- Email: [email protected]
Development of expandability charts for Ankara Kalecik Clay
Applied Clay Science 101 (2014) 533–540
Contents lists available at ScienceDirect
Applied Clay Science journal homepage: www.elsevier.com/locate/clay
Research paper
Development of expandability charts for Ankara Kalecik Clay Ahmet Ozguven ⁎ General Directorate of Mineral Research and Exploration, 06800 Ankara, Turkey
a r t i c l e
i n f o
Article history: Received 7 February 2012 Received in revised form 7 July 2014 Accepted 28 August 2014 Available online 8 October 2014 Keywords: Lightweight aggregate Expanded clay aggregates Working parameters Expandability Optimization
a b s t r a c t The main aim of this study is to establish optimum working conditions and to propose expandability charts for Kalecik Clay using expansion ratio versus unit volume weight values via a statistical computer program (Design Expert 7.0). First of all, some suitable samples were collected from the Kalecik County (Ankara, Turkey). Later, the samples were crushed and milled up to 200 μm clay size in the laboratory. Then, pellets with a 10 mm diameter were dried in ovens and heated inside a furnace to expand in different temperatures. One heating operation began with 900 °C and ended with 1200 °C. In addition to that, the holding time periods were selected as 5, 10, 15 or 20 min for each operation. In the final stage, the unit volume weight and expansion ratio of the produced aggregates were determined. As results of this study, both the optimum furnace holding time and furnace temperature values for the samples were determined in order to maximize the expansion ratio values and minimize the unit volume weight for Kalecik Clays. Some expandability charts were proposed for Kalecik Clay by using the findings. The charts allow one to assign expansion ratio and unit volume weight values in different working conditions including holding time inside the furnace and furnace temperature. The proposed charts are expected to be quite useful for more efficient production of expanded clay aggregates in construction industry. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Lightweight aggregates are defined as natural or artificial materials, which are granular, porous and lightweight (Cougny, 1990). They can originate from different natural sources such as volcanic rocks (pumice, volcanic tuffs), sedimentary rocks (clayey diatomite) and metamorphic rocks (claystones, slates, shales), and from waste material or industrial by-products such as recycled flat glass and fly ash (Fakhfakh et. al., 2007). The lightweight aggregates are used as a loose material — for example in back wall fillers and in agronomic applications — or, with a binder, in the manufacture of plaster, asphalt and lightweight concrete thermo-acoustic insulators, as well as in lightweight structural concrete production. Lightweight expanded aggregates can be formed by a quick heating at high temperature of some rocks which are able to bloat. The raw material must contain substances that develop gas upon heating and, at the same time, the material must transform into a highly viscous plastic mass able to expand by virtue of gas entrapment (de Gennaro et. al, 2007). There are a lot of studies about the production of expanded clay aggregates (Decleer and Viaene, 1993; Bragdon, 1996; Fragoulis et. al., 2004; Pioro and Pioro, 2004; Rattanachan and Lorprayoon, 2005; Fakhfakh et. al., 2007; de Gennaro et. al, 2007; Ozguven, 2009; Bartolini et al, 2010). ⁎ Tel.: +90 312 2011227; fax: +90 312 2878747. E-mail address: [email protected].
http://dx.doi.org/10.1016/j.clay.2014.08.028 0169-1317/© 2014 Elsevier B.V. All rights reserved.
As can be seen from the literature, there are some important studies on the parameters, taken one by one, affecting the expansion of the clays for producing lightweight expanded clay aggregates. However, so far no study has been carried out to investigate the combined influence of working parameters (holding time in the furnace and furnace temperature) on expansion efficiency parameters, jointly expansion ratio and unit volume weight, to determine the optimum working conditions and also to develop the expandability charts with expansion ratio and unit volume weight. For this reason, the purpose of this study is to determine the optimum working conditions and also to develop expandability charts for Kalecik Clay with respect to expansion ratio and unit volume weight by using a special statistical program (Design Expert 7.0) providing the factorial design of the experiment which is very important in the optimization of the experimental results. 2. Material and method The methodology of this study is composed of four steps. 1 Sampling and preparation of the samples 2 Sampling of clay in a constant pellet size of 10 mm and clay grain size of 200 μm. 3 Expansion tests 4 Statistical assessment of the test results Considering these four steps, the expansion operation methodology and the work followed in this study are given in detail in Fig. 1.
534
A. Ozguven / Applied Clay Science 101 (2014) 533–540
Fig. 1. Methodology followed in this study.
The studies were conducted in clays taken from the Ankara City's Kalecik County. From a geological point of view, of the expanding clay fields under study, the Ankara Kalecik field has a grayish metallic luster, while it is seen as dark gray-black when wet. Secondary calcite was developed in some places. It has sandstone-shale alteration with a schist-like view behind limestone lenses. This unit is overlain by grayish-gray color sandstone-shale faces. It is a unit exposed in macro-scale achieving 100 m from place to place along approximately
a 1-km path in the form of 25–30 m lenses along the north–south direction. X-ray diffraction pattern belonging to the samples obtained from the field under study is given in Fig. 2 and the result from the X-ray diffraction analysis is seen in Table 1. Chemical analysis (Table 2) showed that the main constituents of the raw materials are silica, alumina, and iron oxides. Na2O and K2O contents are mainly attributed to the clay minerals and feldspars.
A. Ozguven / Applied Clay Science 101 (2014) 533–540
535
Fig. 2. X-ray diffraction pattern belonging to the sample.
The samples collected from the fields were broken and then milled to −200 μm clay size. Ground clay was mixed with only water without any additive to produce clay dough. Dough preparations were left for maturation for one day and shaped through an extruder. Pellets were prepared in a 10 mm diameter and the diameter/length ratio of the pellets is approximately 1:1. The prepared pellets were dried in ovens and then expanded in a furnace. Firing processes took place at different temperatures beginning from 900 °C up to 1200 °C. Raw pellets were kept inside the furnace for different time periods at the same temperature (5, 10, 15 or 20 min). Fired pellets were removed from the furnace and cooled suddenly. Unit volume weight of the produced aggregates was measured by using the ASTM C493-98 standard (ASTM, 1998). The found unit volume weight of the expanded clay aggregates (UVWexp) was compared with the unit volume weight of the raw pellets (UVWorj) to calculate the expansion ratio. The expansion ratio is calculated as (UVWorj / UVWexp) × 100.
the determination process of the optimum points by means of the derived equations. Imaging the optimum working points and the estimated results obtained as a result of the experiments made on these points is possible with this program (Ozcelik and Kanbir, 2011; Ozcelik et. al. 2012). 3.2. Design summary Before initializing the statistical analysis, information that reflects the properties of each variable regarding the factors and response was analyzed. This information is composed of the data such as mean and standard deviation for describing the frequencies about the variables and also definitions of the methods and models to be used in modeling studies. The design properties used in this study are demonstrated in Table 3 and the descriptive statistical data regarding factors and response are shown in Tables 4 and 5, respectively. In addition to the descriptive statistics, the relationships between unit volume weight and expansion ratio and temperature at various
3. Statistical analysis Within the scope of this study, Design Expert 7.0, which is a special statistical program, was used for data analysis and for determining the optimum values of furnace temperature and holding time in the furnace for the Kalecik Clay sample. The method of the statistical analysis performed in this study is given in Fig. 1. 3.1. Design Expert 7.0 statistical software Design Expert 7.0 is a widely used program which was developed for the experimental optimization process and which can effectively design the experiments in the most suitable way according to different methods. After making the experiments based on the design selected and entering the results obtained in the program, it derives the most suitable equations for dependent variables (response) and can realize
Table 1 XRD analysis result of the clay under study. Fields under study
Minerals
Ankara
Quartz (ASTM No: 5-0490), illite, chlorite, mixed layered clay minerals, feldspar, calcite (ASTM No: 5-0586), (graphite)
Table 2 Chemical composition of the studied samples. Na2O MgO Al2O3 SiO2 P2O5 K2O CaO TiO2 MnO Fe2O3 Fluxing (Na2O + MgO + K2O + CaO + Fe2O3) SiO2/Al2O3 SiO2/fluxing
0.98 2.54 18.16 56.11 0.16 3.73 3.04 0.70 0.19 6.89 17.18 3.09 3.27
Table 3 Statistical design properties used in the study. Study type Initial design Design model Runs Blocks
Factorial D-optimal, Point Exchange 2FI 28 No blocks
536
A. Ozguven / Applied Clay Science 101 (2014) 533–540
Table 4 Descriptive statistical data of the factors. Sample name
Factor name
Units
Type
Low actual
High actual
Mean
Std. dev.
Kalecik Clay
Temperature (Temp) Time (Time)
(°C) (min)
Numeric Numeric
900.00 5.00
1200.00 20.00
1050.000 12.500
100.000 5.590
Table 5 Descriptive statistical data of the responds. Sample name
Response name
Units
Obs
Analysis
Min.
Max.
Mean
Std. dev.
Model
Kalecik Clay
Unit volume weight (UVW) Expansion ratio (ER)
(g/cm3) (%)
28 28
Polynomial Polynomial
0.36 1.00
1.89 5.50
1.20 2.41
0.60 1.61
Cubic Quadratic
Fig. 3. Relationships between unit volume weight and expansion ratio and time at various temperatures for Kalecik Clay.
holding times in the furnace are shown in Fig. 3. As can be seen from Fig. 3, as the furnace temperature increases, the expansion ratio increases while the unit volume weight decreases. The complexity of most scientific mechanisms is such that in order to be able to predict an important response, a well-known multiple regression model is needed. The model formed for regression analysis is a model which involves dependent and independent variables. In such a model, the change in dependent variable is tried to be explained with independent variables. The first phase in forming simple or multiple regression models is to determine the coefficients that form the model and then the statistical test of the validity of the model with variance analysis. The figures demonstrating the results obtained from here are called ANOVA (Analysis of Variance). F value obtained from the figures is the value to be tested for the validity of the general model. F value and the results of ANOVA in regression analysis and in other models show the validity of the model formed, in other words, the representativeness of the system (Ozcelik and Kanbir, 2011; Ozcelik et. al. 2012).
weight model equation was obtained. The cubic model for the Kalecik Clay sample, which was found to be statistically the most significant in the analysis, was chosen as the most suitable model for estimating the unit volume weight (Table 6). The validity of the cubic model was tested with variance analysis. The results were presented in Tables 7 and 8, respectively. The model based on the regression coefficients given in Table 7 is statistically significant at 99% (α = 0.01) confidence level (*P = 0.0001 b α = 0.01). The estimation graph of the developed model is given in Fig. 4. The equation of the cubic model for Kalecik Clay that was obtained according to Table 7 is as follows:
3.3. Statistical assessment related to unit volume weight
Table 7 The results of the multiple regression analysis for unit volume weight.
Statistical analysis was made for the purpose of estimating unit volume weight by using time and temperature and the unit volume
Table 6 Results of the statistical analysis for selecting of a suitable model for unit volume weight. Sample name
Source
R2
Adjusted R2
Predicted R2
Kalecik Clay
Linear 2FI Quadratic Cubic
0.9069 0.9123 0.9237 0.9858
0.8994 0.9013 0.9063 0.9786
0.8833 0.8778 0.8675 0.9506
Suggested
Unit volume weightðUVWÞ ¼ þ1:20−1:41 ðTempÞ−0:16 ðTimeÞ 2 −0:089 ðTempÞ ðTimeÞ− 0:12 ðTempÞ 2 2 þ0:10 ðTimeÞ þ 0:17 ðTempÞ ðTimeÞ 2 3 þ0:047 ðTempÞ ðTimeÞ þ 0:68 ðTempÞ 3 −0:025 ðTimeÞ :
Sample name
Factor
Coefficient estimate
Degree of freedom
Standard Error
Kalecik Clay
Intercept (Temp) (Time) (Temp)(Time) (Temp)2 (Time)2 (Temp)2(Time) (Temp)(Time)2 (Temp)3 (Time)3
1.20 −1.41 −0.16 −0.089 −0.12 0.10 0.17 0.047 0.68 −0.025
1 1 1 1 1 1 1 1 1 1
0.033 0.076 0.085 0.034 0.044 0.038 0.059 0.057 0.082 0.085
A. Ozguven / Applied Clay Science 101 (2014) 533–540
537
Table 8 The ANOVA of the regression model for Kalecik Clay. Sample name
Source
Sum of squares
df
Mean square
F value
P-value prob N F
Kalecik Clay
Model (Temp) (Time) (Temp)(Time) (Temp)2 (Time)2 (Temp)2(Time) (Temp)(Time)2 (Temp)3 (Time)3 Residual Cor total
10.03 2.78 0.030 0.055 0.059 0.058 0.070 5.432E−003 0.56 6.864E−004 0.14 10.18
9 1 1 1 1 1 1 1 1 1 18 27
1.11 2.78 0.030 0.055 0.059 0.058 0.070 5.432E−003 0.56 6.864E−004 8.048E−003
138.49 345.21 3.75 6.81 7.29 7.16 8.76 0.67 68.97 0.085
b0.0001 b0.0001 0.0685 0.0177 0.0147 0.0154 0.0084 0.4221 b0.0001 0.7736
Significant
Fig. 4. The estimation graph of the unit volume weight model for Kalecik Clay.
It is possible to test whether a regression model is statistically significant or not by means of variance analysis method. As well as this, there are different approaches serving for the same purpose. One of them is to examine the scatter diagram between the observed results obtained from the experimental studies and the results obtained from the model equation (Fig. 5). When Fig. 5 is examined, it is seen that the results obtained from the model equation well reflect the real condition.
3.4. Statistical assessment related to expansion ratio Statistical analysis was performed with different kinds of models given in Table 9 for estimating the expansion ratio by using temperature and time. The quadratic model for the Kalecik Clay sample was found to be the best model for estimating the expansion ratio (Table 9). The validity of the quadratic model was tested with variance analysis. Results of the multiple regression analysis are given in Table 10 and the results of the variance analysis are given in Table 11. The model based on the regression coefficients given in Table 10 is statistically significant at 99% (α = 0.01) confidence level (*P = 0.0028 b α = 0.01). The estimation graph of the developed model is given in Fig. 6. The equation of the quadratic model for Kalecik Clay that was obtained according to Table 10 is as follows: Expansion ratioðERÞ ¼ 2:02 þ 2:11 ðTempÞ þ 0:42 ðTimeÞ 2 þ0:54 ðTempÞ ðTimeÞ þ 1:31 ðTempÞ 2 −0:35 ðTimeÞ :
Table 9 Results of the statistical for selecting of a suitable model for expansion ratio.
Fig. 5. The relationships between observed values from experimental studies and predicted values obtained from the unit volume weight model for Kalecik Clay.
Sample name
Source
R2
Adjusted R2
Predicted R2
Kalecik Clay
Linear 2FI Quadratic Cubic
0.7996 0.8271 0.9345 0.9517
0.7835 0.8055 0.9196 0.9275
0.7454 0.7576 0.8802 0.8445
Suggested
538
A. Ozguven / Applied Clay Science 101 (2014) 533–540
Table 10 The results of the multiple regression analysis for expansion ratio. Sample name
Factor
Coefficient estimate
Kalecik Clay Intercept 2.02 (Temp) 2.11 (Time) 0.42 (Temp)(Time) 0.54 2 1.31 (Temp) 2 (Time) −0.35
Degree of freedom
Standard error
1 1 1 1 1 1
0.17 0.13 0.12 0.18 0.23 0.20
Table 11 The ANOVA of the regression model for Kalecik Clay. Sample name
Source
Sum of squares
Kalecik Clay
Model 67.68 (Temp) 55.16 (Time) 2.74 (Temp)(Time) 1.99 (Temp)2 7.09 2 (Time) 0.69 Residual 4.74 Cor total 72.42
df
Mean square
5 13.54 1 55.16 1 2.74 1 1.99 1 7.09 1 0.69 22 0.22 27
F value
P-value prob N F
62.78 b0.0001 255.85 b0.0001 12.73 0.0017 9.24 0.0060 32.87 b0.0001 3.21 0.0871
Significant Fig. 7. Relationships between observed values from experimental studies and predicted values obtained from the expansion ratio model for Kalecik Clay.
Table 12 Design constraints for optimization.
The relationship between the observed values of the expansion ratio from experimental studies and the predicted values from the expansion ratio model was also investigated and the result is given in Fig. 7. Fig. 7 indicates that the results obtained from the model well reflect the real condition. The results obtained from the statistical analysis show that developed unit volume weight and expansion ratio models are statistically significant and unit volume weight and expansion ratio can be modeled by these ways. 3.5. Optimization The main purpose of this study is to determine the optimum time and temperature values that would maximize the expansion ratio values and minimize the unit volume weight values in expansion of Kalecik Clay and also to develop expandability charts for the Kalecik Clay sample. For this purpose, the aforesaid Design Expert 7.0 program was used, and firstly the constraints were defined (Table 12). Later,
Sample name
Name
Goal
Lower limit
Upper limit
Importance
Kalecik Clay
Temperature (°C) Time (min) UVW Expansion ratio (%)
Is in range Is in range Minimize Maximize
900 5 0.36 1
1200 20 1.89 5.5
3 3 5 5
the constraints were used to determine the optimum points for the Kalecik Clay sample. The optimum points were determined by using the Design Expert 7.0 program for the Kalecik Clay sample considering the design constraints. The results are given in Table 13. By using the model equations obtained from the statistical analyses, charts for the Kalecik Clay sample was developed with respect to unit volume weight and expansion ratio separately. The results are given in Fig. 8 for unit volume weight and expansion ratio. The charts show the optimum working conditions and estimated unit volume weight and expansion ratio values that will occur under these conditions.
Fig. 6. The estimation graph of the expansion ratio model for Kalecik Clay.
A. Ozguven / Applied Clay Science 101 (2014) 533–540 Table 13 The optimum working conditions for Kalecik Clay. Temperature (°C)
Time (min)
UVW (g/cm3)
Expansion ratio (%)
Desirability level
1199.26
15.09
0.337518
5.69897
1.00
4. Conclusions In this study, the effects of different operational parameters including holding time in the furnace and furnace temperature on the expansion efficiency, namely expansion ratio and unit volume weight in the expanding operation of Kalecik Clay were investigated. However, optimization of the working parameters in reference to the expanding efficiency parameters and also development of expandability charts for the Kalecik Clay samples were aimed in this study. The results obtained are given as follows. • It was found that there is a characteristic relationship between furnace temperature and expansion ratio and unit volume weight. As the furnace temperature increases, the expansion ratio increases while the unit volume weight decreases. • The quadratic model and cubic model were found to be the best models for predicting expansion ratio and unit volume
539
weight values, respectively, in terms of the R2 value. Independent variables including holding time in the furnace and furnace temperature explain over 90% of both the dependent variables (expansion ratio and unit volume weight) in the developed model equations. The expansion ratio and unit volume weight could be determined by using these two model equations and also developed model graphs for them before starting the expanding operation, because the values of the expansion ratio and unit volume weight obtained by the models reflect the real conditions well. • The optimum working parameters for the Kalecik Clay sample, namely a temperature of 1199.26 °C and a holding time in the furnace of 15.09 min were found. In these optimum working conditions, a unit volume weight of 0.3375 g/cm3 and an expansion ratio of 5.699% were found to be the expanding values. These values are in good agreement with the observed values. • Expandability charts for the Kalecik Clay samples based on expansion ratio and unit volume weight were developed in this study. By using these charts, it is possible to determine expansion ratio and unit volume weight values in different working conditions including holding time in the furnace and furnace temperature. These charts could be of help to the construction industry in terms of a more efficient production of expanded clay aggregates.
Fig. 8. Expandability charts including the optimum working conditions with respect to unit volume weight (a) and expansion ratio (b) for Kalecik Clay.
540
A. Ozguven / Applied Clay Science 101 (2014) 533–540
As a result of this study, it is possible to determine the optimum working parameters for different kinds of clays by using this type of study. Acknowledgments The author wishes to thank Dr. Yilmaz Ozcelik of Hacettepe University for his valuable comments and suggestions in preparing the manuscript. References ASTM, 1998. Standard Test Method for Bulk Density and Porosity of Granular Refractory Materials by Mercury Displacement (Number: ASTM C493-98). Bartolini, R., Filippozzi, S., Princi, E., Schenone, C., Vicini, S., 2010. Acoustic and mechanical properties of expanded clay granulates consolidated by epoxy resin. Appl. Clay Sci. 48, 460–465. Bragdon, P.W.E., 1996. Development of High Performance Lightweight Aggregate from New Brunswick Raw MaterialsMaster Thesis The University of New Brunswick, (Canada). Cougny, G., 1990. Spécifications sur les matières premières argileuses pour la fabrication de granulats légers expansés. Bull. Int. Assoc. Eng. Geol. 41, 47–55.
de Gennaro, R., Cappelletti, P., Cerri, G., de' Gennaro, M., Dondi, M., Graziano, S.F., Langella, A., 2007. Campanian Ignimbrite as raw material for lightweight aggregates. Appl. Clay Sci. 37, 115–126. Decleer, J., Viaene, W., 1993. Rupelian Boom clay as raw material for expanded clay manufacturing. Appl. Clay Sci. 8, 111–128. Fakhfakh, E., Hajjaji, W., Medhioub, M., Rocha, F., López-Galindo, A., Setti, M., Kooli, F., Zargouni, F., Jamoussi, F., 2007. Effects of sand addition on production of lightweight aggregates from Tunisian smectite-rich clayey rocks. Appl. Clay Sci. 35, 228–237. Fragoulis, D., Stamatakis, G., Chaniotakis, E., Columbus, G., 2004. Characterization of lightweight aggregates produced with clayey diatomite rocks originated from Greece. Mater. Charact. 53, 307–316. Ozcelik, Y., Kanbir, E.S., 2011. Determination of optimum working condition and development of cuttability abacus for a limestone and real marble sample in diamond wire cutting. 46th U.S. Rock Mechanics Geomechanics Symposium. Ozcelik, Y., Gursel, M., Ciccu, R., Costa, G., Bortolussi, A., 2012. Optimization of working parameters of water jet cutting in terms of depth and width of cut. J. Process. Mech. Eng. Proc. Inst. Mech. Eng. Part E 226, 64–78. Ozguven, A., 2009. Genleşen Kil Agrega Üretimi ve Endüstriyel Olarak Değerlendirilmesi. Ph.D. Thesis Suleyman Demirel University, Isparta, (in Turkish). Pioro, L.S., Pioro, I.L., 2004. Production of expanded-clay aggregate for lightweight concrete from non-selfbloating clays. Cem Concr Compos 26, 639–643. Rattanachan, S., Lorprayoon, C., 2005. Korat clays as raw materials for lightweight aggregates. ScienceAsia 31, 277–281.
Contents lists available at ScienceDirect
Applied Clay Science journal homepage: www.elsevier.com/locate/clay
Research paper
Development of expandability charts for Ankara Kalecik Clay Ahmet Ozguven ⁎ General Directorate of Mineral Research and Exploration, 06800 Ankara, Turkey
a r t i c l e
i n f o
Article history: Received 7 February 2012 Received in revised form 7 July 2014 Accepted 28 August 2014 Available online 8 October 2014 Keywords: Lightweight aggregate Expanded clay aggregates Working parameters Expandability Optimization
a b s t r a c t The main aim of this study is to establish optimum working conditions and to propose expandability charts for Kalecik Clay using expansion ratio versus unit volume weight values via a statistical computer program (Design Expert 7.0). First of all, some suitable samples were collected from the Kalecik County (Ankara, Turkey). Later, the samples were crushed and milled up to 200 μm clay size in the laboratory. Then, pellets with a 10 mm diameter were dried in ovens and heated inside a furnace to expand in different temperatures. One heating operation began with 900 °C and ended with 1200 °C. In addition to that, the holding time periods were selected as 5, 10, 15 or 20 min for each operation. In the final stage, the unit volume weight and expansion ratio of the produced aggregates were determined. As results of this study, both the optimum furnace holding time and furnace temperature values for the samples were determined in order to maximize the expansion ratio values and minimize the unit volume weight for Kalecik Clays. Some expandability charts were proposed for Kalecik Clay by using the findings. The charts allow one to assign expansion ratio and unit volume weight values in different working conditions including holding time inside the furnace and furnace temperature. The proposed charts are expected to be quite useful for more efficient production of expanded clay aggregates in construction industry. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Lightweight aggregates are defined as natural or artificial materials, which are granular, porous and lightweight (Cougny, 1990). They can originate from different natural sources such as volcanic rocks (pumice, volcanic tuffs), sedimentary rocks (clayey diatomite) and metamorphic rocks (claystones, slates, shales), and from waste material or industrial by-products such as recycled flat glass and fly ash (Fakhfakh et. al., 2007). The lightweight aggregates are used as a loose material — for example in back wall fillers and in agronomic applications — or, with a binder, in the manufacture of plaster, asphalt and lightweight concrete thermo-acoustic insulators, as well as in lightweight structural concrete production. Lightweight expanded aggregates can be formed by a quick heating at high temperature of some rocks which are able to bloat. The raw material must contain substances that develop gas upon heating and, at the same time, the material must transform into a highly viscous plastic mass able to expand by virtue of gas entrapment (de Gennaro et. al, 2007). There are a lot of studies about the production of expanded clay aggregates (Decleer and Viaene, 1993; Bragdon, 1996; Fragoulis et. al., 2004; Pioro and Pioro, 2004; Rattanachan and Lorprayoon, 2005; Fakhfakh et. al., 2007; de Gennaro et. al, 2007; Ozguven, 2009; Bartolini et al, 2010). ⁎ Tel.: +90 312 2011227; fax: +90 312 2878747. E-mail address: [email protected].
http://dx.doi.org/10.1016/j.clay.2014.08.028 0169-1317/© 2014 Elsevier B.V. All rights reserved.
As can be seen from the literature, there are some important studies on the parameters, taken one by one, affecting the expansion of the clays for producing lightweight expanded clay aggregates. However, so far no study has been carried out to investigate the combined influence of working parameters (holding time in the furnace and furnace temperature) on expansion efficiency parameters, jointly expansion ratio and unit volume weight, to determine the optimum working conditions and also to develop the expandability charts with expansion ratio and unit volume weight. For this reason, the purpose of this study is to determine the optimum working conditions and also to develop expandability charts for Kalecik Clay with respect to expansion ratio and unit volume weight by using a special statistical program (Design Expert 7.0) providing the factorial design of the experiment which is very important in the optimization of the experimental results. 2. Material and method The methodology of this study is composed of four steps. 1 Sampling and preparation of the samples 2 Sampling of clay in a constant pellet size of 10 mm and clay grain size of 200 μm. 3 Expansion tests 4 Statistical assessment of the test results Considering these four steps, the expansion operation methodology and the work followed in this study are given in detail in Fig. 1.
534
A. Ozguven / Applied Clay Science 101 (2014) 533–540
Fig. 1. Methodology followed in this study.
The studies were conducted in clays taken from the Ankara City's Kalecik County. From a geological point of view, of the expanding clay fields under study, the Ankara Kalecik field has a grayish metallic luster, while it is seen as dark gray-black when wet. Secondary calcite was developed in some places. It has sandstone-shale alteration with a schist-like view behind limestone lenses. This unit is overlain by grayish-gray color sandstone-shale faces. It is a unit exposed in macro-scale achieving 100 m from place to place along approximately
a 1-km path in the form of 25–30 m lenses along the north–south direction. X-ray diffraction pattern belonging to the samples obtained from the field under study is given in Fig. 2 and the result from the X-ray diffraction analysis is seen in Table 1. Chemical analysis (Table 2) showed that the main constituents of the raw materials are silica, alumina, and iron oxides. Na2O and K2O contents are mainly attributed to the clay minerals and feldspars.
A. Ozguven / Applied Clay Science 101 (2014) 533–540
535
Fig. 2. X-ray diffraction pattern belonging to the sample.
The samples collected from the fields were broken and then milled to −200 μm clay size. Ground clay was mixed with only water without any additive to produce clay dough. Dough preparations were left for maturation for one day and shaped through an extruder. Pellets were prepared in a 10 mm diameter and the diameter/length ratio of the pellets is approximately 1:1. The prepared pellets were dried in ovens and then expanded in a furnace. Firing processes took place at different temperatures beginning from 900 °C up to 1200 °C. Raw pellets were kept inside the furnace for different time periods at the same temperature (5, 10, 15 or 20 min). Fired pellets were removed from the furnace and cooled suddenly. Unit volume weight of the produced aggregates was measured by using the ASTM C493-98 standard (ASTM, 1998). The found unit volume weight of the expanded clay aggregates (UVWexp) was compared with the unit volume weight of the raw pellets (UVWorj) to calculate the expansion ratio. The expansion ratio is calculated as (UVWorj / UVWexp) × 100.
the determination process of the optimum points by means of the derived equations. Imaging the optimum working points and the estimated results obtained as a result of the experiments made on these points is possible with this program (Ozcelik and Kanbir, 2011; Ozcelik et. al. 2012). 3.2. Design summary Before initializing the statistical analysis, information that reflects the properties of each variable regarding the factors and response was analyzed. This information is composed of the data such as mean and standard deviation for describing the frequencies about the variables and also definitions of the methods and models to be used in modeling studies. The design properties used in this study are demonstrated in Table 3 and the descriptive statistical data regarding factors and response are shown in Tables 4 and 5, respectively. In addition to the descriptive statistics, the relationships between unit volume weight and expansion ratio and temperature at various
3. Statistical analysis Within the scope of this study, Design Expert 7.0, which is a special statistical program, was used for data analysis and for determining the optimum values of furnace temperature and holding time in the furnace for the Kalecik Clay sample. The method of the statistical analysis performed in this study is given in Fig. 1. 3.1. Design Expert 7.0 statistical software Design Expert 7.0 is a widely used program which was developed for the experimental optimization process and which can effectively design the experiments in the most suitable way according to different methods. After making the experiments based on the design selected and entering the results obtained in the program, it derives the most suitable equations for dependent variables (response) and can realize
Table 1 XRD analysis result of the clay under study. Fields under study
Minerals
Ankara
Quartz (ASTM No: 5-0490), illite, chlorite, mixed layered clay minerals, feldspar, calcite (ASTM No: 5-0586), (graphite)
Table 2 Chemical composition of the studied samples. Na2O MgO Al2O3 SiO2 P2O5 K2O CaO TiO2 MnO Fe2O3 Fluxing (Na2O + MgO + K2O + CaO + Fe2O3) SiO2/Al2O3 SiO2/fluxing
0.98 2.54 18.16 56.11 0.16 3.73 3.04 0.70 0.19 6.89 17.18 3.09 3.27
Table 3 Statistical design properties used in the study. Study type Initial design Design model Runs Blocks
Factorial D-optimal, Point Exchange 2FI 28 No blocks
536
A. Ozguven / Applied Clay Science 101 (2014) 533–540
Table 4 Descriptive statistical data of the factors. Sample name
Factor name
Units
Type
Low actual
High actual
Mean
Std. dev.
Kalecik Clay
Temperature (Temp) Time (Time)
(°C) (min)
Numeric Numeric
900.00 5.00
1200.00 20.00
1050.000 12.500
100.000 5.590
Table 5 Descriptive statistical data of the responds. Sample name
Response name
Units
Obs
Analysis
Min.
Max.
Mean
Std. dev.
Model
Kalecik Clay
Unit volume weight (UVW) Expansion ratio (ER)
(g/cm3) (%)
28 28
Polynomial Polynomial
0.36 1.00
1.89 5.50
1.20 2.41
0.60 1.61
Cubic Quadratic
Fig. 3. Relationships between unit volume weight and expansion ratio and time at various temperatures for Kalecik Clay.
holding times in the furnace are shown in Fig. 3. As can be seen from Fig. 3, as the furnace temperature increases, the expansion ratio increases while the unit volume weight decreases. The complexity of most scientific mechanisms is such that in order to be able to predict an important response, a well-known multiple regression model is needed. The model formed for regression analysis is a model which involves dependent and independent variables. In such a model, the change in dependent variable is tried to be explained with independent variables. The first phase in forming simple or multiple regression models is to determine the coefficients that form the model and then the statistical test of the validity of the model with variance analysis. The figures demonstrating the results obtained from here are called ANOVA (Analysis of Variance). F value obtained from the figures is the value to be tested for the validity of the general model. F value and the results of ANOVA in regression analysis and in other models show the validity of the model formed, in other words, the representativeness of the system (Ozcelik and Kanbir, 2011; Ozcelik et. al. 2012).
weight model equation was obtained. The cubic model for the Kalecik Clay sample, which was found to be statistically the most significant in the analysis, was chosen as the most suitable model for estimating the unit volume weight (Table 6). The validity of the cubic model was tested with variance analysis. The results were presented in Tables 7 and 8, respectively. The model based on the regression coefficients given in Table 7 is statistically significant at 99% (α = 0.01) confidence level (*P = 0.0001 b α = 0.01). The estimation graph of the developed model is given in Fig. 4. The equation of the cubic model for Kalecik Clay that was obtained according to Table 7 is as follows:
3.3. Statistical assessment related to unit volume weight
Table 7 The results of the multiple regression analysis for unit volume weight.
Statistical analysis was made for the purpose of estimating unit volume weight by using time and temperature and the unit volume
Table 6 Results of the statistical analysis for selecting of a suitable model for unit volume weight. Sample name
Source
R2
Adjusted R2
Predicted R2
Kalecik Clay
Linear 2FI Quadratic Cubic
0.9069 0.9123 0.9237 0.9858
0.8994 0.9013 0.9063 0.9786
0.8833 0.8778 0.8675 0.9506
Suggested
Unit volume weightðUVWÞ ¼ þ1:20−1:41 ðTempÞ−0:16 ðTimeÞ 2 −0:089 ðTempÞ ðTimeÞ− 0:12 ðTempÞ 2 2 þ0:10 ðTimeÞ þ 0:17 ðTempÞ ðTimeÞ 2 3 þ0:047 ðTempÞ ðTimeÞ þ 0:68 ðTempÞ 3 −0:025 ðTimeÞ :
Sample name
Factor
Coefficient estimate
Degree of freedom
Standard Error
Kalecik Clay
Intercept (Temp) (Time) (Temp)(Time) (Temp)2 (Time)2 (Temp)2(Time) (Temp)(Time)2 (Temp)3 (Time)3
1.20 −1.41 −0.16 −0.089 −0.12 0.10 0.17 0.047 0.68 −0.025
1 1 1 1 1 1 1 1 1 1
0.033 0.076 0.085 0.034 0.044 0.038 0.059 0.057 0.082 0.085
A. Ozguven / Applied Clay Science 101 (2014) 533–540
537
Table 8 The ANOVA of the regression model for Kalecik Clay. Sample name
Source
Sum of squares
df
Mean square
F value
P-value prob N F
Kalecik Clay
Model (Temp) (Time) (Temp)(Time) (Temp)2 (Time)2 (Temp)2(Time) (Temp)(Time)2 (Temp)3 (Time)3 Residual Cor total
10.03 2.78 0.030 0.055 0.059 0.058 0.070 5.432E−003 0.56 6.864E−004 0.14 10.18
9 1 1 1 1 1 1 1 1 1 18 27
1.11 2.78 0.030 0.055 0.059 0.058 0.070 5.432E−003 0.56 6.864E−004 8.048E−003
138.49 345.21 3.75 6.81 7.29 7.16 8.76 0.67 68.97 0.085
b0.0001 b0.0001 0.0685 0.0177 0.0147 0.0154 0.0084 0.4221 b0.0001 0.7736
Significant
Fig. 4. The estimation graph of the unit volume weight model for Kalecik Clay.
It is possible to test whether a regression model is statistically significant or not by means of variance analysis method. As well as this, there are different approaches serving for the same purpose. One of them is to examine the scatter diagram between the observed results obtained from the experimental studies and the results obtained from the model equation (Fig. 5). When Fig. 5 is examined, it is seen that the results obtained from the model equation well reflect the real condition.
3.4. Statistical assessment related to expansion ratio Statistical analysis was performed with different kinds of models given in Table 9 for estimating the expansion ratio by using temperature and time. The quadratic model for the Kalecik Clay sample was found to be the best model for estimating the expansion ratio (Table 9). The validity of the quadratic model was tested with variance analysis. Results of the multiple regression analysis are given in Table 10 and the results of the variance analysis are given in Table 11. The model based on the regression coefficients given in Table 10 is statistically significant at 99% (α = 0.01) confidence level (*P = 0.0028 b α = 0.01). The estimation graph of the developed model is given in Fig. 6. The equation of the quadratic model for Kalecik Clay that was obtained according to Table 10 is as follows: Expansion ratioðERÞ ¼ 2:02 þ 2:11 ðTempÞ þ 0:42 ðTimeÞ 2 þ0:54 ðTempÞ ðTimeÞ þ 1:31 ðTempÞ 2 −0:35 ðTimeÞ :
Table 9 Results of the statistical for selecting of a suitable model for expansion ratio.
Fig. 5. The relationships between observed values from experimental studies and predicted values obtained from the unit volume weight model for Kalecik Clay.
Sample name
Source
R2
Adjusted R2
Predicted R2
Kalecik Clay
Linear 2FI Quadratic Cubic
0.7996 0.8271 0.9345 0.9517
0.7835 0.8055 0.9196 0.9275
0.7454 0.7576 0.8802 0.8445
Suggested
538
A. Ozguven / Applied Clay Science 101 (2014) 533–540
Table 10 The results of the multiple regression analysis for expansion ratio. Sample name
Factor
Coefficient estimate
Kalecik Clay Intercept 2.02 (Temp) 2.11 (Time) 0.42 (Temp)(Time) 0.54 2 1.31 (Temp) 2 (Time) −0.35
Degree of freedom
Standard error
1 1 1 1 1 1
0.17 0.13 0.12 0.18 0.23 0.20
Table 11 The ANOVA of the regression model for Kalecik Clay. Sample name
Source
Sum of squares
Kalecik Clay
Model 67.68 (Temp) 55.16 (Time) 2.74 (Temp)(Time) 1.99 (Temp)2 7.09 2 (Time) 0.69 Residual 4.74 Cor total 72.42
df
Mean square
5 13.54 1 55.16 1 2.74 1 1.99 1 7.09 1 0.69 22 0.22 27
F value
P-value prob N F
62.78 b0.0001 255.85 b0.0001 12.73 0.0017 9.24 0.0060 32.87 b0.0001 3.21 0.0871
Significant Fig. 7. Relationships between observed values from experimental studies and predicted values obtained from the expansion ratio model for Kalecik Clay.
Table 12 Design constraints for optimization.
The relationship between the observed values of the expansion ratio from experimental studies and the predicted values from the expansion ratio model was also investigated and the result is given in Fig. 7. Fig. 7 indicates that the results obtained from the model well reflect the real condition. The results obtained from the statistical analysis show that developed unit volume weight and expansion ratio models are statistically significant and unit volume weight and expansion ratio can be modeled by these ways. 3.5. Optimization The main purpose of this study is to determine the optimum time and temperature values that would maximize the expansion ratio values and minimize the unit volume weight values in expansion of Kalecik Clay and also to develop expandability charts for the Kalecik Clay sample. For this purpose, the aforesaid Design Expert 7.0 program was used, and firstly the constraints were defined (Table 12). Later,
Sample name
Name
Goal
Lower limit
Upper limit
Importance
Kalecik Clay
Temperature (°C) Time (min) UVW Expansion ratio (%)
Is in range Is in range Minimize Maximize
900 5 0.36 1
1200 20 1.89 5.5
3 3 5 5
the constraints were used to determine the optimum points for the Kalecik Clay sample. The optimum points were determined by using the Design Expert 7.0 program for the Kalecik Clay sample considering the design constraints. The results are given in Table 13. By using the model equations obtained from the statistical analyses, charts for the Kalecik Clay sample was developed with respect to unit volume weight and expansion ratio separately. The results are given in Fig. 8 for unit volume weight and expansion ratio. The charts show the optimum working conditions and estimated unit volume weight and expansion ratio values that will occur under these conditions.
Fig. 6. The estimation graph of the expansion ratio model for Kalecik Clay.
A. Ozguven / Applied Clay Science 101 (2014) 533–540 Table 13 The optimum working conditions for Kalecik Clay. Temperature (°C)
Time (min)
UVW (g/cm3)
Expansion ratio (%)
Desirability level
1199.26
15.09
0.337518
5.69897
1.00
4. Conclusions In this study, the effects of different operational parameters including holding time in the furnace and furnace temperature on the expansion efficiency, namely expansion ratio and unit volume weight in the expanding operation of Kalecik Clay were investigated. However, optimization of the working parameters in reference to the expanding efficiency parameters and also development of expandability charts for the Kalecik Clay samples were aimed in this study. The results obtained are given as follows. • It was found that there is a characteristic relationship between furnace temperature and expansion ratio and unit volume weight. As the furnace temperature increases, the expansion ratio increases while the unit volume weight decreases. • The quadratic model and cubic model were found to be the best models for predicting expansion ratio and unit volume
539
weight values, respectively, in terms of the R2 value. Independent variables including holding time in the furnace and furnace temperature explain over 90% of both the dependent variables (expansion ratio and unit volume weight) in the developed model equations. The expansion ratio and unit volume weight could be determined by using these two model equations and also developed model graphs for them before starting the expanding operation, because the values of the expansion ratio and unit volume weight obtained by the models reflect the real conditions well. • The optimum working parameters for the Kalecik Clay sample, namely a temperature of 1199.26 °C and a holding time in the furnace of 15.09 min were found. In these optimum working conditions, a unit volume weight of 0.3375 g/cm3 and an expansion ratio of 5.699% were found to be the expanding values. These values are in good agreement with the observed values. • Expandability charts for the Kalecik Clay samples based on expansion ratio and unit volume weight were developed in this study. By using these charts, it is possible to determine expansion ratio and unit volume weight values in different working conditions including holding time in the furnace and furnace temperature. These charts could be of help to the construction industry in terms of a more efficient production of expanded clay aggregates.
Fig. 8. Expandability charts including the optimum working conditions with respect to unit volume weight (a) and expansion ratio (b) for Kalecik Clay.
540
A. Ozguven / Applied Clay Science 101 (2014) 533–540
As a result of this study, it is possible to determine the optimum working parameters for different kinds of clays by using this type of study. Acknowledgments The author wishes to thank Dr. Yilmaz Ozcelik of Hacettepe University for his valuable comments and suggestions in preparing the manuscript. References ASTM, 1998. Standard Test Method for Bulk Density and Porosity of Granular Refractory Materials by Mercury Displacement (Number: ASTM C493-98). Bartolini, R., Filippozzi, S., Princi, E., Schenone, C., Vicini, S., 2010. Acoustic and mechanical properties of expanded clay granulates consolidated by epoxy resin. Appl. Clay Sci. 48, 460–465. Bragdon, P.W.E., 1996. Development of High Performance Lightweight Aggregate from New Brunswick Raw MaterialsMaster Thesis The University of New Brunswick, (Canada). Cougny, G., 1990. Spécifications sur les matières premières argileuses pour la fabrication de granulats légers expansés. Bull. Int. Assoc. Eng. Geol. 41, 47–55.
de Gennaro, R., Cappelletti, P., Cerri, G., de' Gennaro, M., Dondi, M., Graziano, S.F., Langella, A., 2007. Campanian Ignimbrite as raw material for lightweight aggregates. Appl. Clay Sci. 37, 115–126. Decleer, J., Viaene, W., 1993. Rupelian Boom clay as raw material for expanded clay manufacturing. Appl. Clay Sci. 8, 111–128. Fakhfakh, E., Hajjaji, W., Medhioub, M., Rocha, F., López-Galindo, A., Setti, M., Kooli, F., Zargouni, F., Jamoussi, F., 2007. Effects of sand addition on production of lightweight aggregates from Tunisian smectite-rich clayey rocks. Appl. Clay Sci. 35, 228–237. Fragoulis, D., Stamatakis, G., Chaniotakis, E., Columbus, G., 2004. Characterization of lightweight aggregates produced with clayey diatomite rocks originated from Greece. Mater. Charact. 53, 307–316. Ozcelik, Y., Kanbir, E.S., 2011. Determination of optimum working condition and development of cuttability abacus for a limestone and real marble sample in diamond wire cutting. 46th U.S. Rock Mechanics Geomechanics Symposium. Ozcelik, Y., Gursel, M., Ciccu, R., Costa, G., Bortolussi, A., 2012. Optimization of working parameters of water jet cutting in terms of depth and width of cut. J. Process. Mech. Eng. Proc. Inst. Mech. Eng. Part E 226, 64–78. Ozguven, A., 2009. Genleşen Kil Agrega Üretimi ve Endüstriyel Olarak Değerlendirilmesi. Ph.D. Thesis Suleyman Demirel University, Isparta, (in Turkish). Pioro, L.S., Pioro, I.L., 2004. Production of expanded-clay aggregate for lightweight concrete from non-selfbloating clays. Cem Concr Compos 26, 639–643. Rattanachan, S., Lorprayoon, C., 2005. Korat clays as raw materials for lightweight aggregates. ScienceAsia 31, 277–281.