength of hardened concrete then was tested on 3, 7, 14, 28 and 56 days of concrete age. The results are listed on Table 8.

Table 8. Test Result for Compressive strength

No Code Sample Compressive Strength (Mpa)

3 day 7 day 14 day 28 day 56 day

1 P.N.3-FA0SF0 35,424 44,418 57,655 61,303 66,182

2 P.N.5-FA0SF0 21,106 22,994 26,770 33,621 34,788

3 P.N.7-FA0SF0 8,358 10,632 13,936 17,012 18,412

4 P.S.3-FA0SF0 36,994 40,133 50,442 54,303 64,739

5 P.S.5-FA0SF0 28,382 35,085 38,139 41,258 48,236

6 P.S.7-FA0SF0 12,621 17,182 25,455 29,888 35,085

7 P.S.3-FA10SF0 44,800 50,973 53,603 58,524 73,182

8 P.S.5-FA10SF0 21,552 28,106 33,558 41,045 45,861

9 P.S.7-FA10SF0 12,961 19,261 23,036 33,982 38,012

10 P.S.3-FA10SF10 26,282 35,679 39,773 47,855 61,664

11 P.S.5-FA10SF10 19,452 23,461 30,164 37,333 46,200

12 P.S.7-FA10SF10 13,321 17,394 20,448 23,609 29,442

13 P.S.3-FA10SF20 30,970 39,773 48,109 51,779 59,818

14 P.S.5-FA10SF20 23,758 28,255 38,139 45,182 54,091

15 P.S.7-FA10SF20 10,182 13,152 17,309 23,715 29,400

16 P.S.3-FA30SF0 27,279 33,515 43,898 50,294 60,455

17 P.S.5-FA30SF0 16,588 25,667 32,582 36,909 46,285

18 P.S.7-FA30SF0 9,715 14,403 22,039 27,236 38,500

19 P.S.3-FA30SF10 28,573 32,964 49,827 57,336 76,682

20 P.S.5-FA30SF10 21,912 25,348 33,600 44,079 54,855

21 P.S.7-FA30SF10 13,258 19,748 21,594 28,000 36,909

22 P.S.3-FA30SF20 31,415 43,633 53,412 65,333 77,191

23 P.S.5-FA30SF20 19,112 23,058 29,803 40,176 59,564

24 P.S.7-FA30SF20 9,970 15,145 23,588 32,115 33,727

25 P.S.3-FA50SF0 32,730 39,794 50,145 59,182 61,982

26 P.S.5-FA50SF0 15,421 22,442 29,803 39,667 47,536

27 P.S.7-FA50SF0 10,458 13,152 16,885 34,703 42,085

28 P.S.3-FA50SF10 31,648 37,545 44,142 49,700 61,091

29 P.S.5-FA50SF10 18,539 22,782 27,830 38,479 45,861

30 P.S.7-FA50SF10 10,436 13,194 18,370 30,800 40,388

31 P.S.3-FA50SF20 25,624 38,945 44,630 52,627 70,339

32 P.S.5-FA50SF20 18,667 27,045 32,900 37,164 45,818

33 P.S.7-FA50SF20 11,264 14,467 19,939 23,779 27,427

3.3 Data analysis

Predictor variables used are cement, water, sand, silica fume, fly ash, and viscocrete ironslag. All seven predictor variables were used to predict the compressive strength of concrete for the measurement started 3 days, 7 days, 14 days, 28 days and 56 days. The first step taken is to analyze the correlation between variables in the study; response and predictor variables.

Based on the results of correlation analysis from SPSS 19, These results indicate that the existence of the alleged multicolinearitas. The existence of this multico cases cause regression models generated by multiple regression analysis is not appropriate. Then performed regression analysis / principal components (PCR) to be the solution of the case multico. The first step is get eigen values/component which is a representation of variance showed that the amount of variation can be explained by the formation of the components. The cumulative number of eigen values / components produced must equal the total number of variables included in the model (seven variables). Of the seven variables taken eigenvalues have a value greater than 1, which indicates the form of several predictor variables that are correlated origin. So that these variables will be represented by a new variable based on these eigenvalues. This representation is realized by a new equation that each independent predictor variable. factor which is the equation that contains the eigen vectors as linear combinations of variables are not correlated principal components / mutually independent

Tabel 9. Principal Components Analysis

Total Variance Explained

Initial Eigen values Rotation Sums of Squared Loadings

Component Total % of Variance Cumulative % Total % of Variance Cumulative %

1 3.909 48.860 48.860 3.538 44.226 44.226

2 2.369 29.619 78.479 2.740 34.253 78.479

3 .766 9.570 88.048

4 .536 6.700 94.749

5 .338 4.228 98.976

6 .071 .893 99.870

7 .010 .130 100.000

8 6,14E-07 7,68E-06 100.000

Results of analysis of main components based on Table 10 shows the predictor variables is reduced from seven dimensions to 2 dimensions. Based on the number of main components of a larger value, namely the first component is 3.909 and the second component is 2,369. These two components are able to represent 78.479% of the variation of data from seven predictor variable origin. The first component represents 44.26% of the variation is a combination of predictor variables from water, sand, silica fume, fly ash and viscocrete. While the second component of the combined variable of cement, crushed stone, and ironslag

Table 10. Matrix Component 1 and 2

Rotated Component Matrix" Component Score Coefficient Matrix

Component _Component

_1_2 _1_2_

CEMENT -.006 -.945 CEMENT -.069 -.361

WATER -.876 -.387 WATER -.210 -.287

SAND -.687 .657 SAND -.157 .202

COARSE -.502 .853 COARSE -.088 .290

SILICAFUME .720 .006 SILICAFUME .213 .053

FLYASH .930 -.031 FLYASH .273 .054

IRONSLAG .253 .650 IRONSLAG .121 .266

SUPERPLAS .774 -.339 SUPERPLAS .205 -.075

From Table 10 we will get a new equation between existing independent predictor variables based on the value of the coefficient matrix components.

Component 1 = -0.069 cement - 0.210 Water - 0.157 sand - 0.088 stone crush + 0. 213 silicafume + 0.273 fly ash + 0.121 ironslag + 0.205 superlasticizer

Component 2 = -0361 cement - 0.287 water + 0.202 sand + 0.290 stone crush + 0. 053 Silicafume + 0.054 fly ash + 0.054 ironslag - 0.075 superplasticizer

Both of the above equations are used to calculate the value of the two new variables formed by inserting the value of the origin of the seven predictor variables of origin. The results of the two variables are that the new predictor variables for regression analysis with fixed response variable compressive strength of concrete. This is the second step in the regression of the primary component. The results of regression analysis of the major components can be seen in Table 11.

Table 11. Principal Components Regression

Compressive Strength 3 Day (SIG. 5%) 7 Day (SIG. 5%) 14 Day (SIG. 5%) 28 Day (SIG. 5%) 56 Day (SIG. 5%)

CONCURRENT MODEL F COUNT 47.279 48.339 56.776 45.083 37.29

SIGINIFICANT sig. sig. sig. sig. sig.

MODEL PARTIAL

(Constant) component 1 component 2 21.175 (Sig.) 1.818 (Sig.) -8.038 (Sig.) 26.95 (Sig.) 2.954 (Sig.) -9.071 (Sig.) 33.683 (Sig.) 3.724 (Sig.) -10.5 (Sig.) 40.848 (Sig.) 4.926 (Sig.) -9.53 (Sig.) 49.449 (Sig.) 7.207 (Sig.) -10.604 (Sig.)

DIAGNOSIS MODEL R2 75.9% 76.3% 79.1% 75% 71.3%

Adj R2 74.3% 74.7% 77.7% 73.4% 69.4%

Normal error correct correct correct correct correct

Identik error correct correct correct correct correct

Indep. Error correct correct correct correct correct

Regression equation between the main components of the response variable with the component, which is entirely in units of kilograms are as follows :

Compressive Strength28 days = 40.848 + 4.926 component l - 9.53 component2. (7)

Then the regression model with predictor variables included the origin of the component, we substitute into:

SCC's compressive strength 28 days = 40.848 + 3.100395 cement - 0.587366 water - 2.69854 sand -3.19724 stone crush + 0.544276 silica fume + 0.830342 fly ash = 1.93886 ironsla + 1.724703. superplaticizer (8)

4. Conclusions And Recommendations

The conclusions of this work are

• The formula of SCC compressive strength at 28 of concrete age can be drawn

SCC compressive Strength 28 days = 40.848 + 3.100395 cement - 0.587366 water - 2.69854 sand -3.19724 stone crush + 0.544276 silicafume + -1.93886 fly ash + 0.830342 iron slag + 1.724703 viscocrete (kg)

• This obtained formula is valid for the used data in this research, since the accuracy of some of data is questionable due to the enviroment temperature, air moisture, scale of measuring materials, etc, therefore this formula needs to be proved using other composition for further research.

• Since there is a trend of a significat increase of SCC's compressive strength at 56 days of concrete age, It is advised to do the SCC's compressive strength test at 91-days of concrete age.

References

Douglas M. & a. Peck J.W and son, lnc (1992)," Introduction to linear regression analysis". 2nd edition

Subakti, A.(1998), "Mixed Normal Concrete Design with DOE & ACI Methode", ISBN 979-15187-4-2, ITS Press.

The European Guidelines for Self Compacting Conncrete (2005), http://www.efnarc.org/pdf/SCCGuidehnesMav2005.pdf