Performance of Lime Kiln Dust as Cementitious Material

Available online at www.sciencedirect.com

ScienceDirect Procedia Engineering 125 (2015) 780 – 787

The 5th International Conference of Euro Asia Civil Engineering Forum (EACEF-5)

Performance of lime kiln dust as cementitious material Masimawati Abdul Latifa,b,*, Sivakumar Naganathanb, Hashim Abdul Razakc, Kamal Nasharuddin Mustaphab a

Human Resource Department, Ministry of Education, Level 15, No. 2, Menara 2, Jalan P5/6, Presint 5, 62200 W.P. Putrajaya b Department of Civil Engineering, Universiti Tenaga Nasional, Jalan IKRAM-UNITEN, 43000 Kajang, Selangor, Malaysia c StrucHMRS Group, Department of Civil Engineering, Faculty of Engineering, University of Malaya, 50603 Kuala Lumpur, Malaysia

Abstract Lime Kiln Dust (LKD) is the by-product of Quick Lime (QL) production which is mostly disposed off into landfills globally. Therefore, LKD needs to be reused to reduce environmental problems and promote sustainability. This paper presents preliminary investigations on the use of LKD as a cementitious material in mortar. LKD was added to mortar mixtures to study its effect on the strength by replacing ordinary Portland cement (OPC) up to 60% by weight with LKD. Experimental results on physical and engineering properties, and compressive strengths of mortar samples containing LKD are reported. Results indicated that the initial and final setting time of the LKD mortar tested ranged from 1.08 to 2.3h and 4.58 to 5.8h respectively, and the strength values from 36.35 to 68.08MPa for 90 days strength. Furthermore, addition of LKD reduced the strength of the mortar but with a faster setting time. It can be concluded that there is a good potential for the use of LKD as a cementitious material with accelerated hydration rates. © 2015 2015The TheAuthors. Authors. Published Elsevier © Published by by Elsevier Ltd.Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of organizing committee of The 5th International Conference of Euro Asia Civil Engineering Peer-review under responsibility of organizing committee of The 5th International Conference of Euro Asia Civil Engineering Forum (EACEF-5). Forum (EACEF-5)

Keywords: Lime Kiln Dust; Cementitious Material; Physical Properties; Density; Compressive Strength.

1. Introduction Portland cement is one of the important materials in the construction industry due to its economic advantage and the quick-setting property [1]. The production of 1 ton of Portland cement will contribute to the emissions of 0.7 –

* Corresponding author. Tel.: +6019-5737197; E-mail address: [email protected]

1877-7058 © 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license

(http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of organizing committee of The 5th International Conference of Euro Asia Civil Engineering Forum (EACEF-5)

doi:10.1016/j.proeng.2015.11.135

781

Masimawati Abdul Latif et al. / Procedia Engineering 125 (2015) 780 – 787

1.1 tons of carbon dioxides (CO2) in the atmosphere [2]. In 1995, the world produced about 1.4 billion tons and increased to 4 billion tons of cement in the year 2013 [3]. This will also increase CO2 and dust emissions polluting the air, which causes the greenhouse effect and global warming [4]. As such, the production of Portland cement contributes to the world’s 5 percent of CO2 emissions. Additionally, the cement production is predicted to increase to 4,380 Mt by 2050 [5]. Hence there is need to reduce the use of Portland cement by introducing new technologies and strategy in cement production process to reduce CO2 emission and produce more energy efficient products to the market[5]. One of the strategy is the usage of supplementary cementitious material such as industrial waste as a partial cement replacement[6]. Globally, there are many types of industrial wastes produced annually. There are varieties of waste that can be used as supplementary pozzolanic and cementing materials such as Calcium Carbide Residue [6], Fly Ash [7], Cement Kiln Dust [8]and many others. Most of these waste are waste from by-products, industrial and agricultural waste. Since most wastes have no economically beneficial use, approximately 4.2 billion tons of non-hazardous byproducts has been disposed into landfills [9]. In 1994, Malaysia generated 7,721.58 ton/day of industrial waste and it has increased to 11,519.24 ton/day in 2005[10]. Industrial waste contributed to 25% off Malaysia waste compositions [11]. Lime Kiln Dust (LKD) is the by-product of Quick Lime (QL) production. QL used as a material in purification of steel, manufacturing of Calcium Carbide, effluent treatment for waste water and many more. QL is white in color and granular. QL production requires lime stone as its raw material. The process involves using natural gas to heat the lime stone (CaCO3) to the temperature of 800°C to 1000°C to turn CaCO3 into QL (CaO) as shown in equation 1. ‫ܱܥܽܥ‬ଷ ൅  ο՜ ‫ ܱܽܥ‬൅  ‫ܱܥ‬ଶ

(1)

The heating process is a continuous process throughout the lime kiln. This also generates CO 2 gases with dust or particulate matter (PM). The gas is filtered using fabric dust filter collector. The PM that is captured is called LKD [12]. LKD chemical compositions may vary for different plants, because it is influenced by type of lime stones, kiln, fuel used, and also the kiln operating parameters [12]. However, it generally contains relatively high percentage of CaO. It has been estimated that about 2.5 million metric tons of LKD produce in United States annually [12] and a single factory in Malaysia produces about 1.5 tons per day but there is still very few published research on the use of LKD as a cement replacement material. Globally, massive quantity of LKD is produced. Therefore, the needs to research about the performance of mortars containing LKD is increasing. The aim of this study is to experimentally investigate the potential benefits of replacing Ordinary Portland Cement (OPC) by up to 60% wt. with LKD in cement mortar. This was achieved by studying its specific gravity, Blaine fineness, particle size, standard consistency, setting time, density and strength for LKD mortar. 2. Materials and Methods 2.1. Materials Ordinary Portland cement (OPC) Type 1, LKD, manufactured silica sand and normal tap water were used in the investigation. LKD was collected from a factory in Taiping, Perak, Malaysia. The factory produces 250 tonnes/day QL and discharge approximately 1.5tonnes LKD per day which is disposed in landfill. In 2005, there was approximately 168 landfills area in Malaysia and 80% of them were filled and closed [11]. The collected LKD was stored in air tight container to avoid contacts with moisture and air because it can cause LKD to become agglomerate. In order to study the effect of LKD incorporation on the strength of cement mortar, OPC was replaced with LKD from 5% to 60% by cement weight. Manufactured silica sand (MS) was purchased from L&T Minerals, Malaysia and used as fine aggregates. The sand has specific gravity, fineness modulus and the maximum grain size of 2.69, 2.61 and 2.36 mm respectively. The sand comes in 4 different sizes of 8/16, 16/30, 30/60 and 50/100. Fig. 1 represents the particle size distribution of the sand in accordance with ASTM C33-03 [13]. Water absorptions of MS was tested in accordance with BS 812: Part 2: 1995 [14] is 2.1%.

782

Masimawati Abdul Latif et al. / Procedia Engineering 125 (2015) 780 – 787

100 90 Cumulative Passing (%)

80 70

60 50 40 30

Upper Limit

20

Manufactured silica sand Lower limit

10 0 0.150

0.300

0.600

1.18

2.36

4.75

Sieve size (mm) Fig. 1. Manufactured Silica Sand Sieve Analysis.

2.2. Admixtures Sika Viscocrete® 2199 is a modified polycarboxylate type super plasticizer (SP) and it was used throughout this study. It is high range water reducing admixture and were used when the mix did not achieve the target mortar spread. The SP used in this study was less than 0.6% of cement weight.This SP is compatible with all types of Portland cement including Sulfate Resistant Cement (SRC) and it is chloride free according to BS 5075 [15]. 2.3. Physical Properties of OPC and LKD Physical properties of OPC and LKD were summarized in Table 1. All LKD used in the mixes was sieved using sieve size 1.18mm only to remove impurities and there was 100% particles passing thru the sieve. In other words, the original LKD were used directly as a cementing agent. Materials specific gravity (SG) were determined according to the BS812: Part 2:1995[16]. OPC has higher SG and Blaine Fineness compared to LKD due to the high fine particles and denser nature of OPC over LKD. Table 1. Physical properties of OPC and LKD. Materials

Specific Gravity

Mean Particle Size, d50 (micron)

Blaine fineness (cm2/g)

OPC

3.16

29.67

3600

LKD

2.54

40.29

1747

Malvern Instrument Particle Size Analyser was used to determine the material grain size distribution by laser particle analysis as shown in Fig. 2. The LKD average grain size d50 is 40.29 μm compared to OPC d50 which is 29.67 μm. The grain size indicated that LKD is coarser with bigger particle sizes compared to OPC.

783

Masimawati Abdul Latif et al. / Procedia Engineering 125 (2015) 780 – 787

100 90 80

Percent Finer (%)

70 60 50 40 30 20 OPC LKD

10 0 0.01

0.1

1

10

100

1000

Particle Diameter, D (μm) Fig. 2. Grain Size Distribution of the OPC and LKD.

2.4. Chemical Compositions of OPC and LKD Bruker S8 Tiger, X-Ray Efflorescence (XRF) was used to analyse the chemical compositions of OPC and LKD presented in Table 2. The XRF was carried out 3 times for the same sample and the average is reported. The results show that, the main composition of LKD is CaO and SiO2. LKD has lower compositions in terms of CaO in comparison with OPC. LKD has 54.88% content of CaO while OPC has 62.79% CaO. The trend is the same for SiO2, Al2O3, Fe2O3, MgO and K2O with lower compositions for LKD compared to OPC. However, MgO content in LKD still conforms with BS12:1996 standard [17]. LKD has 0.3% SO3 which is less than 3.5% as stated in BS12:1996 [17]. This indicates that LKD has potential to be used partial OPC replacement as it has pozzolanic property. The LKD loss of ignition (L.O.I) is 38.59 which is significantly higher than OPC L.O.I (2.32). Therefore LKD has exceeded the maximum L.O.I which is less than 3% as the limit stated in BS12:1996. This is similar with Najim (2014) research on Cement kiln Dust (CKD) replacement of OPC whereby the CKD L.O.I exceeded the limit too [18]. Table 2. Chemical Properties of OPC and LKD. Oxides

CaO

SiO2

Al2O3

SO3

Fe2O3

MgO

K2O

Na2O

TiO2

P2O5

LOI

OPC

62.79

18.72

4.44

3.28

3.22

2.97

0.19

0.14

0.13

0.13

2.32

LKD

54.88

0.67

0.68

0.3

0.38

0.95

0.85

0.01

0.01

0.01

38.59

2.5. Standard Consistency and Setting Times LKD mixes standard consistency and the setting times were done according to the ASTM C191[19] as shown in Table 3. It can be seen that when LKD is increased, the water content required increases. It shown that in parallel with the increasing content of LKD, more water was needed to be added. This indicates that in order, to obtain standard consistency cement paste the water/cement (W/C) ratio needs to be increased. LKD absorbs more water compared to OPC probably because LKD particles is porous and bigger in size. This increased W/C ratio accelerates the mixes initial and final setting time and vice versa. The findings is in agreement with previous research on CKD as cement replacement [18], [20]. As such, the LKD mixes initial and final setting times conforms with ASTM C150-81 [21] standards, even at 60% LKD replacements.

784

Masimawati Abdul Latif et al. / Procedia Engineering 125 (2015) 780 – 787

Table 3. The effect of LKD replacement on Consistency and Setting Time. LKD Mix ID

Standard consistency of cement paste Water

Replacement

Setting Time (min)

(%)

Water (ml)

W/C

Initial

Final

LKD0

0

102

0.255

175

235

LKD5%

5

108

0.27

170

225

LKD10%

10

116

0.29

160

220

LKD15%

15

120

0.3

150

215

LKD20%

20

125

0.313

135

200

LKD30%

30

139

0.348

124

185

LKD40%

40

145

0.363

115

176

LKD50%

50

150

0.375

110

157

LKD60%

60

158

0.395

101

140

2.6. Samples Preparation and Test Methods Table 4 shows the LKD mortar mix proportions. A total of 9 mixes were prepared by increasing the LKD replacement percentage. All mixes used constant water binder ratio (W/B) = 0.4 and binder to sand ratio (B: S) = 1:2. The studied OPC replacement was increased gradually from 5% to 60%. Firstly, four groups of silica sands were mixed based on percentage derived from the sieve analysis. After that, OPC was put into the mixture for another 2 minutes of mixing. LKD was added after OPC and mixed for two minutes. Later, water was added and continue mixing for 2 minutes. The mixes were checked for its workability with the flow table. Super plasticizer was added when required then continued mixing for 2 minutes. Lastly, two layers of fresh mortar were filled into 50mm X 50mm X 50mm cubes moulds. The mould was then vibrated on the vibrating table for every layer. The excess mortar was removed. Then, all moulds were covered by gunny sacks to reduce water evaporation. After 24hours, the samples were dismantled and cured in water. At least 3 samples were tested for each curing time and replacement percentage to check for consistency of the test. The targeted mortar spread was fixed at 150mm – 200mm and flow table test has been done according to ASTM C230/C230M-08 [22]. The compressive strength test was tested after 1 day, 3 days, 7 days, 14 days, 28 days and 90 days of curing according to ASTM C109M-13 [23]. Each type of LKD mixes samples were tested with three samples at each testing age and the average results were reported. The LKD mortar 28-day density was determined according to BS 1881-114:1983[24]. 3. Results and discussion Table 3 shows the mixes consistency and setting times with the influence of LKD replacement. The results indicated that, all LKD mixes has faster initial and final setting times when compared with control mix (LKD0). LKD setting times still conforms with ASTM C 150 which initial setting times < 45 minutes and final setting times > 375minutes [25]. Binder with LKD has faster hydration as reported by Maslehuddin (2009) for Cement Kiln Dust (CKD) [26]. The lime content increases with the increase of LKD. Therefore, LKD reacted with water and accelerated its hydration. Similar findings was reported by R. Siddique (2008) that, OPC rheological properties was affected by the CKD. It also stated that water requirement for normal consistency increased with the increase in CKD content [27].Consistency and setting time results are consistent with LKD flow spread as shown in Fig. 3. The flow decreased with the increased of LKD content. Therefore, SP was added for LKD mixes from 30% until LKD60% replacement to achieve the targeted flow diameter of 150mm to 200mm. Fig. 3 also showed that LKD mortar density decreased with increment of LKD replacement. This is because LKD particles is mostly porous. Porous particles will decrease the LKD mortar density as its porosity increased. This also supported by LKD paste accelerated setting time.

785

Masimawati Abdul Latif et al. / Procedia Engineering 125 (2015) 780 – 787

Table 4. Mix proportions and basic properties of LKD Mixes. LKD Replacement (%)

OPC (kg/m3)

LKD (kg/m3)

Water (kg/m3)

Manufactured Silica Sand (kg/m3)

SP (Sika VS2199) (%)

LKD0

0

550

0

220.1

1100

0

LKD5%

5

522.5

27.5

220.1

1100

0

LKD10%

10

495.0

55.0

220.1

1100

0

LKD15%

15

467.5

82.5

220.1

1100

0

LKD20%

20

440.0

110.0

220.1

1100

0

LKD30%

30

385.0

165.0

220.1

1100

0.145

LKD40%

40

330.0

220.0

220.1

1100

0.257

LKD50%

50

275.0

275.0

220.1

1100

0.419

LKD60%

60

220.0

330.0

220.1

1100

0.568

Mix ID

2250

170 Mix Density Flow Spread 165

2150 160 2100 155 2050

Flow Spread (mm)

Density (kg/m3)

2200

150

2000

145

1950 LKD0

LKD5% LKD10% LKD15% LKD20% LKD30% LKD40% LKD50% LKD60% Mix ID Fig. 3. LKD Mixes Density and Flow Spread

Table 5 showed the effect of cement content on the compressive strength. The compressive strength increased with mortar age for all samples. However, compressive strength decreased considerably with the increasing of LKD at every replacement level. The strength reduction is influenced by the reduction in the OPC content being substituted by LKD since C3S and C2S content present in OPC contributes to mortar strength. Generally, higher amount of LKD affects the overall strength performance. Each measurement was taken from average of three specimens. Furthermore, the porous nature of LKD also influence the lower compressive strength. This is in agreement with Najim (2014). His study stated that, the internal stress generated due to volume increased in porous material deteriorate the mortar strength [18]. LKD mortar early age at 3 days and 7 days is within 15.46MPa–44.64MPa and 17.3MPa–53.41MPa respectively. LKD mortar early strength shows similar traits with Type I cement in ASTM C150 with strength of ≥12MPa for 3days and strength ≥19MPa for 7days, similar with CKD mortar strength [26]. In this research, the rate of early strength decreased with increased of LKD content. However, the strength increased with age. Replacement of LKD

786

Masimawati Abdul Latif et al. / Procedia Engineering 125 (2015) 780 – 787

above 20% reduced the compressive strength of the mixes significantly. However, at 28 days, mixes with 5%, 10%, 15% and 20% LKD replacement reduced the strength only about 20% when compared to the control. Many researcher has observed the same trend of strength reductions with pozzolans reaction in paste when OPC is substituted [26,28]. In this research, at age 28 days, almost all LKD mixes obtained strength above 50 MPa. The normalized compressive strength value indicated that LKD can be use as cement replacement up to 50%. Mortar mix with the same B/S ratio and W/B ratio while increasing the LKD replacement ratio generally reduces the compressive strength of LKD mortar. Table 5. Compressive and normalized strength value of blended mortar Compressive Strength (MPa)

Normalized Compressive Strength (MPa)

Mix ID 1

3

7

14

28

90

1

3

7

14

28

90

LKD0

38.33

48.06

54.49

57.96

69.23

74.96

100

100

100

100

100

100

LKD5%

19.92

44.64

53.41

55.46

61.98

67.86

51.97

92.88

98.02

95.68

89.53

90.53

LKD10%

18.55

43.15

52.12

53.86

59.08

66.98

48.40

89.79

95.66

56.29

85.34

89.36

LKD15%

18.31

32.47

51.45

52.78

59.05

63.96

47.76

67.56

94.43

93.77

85.30

85.33

LKD20%

18.10

31.88

49.21

50.31

55.88

61.07

47.23

66.35

90.31

53.66

80.72

81.47

LKD30%

17.69

31.08

34.88

46.73

54.56

57.03

46.16

64.66

64.01

87.09

78.81

76.08

LKD40%

16.84

30.85

32.68

37.84

53.27

56.80

43.93

64.18

59.97

43.45

76.95

75.77

LKD50%

16.81

21.00

24.90

30.28

52.11

54.59

43.86

43.70

45.71

69.69

75.27

72.83

LKD60%

13.87

15.46

17.30

22.38

31.93

36.35

36.17

32.17

31.76

32.11

46.12

48.49

4. Conclusions The conclusions from this experimental study can be drawn are: x x x x x

Addition of LKD reduces the mortar strength. A strength of above 50 MPa for 28days age can be achieved by replacing cement with LKD up to 50%. Increase of LKD increases water demand. Addition of LKD significantly accelerates the initial and final setting times. LKD also reduced the density of the mortar.

This study results are motivations for all researcher to conduct an investigation on LKD as a cement replacement material in producing mortar for sustainable constructions materials. LKD in mortar requires more studies on durability aspects of the LKD mixes mortars. Acknowledgements The authors would like to thank the University of Malaya and the Ministry of Education, Malaysia, for supporting this research work through research grants UM.C/625/1/HIR/ MOHE/ENG/56. Appreciation is also due to all the staff members of the Concrete Laboratory, University of Malaya for assisting with the laboratory experiments. References [1]

G. Matias, P. Faria, and I. Torres, “Lime mortars with heat treated clays and ceramic waste: A review,” Constr. Build. Mater., vol. 73, pp. 125–136, Dec. 2014.

Masimawati Abdul Latif et al. / Procedia Engineering 125 (2015) 780 – 787 [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28]

V. Sata, A. Sathonsaowaphak, and P. Chindaprasirt, “Resistance of lignite bottom ash geopolymer mortar to sulfate and sulfuric acid attack,” Cem. Concr. Compos., vol. 34, no. 5, pp. 700–708, May 2012. U.S. Geological Survey, “U.S. Geological Survey, 2014, Mineral commodity summaries 2014,” 2014. S. Horpibulsuk, A. Cholphatsorn, and A. Chinkulkijniwat, “Strength of Concrete Block Manufactured from Calcium Carbide Residue and Fly Ash,” Maejo Int. J. Sci. Technol., vol. 6, no. 01, pp. 1–9, 2012. A. Hasanbeigi, L. Price, and E. Lin, “Emerging energy-efficiency and CO2 emission-reduction technologies for cement and concrete production: A technical review,” Renew. Sustain. Energy Rev., vol. 16, no. 8, pp. 6220–6238, Oct. 2012. C. Rattanashotinunt, P. Thairit, W. Tangchirapat, and C. Jaturapitakkul, “Use of Calcium Carbide Residue and Bagasse Ash Mixtures as a New Cementitious Material in Concrete,” Mater. Des., vol. 46, pp. 106–111, Apr. 2013. N. Makaratat, C. Jaturapitakkul, and T. Laosamathikul, “Effects of Calcium Carbide Residue – Fly Ash Binder on Mechanical Properties of Concrete,” J. Mater. Civ. Eng., vol. 22, no. November, pp. 1164–1170, 2010. R. Siddique and A. Rajor, “Use of cement kiln dust in cement concrete and its leachate characteristics,” "Resources, Conserv. Recycl., vol. 61, pp. 59–68, 2012. M. A. Issa, “Efficient and Beneficial Use of Industrial By-Products In Concrete Technology,” Geo-Frontiers 2011, pp. 1182–1191, 2011. A. F. Mohamed, “Recycling System in Malaysia: Case studies on Industrial Waste,” 2009. V. L. Lau, “Case Study on the Management of Waste Materials in Malaysia,” Forum Geookol, vol. 15, no. 2, pp. 7–14, 2004. B. M. M. Miller and R. M. Callaghan, “U.S. Geological Survey - Lime Kiln Dust as a Potential Raw Material in Portland Cement Manufacturing,” 2004. American Society for Testing and Materials, “ASTM C33- 03 : Standard Spesification for Concrete Aggregate,” ASTM Stand. B., vol. 04, pp. 1–11, 2001. British Standards Institute, “BS 812: Part 2: 1995 Testing Aggregates,” no. August, 2002. B. Standard, “BS 5075-1: 1982 Concrete admixtures - Part 1: Specification for accelerating and retarding water reducing admixtures,” no. September, 2004. British Standards Institute, “BS812-Part 2. Method for determination of density,” no. 3, 1999. B. S. Institute, “BS12:1996. Specification for Portland Cement,” 1996. K. B. Najim, Z. S. Mahmod, and A.-K. M. Atea, “Experimental investigation on using Cement Kiln Dust (CKD) as a cement replacement material in producing modified cement mortar,” Constr. Build. Mater., vol. 55, pp. 5–12, Mar. 2014. American Society for Testing and Materials, “ASTM C191 : Standard test method for normal consistency by Vicat needle.,” vol. 4, 2004. S. A. El-aleem and H. El Didamony, “Effect of cement kiln dust sustitution on chemical and phhysical properties and compressive strength of portland and slag cements,” Arab. J. Sci. Eng., vol. 30, no. 2, pp. 263–273, 2005. American Society for Testing and Materials, “ASTM C150-81: Standard Specification for Portland Cement,” pp. 153–159, 1981. American Society for Testing and Materials, “ASTM C1437-13: Standard Test Method for Flow of Hydraulic Cement Mortar,” pp. 1–2, 2014. American Society for Testing and Materials, “ASTM C109/C109M -13: Standard Test Method for Compressive Strength of Hydraulic Cement Mortars ( Using 2-in . or [ 50-mm ] Cube Specimens ),” pp. 1–10, 2014. British Standards Institute, “BS1881-114. Method for determination of density of hardened concrete.,” no. 1, 1983. American Society for Testing and Materials, “ASTM C150/C150M -12: Standard Specification for Portland Cement,” pp. 1–9, 2014. M. Maslehuddin, O. S. B. Al-Amoudi, M. K. Rahman, M. R. Ali, and M. S. Barry, “Properties of cement kiln dust concrete,” Constr. Build. Mater., vol. 23, no. 6, pp. 2357–2361, Jun. 2009. R. Siddique, “Chapter 11 - Cement Kiln Dust,” in Waste Materials and By-Products in Concrete., 2008, pp. 351–380. S. Pavía and D. Regan, “Influence of cement kiln dust on the physical properties of calcium lime mortars,” Mater. Struct., vol. 43, pp. 381– 391, 2010.

787