2nd Asian Concrete Federation Conference_Bali_Nov 20-21 2006

2nd ASIAN CONCRETE FEDERATION CONFERENCE – BALI, INDONESIA, NOVEMBER 20-21, 2006

CMT -

DEVELOPMENT OF FLY ASH-BASED GEOPOLYMER CONCRETE: PROGRESS AND RESEARCH NEEDS

Djwantoro Hardjito1 and B.Vijaya Rangan2

ABSTRACT : Research on low-calcium fly ash-based geopolymer concrete, sometimes called ‘concrete with no cement’, has been attracting significant attention. Geopolymer concrete offers a solution for the need of ‘greener’ construction material in the midst of the environmental concern on the production of ordinary Portland cement (OPC). In the last five years, numerous reports on geopolymer concrete have been published. This paper reviews the progress of development, discusses the present status, and explores the research needs.

KEYWORDS: Concrete, Environmental Issue, Fly Ash, Geopolymer, Mechanical Properties.

1. INTRODUCTION

The development of fly ash-based geopolymer concrete is an attempt to answer the challenge to produce more environmentally friendly concrete. The use of by-product material, i.e. fly ash, as a base material for concrete binder to totally replace the use of Portland cement through geopolymerisation process has been attracting a lot of attention globally. This attempt results in twofold benefits, i.e. to provide a solution with regard to the concern on the carbon dioxide emission from Portland cement production, and to provide way to effectively use fly ash. Fly ash, the by-product material from burning coal especially in power stations, is available abundantly worldwide. Its availability is increasing and yet its utilization to date is still very low. Without proper plan, the management of fly ash may incur cost, and potentially harm the natural environment. In the last five years, a significant progress has been made in the development of fly ash-based geopolymer concrete, understanding its properties, and application of geopolymer concrete in reinforced structural members. This paper reviews the progress of development, discusses the present status, and explores the research needs.

2. MATERIALS AND TECHNOLOGY

The geopolymer technology proposed by Davidovits (1988; 1988) shows considerable promise for application in concrete industry as an alternative binder to the Portland cement. In terms of reducing the global warming, the geopolymer technology could reduce the CO2 emission to the atmosphere caused by cement and aggregates industries by about 80% [3]. In this technology, the source material that is rich in silicon (Si) and Aluminium (Al) is reacted with a highly alkaline solution through the process of geopolymerisation to produce the binding material. The term ‘geopolymer’ describes a family of mineral binders that have a polymeric silicon-oxygen-aluminium framework structure, similar to that found in zeolites, but without the crystal structure. The polymerisation process involves a substantially fast chemical reaction under highly alkaline condition on Si-Al minerals, that results in a three-dimensional polymeric chain and ring structure consisting of Si-O-Al-O bonds, as follows (Davidovits 1999):

1 Lecturer, School of Engineering and Science, Curtin University of Technology, Sarawak, Malaysia 2 Emeritus Professor, Faculty of Engineering, Curtin University of Technology, Perth, Australia

1 CMT - 1


CMT -

M n [-(SiO2) z –AlO2] n . wH 2 O (1)

Where: M = the alkaline element or cation such as potassium, sodium or calcium; the symbol – indicates the presence of a bond, n is the degree of polycondensation or polymerisation; z is1,2,3, or higher, up to 32. As source material, the calcined kaolin has been widely investigated. However, from the economic point of view, other Si-Al materials including by-product materials such as fly ash and blast furnace slag have been studied as well. As for the alkaline liquid, the most common alkaline solution used in geopolymerisation is a combination of sodium hydroxide (NaOH) or potassium hydroxide (KOH) and sodium silicate or potassium silicate (Davidovits 1999; Palomo, Grutzeck et al. 1999; Barbosa, MacKenzie et al. 2000; Xu and van Deventer 2000; Swanepoel and Strydom 2002; Xu and van Deventer 2002). The use of a single alkaline liquid has been reported (Palomo, Grutzeck et al. 1999; Teixeira-Pinto, Fernandes et al. 2002), Palomo et al (1999) concluded that the type of alkaline liquid plays an important role in the polymerisation process. Reactions occur at a high rate when the alkaline liquid contains soluble silicate, either sodium or potassium silicate, compared to the use of only alkaline hydroxides. Xu and van Deventer (2000) confirmed that the addition of sodium silicate solution to the sodium hydroxide solution enhanced the reaction between the source material and the solution. Furthermore, after a study of the geopolymerisation of sixteen natural Al-Si minerals, they found that generally the NaOH solution caused a higher extent of dissolution of minerals than the KOH solution. Inspired by the geopolymer technology and the fact that fly ash is a waste material abundantly available, fly ash-based geopolymer concrete has been gaining much attention as a solution for the need of ‘greener’ construction material, and research in this field has been becoming attractive.

3. SHORT TERM PROPERTIES

Of the short-term properties, compressive strength that has been studied widely. This is understandable, as the compressive strength usually gives an overall picture of concrete quality, and used as a salient parameter in design and specified for compliance purposes (Neville 2000). Several factors have been reported to influence the compressive strength of fly ash-based geopolymer concrete, i.e.:

- H20-to-Na2O molar ratio, with smaller ratio tends to result in higher strength (Barbosa, MacKenzie et al. 2000; Hardjito and Rangan 2005)

- Curing time, with longer curing period results in higher compressive strength (Hardjito and Rangan 2005)

- Curing temperature, with increase in the curing temperature increases the compressive strength (Palomo, Grutzeck et al. 1999; Hardjito and Rangan 2005)

Mixing time (Hardjito, Wallah et al. 2004) and rest period (Hardjito, Wallah et al. 2004; Bakharev 2005) were also found to influence the compressive strength of fly ash-based geopolymer concrete. Based on the study of geopolymerisation of sixteen natural Si-Al minerals, Xu and van Deventer (Xu and van Deventer 2000) reported that factors such as the percentage of CaO, K2O, and the molar Si-to-Al ratio in the source material, the type of alkaline liquid, the extent of dissolution of Si, and the molar Si-to-Al ratio in solution significantly influenced the compressive strength of geopolymers. Also, based on a statistical study of the effect of parameters on the polymerisation process of metakaolin-based geopolymers, Barbosa et al (1999; 2000) reported the importance of the molar composition of the oxides present in the mixture and the water content. The other properties of hardened concrete properties, such as the tensile strength, the stress-strain behavior, the Poisson’s ratio and the Young modulus elasticity, have been reported (Hardjito and

1 CMT - 2


CMT -

Rangan 2005). The workability of the fresh concrete; measured in term of slump value; is mostly affected by the water content in the mixture (Hardjito and Rangan 2005).

4. LONG TERM PROPERTIES

It was reported that heat-cured fly ash-based geopolymer concrete undergoes low creep and very little drying shrinkage (Palomo, Fernandez-Jimenez et al. 2004; Gourley and Johnson 2005; Wallah, Hardjito et al. 2005; Wallah and Rangan 2006). The creep factor (i.e. ratio between creep strain and elastic strain) is about 0.44-0.63 after one year for concrete with compressive strength about 60 MPa, and the drying shrinkage strain is about 100x10-6 also after one year (Wallah, Hardjito et al. 2005; Wallah and Rangan 2006). The reason for the low value of drying shrinkage may be due to the presence of very few hydrates or gross capillaries in fly ash-based geopolymer concrete (Gourley and Johnson 2005). There is also suggestion that the corrosion mechanism in fly ash-based geopolymer concrete is very different to those in ordinary Portland cement concrete (Gourley and Johnson 2005). It was reported that fly ash-based geopolymer concrete specimens performed very well after soaking in 5% sodium sulfate solution for one year. The properties investigated were the changes in compressive strength, mass, and length of test specimens. Geopolymer concrete performed far better than OPC concrete when exposed to sulfuric acid solution with concentrations ranging from 0.5% to 10.0% [18, 19].

5. MIX DESIGN

Figure 1. Preliminary Mixture Design Process

A preliminary mixture design process has been proposed (Hardjito and Rangan 2005). It is based on the assumption that the influence of the aggregates is the same as in OPC concrete. The compressive strength of hardened concrete and the workability of fresh concrete are selected as the performance criteria (see Figure 1).

Workability

1. Water Content 2. Super plasticiser

1.H 2 O-to-Na 2 O molar ratio or water-to- geopolymer solids ratio 2. Curing time and curing temperature

Mixture Proportion

Mean Compressive Strength

1 CMT - 3


CMT -

6. REINFORCED STRUCTURAL MEMBERS

Studies on reinforced fly ash-based geopolymer concrete railway sleepers (Palomo, Fernandez-Jimenez et al. 2004; Gourley and Johnson 2005), beams and columns (Sumajouw, Hardjito et al. 2005; Sumajouw, Hardjito et al. 2006; Sumajouw and Rangan 2006) beam-column joints (Brooke, Keyte et al. 2005), wall panels and sewer pipes (Gourley and Johnson 2005) have been reported. Pre-stressed fly ash-based geopolymer concrete railway sleepers have been manufactured in a factory without using any significant change from the conventional process, and adapted regular pre-stressing procedure (Palomo, Fernandez-Jimenez et al. 2004). It has been reported that the geopolymer concrete used in the manufacture of the sleepers experienced very little drying shrinkage (Palomo, Fernandez-Jimenez et al. 2004), very low creep (Gourley and Johnson 2005), and showed excellent bonding with the pre-stressed steel bars showing no slippage at ultimate load (Palomo, Fernandez-Jimenez et al. 2004; Gourley and Johnson 2005). The geopolymer concrete railway sleepers also more than adequately met all the relevant Australian Standards for static and cyclic load tests and showed excellent performance in the trial in mainline tracks (Gourley and Johnson 2005). It has been reported that the structural behavior and strength of reinforced geopolymer concrete beams and columns were similar to those made of Portland cement concrete (Sumajouw, Hardjito et al. 2005; Sumajouw, Hardjito et al. 2006; Sumajouw and Rangan 2006). The seismic behavior of reinforced geopolymer concrete beam-column joints is similar to that of OPC members (Brooke, Keyte et al. 2005). Therefore, the design provisions contained in the current codes and standards for concrete structures are applicable to reinforced geopolymer concrete members. Successful manufacturing of geopolymer concrete sewer pipes using the conventional pipe making processes has been reported; the geopolymer concrete pipes showed excellent performance and met the requirements set by the relevant Australian Standards (Gourley and Johnson 2005). On the resistance to high temperature, geopolymer concrete wall panels (3 meters long, 0.6 wide, and 60 mm thickness) have been set up to fire test, and showed excellent performance (Gourley and Johnson 2005).

7. RESEARCH NEEDS

In most reports, fly ash-based geopolymer concrete was cured at elevated temperature. Although steam- curing is common in pre-cast concrete industry, it will be beneficial if fly ash-based geopolymer concrete can be manufactured in ambient temperature. This will enhance the usage of geopolymer concrete in industrial applications. Research in the mechanism of geopolymerisation, especially with fly ash as source material, is needed. It is suggested that the use of potassium silicate solution as reactor may enable fly ash-based geopolymer concrete undergoes curing in room temperature. More data are needed on the long-term properties of geopolymer concrete.

8. CONCLUDING REMARKS

This paper presented a summary on the progress and current status on the development of geopolymer concrete using low-calcium fly ash as a source material. Significant information has been gathered and reported. The excellent potential of geopolymer concrete as a construction material has been exposed. This paper also presented several areas where research is needed to further develop this material, and to make it widely usable in many applications.

1 CMT - 4


CMT -

9. REFERENCES

Bakharev, T. (2005). "Geopolymeric Materials Prepared Using Class F Fly Ash and Elevated Temperature Curing." Cement and Concrete Research 35(6): 1224-1232.

Barbosa, V. F. F., K. J. D. MacKenzie, et al. (1999). "Synthesis and Characterisation of Sodium Polysialate Inorganic Polymer Based on Alumina and Silica." Geopolymer '99 International Conference, France.

Barbosa, V. F. F., K. J. D. MacKenzie, et al. (2000). "Synthesis and Characterisation of Materials Based on Inorganic Polymers of Alumina and Silica: Sodium Polysialate Polymers." International Journal of Inorganic Materials 2(4): 309-317.

Brooke, N. J., L. M. Keyte, et al. (2005). "Seismic Performance of 'Green Concrete' Interior Beam-Column Joints." Australian Structural Engineering Conference, Newcastle, Australia.

Davidovits, J. (1988). "Geopolymer Chemistry and Properties." Geopolymer '88, First European Conference on Soft Mineralurgy, Compiegne, France, The Geopolymer Institute.

Davidovits, J. (1988). "Soft Mineralurgy and Geopolymers." Geopolymer '88, First European Conference on Soft Mineralurgy, Compiegne, France, The Geopolymer Institute.

Davidovits, J. (1999). "Chemistry of Geopolymeric Systems, Terminology." Geopolymer '99 International Conference, France.

Gourley, J. T. and G. B. Johnson (2005). "Development in Geopolymer Precast Concrete. Geopolymer, Green Chemistry and Sustainable Development Solutions." J. Davidovits. Saint Quentin, France, Geopolymer Institute: 139-143.

Hardjito, D. and B. V. Rangan (2005). "Development and Properties of Low Calcium Fly Ash-Based Geopolymer Concrete." Research Report GC-1, Perth, Australia, Faculty of Engineering, Curtin University of Technology: 94.

Hardjito, D., S. E. Wallah, et al. (2004). "On The Development of Fly Ash-Based Geopolymer Concrete." ACI Materials Journal 101(6): 467-472.

Neville, A. M. (2000). "Properties of Concrete." Prentice Hall. Palomo, A., A. Fernandez-Jimenez, et al. (2004). "Precast Elements Made of Alkali-Activated Fly Ash

Concrete." Eighth CANMET/ACI International Conference on Fly Ash, Silica Fume, Slag, and Natural Pozzolans in Concrete, Las Vegas, USA.

Palomo, A., M. W. Grutzeck, et al. (1999). "Alkali-Activated Fly Ashes, A Cement for the Future." Cement and Concrete Research 29(8): 1323-1329.

Sumajouw, D. M. J., D. Hardjito, et al. (2005). "Behaviour and Strength of Reinforced Fly Ash-Based Geopolymer Concrete Beams." Australian Structural Engineering Conference 2005 "Structural Engineering - Preserving and Building into the Future", Newcastle, NSW, Australia.

Sumajouw, D. M. J., D. Hardjito, et al. (2006). "Fly Ash-Based Geopolymer Concrete: Study of Slender Reinforced Columns." Journal of Material Science Accepted for publication.

Sumajouw, D. M. J. and B. V. Rangan (2006). "Low-Calcium Fly Ash-Based Geopolymer Concrete: Reinforced Beams and Columns." Research Report GC-3, Perth, Australia, Faculty of Engineering, Curtin University of Technology: 120.

Swanepoel, J. C. and C. A. Strydom (2002). "Utilisation of fly ash in a geopolymeric material." Applied Geochemistry 17(8): 1143-1148.

Teixeira-Pinto, A., P. Fernandes, et al. (2002). "Geopolymer Manufacture and Application - Main problems When Using Concrete Technology." Geopolymers 2002 International Conference, Melbourne, Australia, Siloxo Pty. Ltd.

Wallah, S. E., D. Hardjito, et al. (2005). "Creep and Drying Shrinkage Behaviour of Fly Ash-Based Geopolymer Concrete." Concrete '05, Melbourne, Australia, Concrete Institute of Australia.

Wallah, S. E. and B. V. Rangan (2006). "Low-Calcium Fly Ash-Based Geopolymer Concrete: Long-Term Properties." Research Report GC-2, Perth, Australia, Faculty of Engineering, Curtin University of Technology: 97.

Xu, H. and J. S. J. van Deventer (2000). "The Geopolymerisation of Alumino-Silicate Minerals." International Journal of Mineral Processing 59(3): 247-266.

Xu, H. and J. S. J. van Deventer (2002). "Geopolymerisation of Multiple Minerals." Minerals Engineering 15(12): 1131-1139.

1 CMT - 5


CMT -

BOTTOM ASH RECYCLED MATERIAL USED CEMENTITIOUS FIRE PROTECTION COATING APPLIED CONCRETE TUNNEL LINING BEHAVIOR

UNDER RABT FIRE LOADING

Jang-Ho Jay Kim1, Hae-Geun Park2, Myeong-Sub Lee3, Jong-Pil Won4, Yun-Mook Lim5

ABSTRACT: Most tunnels exposed to prolonged fires are heavily damaged or collapsed. So important point to consider is not only minimizing human fatalities but also minimizing economical loss from tunnel fire. When compared to repairing of fire damaged tunnel lining, the application of fire protection coating is relatively simple and less expensive in construction as well as effective in protecting tunnels from collapsing due to fire. However, the tunnel coating materials available in the market are still expensive in cost and have low strength. In order to enhance the current coating materials’ weaknesses, the newly developed coating material is low in cost and higher in strength by using bottom ash from coal generated electric power plant with special mixture design. For the experiments, the general NATM tunnel concrete lining with the coating material is fire tested using fire loading curve of RABT. The developed fire protection coating material is applied on the concrete tunnel lining surface. Also, thermo-couples are embedded in the specimens to measure temperatures during fire loading. This paper describes the results of fire tests of the newly developed cementitious fire protection coating material applied concrete tunnel lining specimens for the application in tunnels. From this test result of RC tunnel lining. The newly developed fire protection coating material enhances fire protection ability. KEYWORDS: Fire protection, Newly developed coating material

1. INTRODUCTION

In recent years, many tunnel users and fire fighters have lost their lives in a number of tunnel fires. So tunnels are to meet very stringent fire safety requirements. Underground spaces, like tunnels, have confined conditions, so tunnel fires can result in severe human casualties and structural damages. Therefore, tunnels are enforced with very stringent fire safety requirements and preventive measures, which are gaining great importance in tunnel design. In recent years, in order to prevent these problems, clients are requesting that tunnel linings should be fire-resistant and new and old tunnels are applied with fire protection coating on the surface of concrete tunnel lining.

2. ADVANTAGES OF NEWLY DEVELOPED FIRE PROTECTION MATERIAL

Newly developed cementitious fire protection coating material is produced to meet safety requirement of tunnel structure for the fire damage and to facilitate casting and coating on construction site. Especially, this fire protection coating material is developed focusing on improving strength to resist spalling or exfoliation due to severe vibration induced by train and traffic or by wind pressure. This new coating material is primarily composed of Type Ι ordinary Portland cement, PP (polypropylene) fibers and bottom ash. Bottom ash from coal generated electric power plant is used as light weight fire proof aggregate over shell sand, which is used broadly in Europe and Japan. Bottom ash is 20% of burnt coal ash, which remains at the bottom of coal burner and has porous micro structure. Because of porous micro structure, bottom ash has superior heat insulating characteristic. Also, utilization of bottom ash is environmentally beneficial since it is a waste material, which needs to be disposed. Because density of bottom ash is higher than shell sand, it enhances the strength of coating material. Accelerated setting agent is used to minimize the reduction of bond strength. In order 1 Associate Professor, Department of Civil & Environmental Engineering, Sejong University, Korea. 2 Civil Division, Samsung Engineering & Construction, Korea. 3 Civil Division, Samsung Engineering & Construction, Korea. 4 Professor, Department of Civil & Environmental System Engineering, Konkuk University, Korea. 5 Professor, Department of Civil & Environmental Engineering, Yonsei University, Korea

2 CMT - 6


CMT -

to prevent catastrophic spalling, fiber length and thicknesses of 18 mm and 2.1 denier, respectively, are used to form effective passageway to release steam. This fireproof material is pre-mixing type where shotcreting and on site casting are possible. Material mixture design of cement to aggregate ratio of 1 to 1.5 and the fiber volume ratio of 0.25% are used. Also, acceleration setting agent of 1 volume percent is used to improve bond strength of the material for shotcreting. The water to coating material ratio of 0.395 is used for the mixture design. Since the material emphasizes strength as the main material property enhancement, compressive, flexural, and bond strength experiments are performed. The test datum are compared with the properties of the popular European fire proof coating material for shotcreting and on site casting. Figures 1, 2 and 3 show experiment results. As shown in Figure 1, 28 day compressive strengths of the newly developed material and the commercially available material are 20.89 MPa and 7.44 MPa, respectively. Also, as shown in Figure 2, the 28 day flexural strengths of the newly developed material and the commercially available material are 4.30MPa and 2.22MPa, respectively. Finally, as shown in Figure 3, the bond strength of the newly developed material the commercially available material are 1.83MPa and 1.34MPa, respectively. The strength test results show that the newly developed material has far greater strength capacities than the material available in the market.

0

5

10

15

20

25

Com

pre

ssiv

e S

trength

(MPa)

commercially available material

newly developed material

Figure 1. Comparison of compressive strength

of commercially available material and newly developed material

0

1

2

3

4

5Fle

xura

l Strength

(MPa)



Figure 2. Comparison of flexural strength of

commercially available material and newly developed material

0

0.5

1

1.5

2

Bond S

trength

(MPa)



Figure 3. Comparison of bond strength of

commercially available material and newly developed material

3. TESTING PROCEDURE

3.1. Production of specimens Concrete lining specimens are designed based on a general NATM tunnel concrete lining. The mix design of concrete of Korean Highway Corporation is used to cast original RC tunnel lining. Table 1 shows mixture design contents of concrete of RC tunnel lining. The expected 28 day compressive strength of concrete is 24MPa.

2 CMT - 7


CMT -

Table 1. Concrete tunnel lining mix design

Unit Weight (kg/m3)

Gmax (mm)

Slump (cm)

W/C (%)

Air (%) Water Cement

Fine aggregate

(S)

Coarse aggregate

(G)

high-range water-

reducing agent (%)

AE agent (%)

25 14 0.5 4.5 167 334 728.33 1025.13 0.2 0.03

Specimen panel size is 1400mm×1000mm×400mm. D16 and D13 steel bars are used as main reinforcement and hoop reinforcement, respectively. K-type sheathed thermocouples were embedded at specified locations in the tunnel lining specimen to obtain the temperature data. Five thermo couples for fire protection coated specimens and four thermo-couples for uncoated concrete tunnel lining specimen are placed at specific locations of specimen to measure temperatures. Thermo-couple locations are ①interface between concrete and fire protection coating, ②mid-depth of concrete cover thickness, ③surface of bottom steel reinforcing bar, ④mid-depth of specimen, and ⑤mid-depth of back concrete lining. Figures 4 and 5 show schematic drawings of reinforcement arrangement of the specimen and sheathed thermocouple locations, respectively.

Figure 4. Schematic drawing of reinforcement arranged in the specimen

2 CMT - 8


CMT -

Figure 5. Schematic drawing of sheathed thermocouples locations

3.2. Coating of fire protection material The newly developed fire protection coating material is applied on the concrete tunnel lining surface. Coating thicknesses of 20mm, 30mm and 40mm were selected to verify the fire protection performance according to the thickness of fire protection coating. First, 1400mm×1000mm size of steel wire mesh was attached on the surface to enhance interface bonding. And after 28 days, dry cured specimens were coated with fire protection material by hand. Single specimen was produced for each thickness and single non-coated specimen was tested as a control specimen. Figures 6 is the photo of applying newly developed fire protection material on the specimen.

Figure 6. Coating the fire protection material by hand

2 CMT - 9


CMT -

3.3. Fire loading curve and fire test Fire test was carried out in KICT (Korea Institute of Construction Technology) after 28 days of dry curing of fire protection material. Fire was applied to the coated face of the specimen using LPG furnace. RABT fire curve was used as control temperature and temperatures obtained from thermo couples were recorded. RABT is fire load regulation for road tunnel in Germany. Figure 7 shows RABT time-temperature curve. Figure 8 is a photo of specimen setup on furnace.

0

200

400

600

800

1000

1200

1400

0 20 40 60 80 100 120 140 160 180

Time(min)

Tem

pera

ture

(C)

Figure 7. RABT time - temperature curve

Figure 8. Specimen set on the furnace

4．FIRE TEST RESULTS

4.1 Concrete lining specimen without fire protection coating (control specimen) Figures 9 is photo of temperature versus time fire test result of concrete lining specimen without fire protection coating. The furnace temperature-time curve is shown for comparison purpose. As shown in Figure 9, when the furnace temperature reaches 1200℃, temperature at mid depth of concrete cover abruptly increased due to the cover thickness spalling caused by high temperature. About 20 minutes after the start of the test, concrete surrounding the reinforcements also spalled off and temperature of the bottom reinforcement increased rapidly. During the test, water and steam were continuously released from the cracks of the side surfaces of specimen. The cracks were propagated toward the

2 CMT - 10


CMT -

upper part of the specimen at the side surfaces. The highest temperature measured at mid-depth of concrete cover was 1197℃, at surface of bottom reinforcing bar was 1075℃, at the mid-depth of specimen was 111℃, and at the mid-depth of back concrete lining was 33℃. Figure 10 shows surface of concrete lining cover after the test. Concrete lining cover was severely damaged by spalling and main reinforcement bar was exposed.

0

200

400

600

800

1000

1200

1400

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180

time(min)

tem

pera

ture

(℃)

mid-depth of concrete cover(37.5mm)

surface of bottom reinforcing bar(75mm)

mid-depth of specimen(207.5mm)

mid-depth of back concrete lining(370mm)

furnace

Figure 9. Test result of concrete lining specimen without fire protection coating

Figure 10. Bottom surface of the specimen after test

4.2 Concrete lining specimen with thickness of 20mm newly developed fire protection coating Figures 11 is obtained temperature-time test results of concrete lining specimen with newly developed fire protection coating material of 20mm. In the test of concrete lining specimen with coating thickness of 20mm, the fire protection coating exploded abruptly by spalling after 50 minutes from the start of the test. Most of the fire protection coating spalled off and test was stopped for a safety reason. When the test was stopped, temperatures at all thermo-couple locations were continuously increasing. The highest temperature at interface, mid-depth of concrete cover, surface of bottom reinforcing bar, mid-depth of specimen and mid-depth of back concrete lining was 839℃, 151℃, 85℃, 11℃ and 12℃, respectively. Figure 12 is a photo of bottom surface of the specimen after the test. The figure shows that concrete lining cover and fire protection coating was severely damaged by spalling. And steel wire mesh for strengthening the interface was separated from the specimen.

2 CMT - 11


CMT -

0

200

400

600

800

1000

1200

1400

0 5 10 15 20 25 30 35 40 45 50 55 60

time(min)

tem

pera

ture

(℃)

interface





furnace

Figure 11. Test result of concrete lining specimen with thickness of 20mm

newly developed fire protection coating

Figure 12. Separated fire protection coating

4.3 Concrete lining specimen with thickness of 30mm newly developed fire protection coating Figures 13 is obtained temperature-time test results of concrete lining specimen with newly developed fire protection coating material of 30mm. The behavior of concrete lining specimen with thickness of 30mm fire protection coating, which was different than that of the specimen with thickness of 20mm fire protection coating, was stable during the test. 20 minutes later after the start of the test, temperature at interface between concrete lining cover and fire protection coating increased significantly. Significant amount of steam is released from the gaps between thermo-couples and specimen during the test, but the amount of steam is less than that observed from the specimen with thickness of 20mm coating. There were no cracks found on the specimen. The highest temperature measured at interface were 473℃, at mid-depth of concrete cover was 163℃, at surface of lower reinforcing bar was 129℃, at the mid-depth of specimen was 40℃, and at the mid-depth of back concrete lining was 16℃. Figure 14 is of a photo of the fire protection coating after the test and no significant damages except slight color change were found at fire applied surface.

2 CMT - 12


CMT -

0

200

400

600

800

1000

1200

1400

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190

time(min)

tem

pera

ture

(℃)

interface





furnace



Figure 14. Surface of fire protection coating after test

4.4 Concrete lining specimen with thickness of 40mm newly developed fire protection coating Figures 15 is obtained temperature-time test results of concrete lining specimen with newly developed fire protection coating material of 40mm. Figures 15 shows very similar behavior to that of the specimen with thickness of 30mm fire protection coating. However, the temperatures were lower than that of 30mm coated specimen. Also, lesser amount of steam release was observed from the specimen and fire protection coating. The highest temperature measured at interface were 252℃, at mid-depth of concrete cover was 117℃, at surface of lower reinforcing bar was 93℃, at the mid-depth of specimen was 24℃, and at the mid-depth of back concrete lining was 12℃. Figure 16 shows a photo of the surface of fire protection coating after the test. There were no significant damages at surface of fire protection coating.

2 CMT - 13


CMT -

0

200

400

600

800

1000

1200

1400

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170

time(min)

tem

pera

ture

(℃)

interface





furnace



Figure 16. Surface of fire protection coating after test

5．CONCLUSION

This fire tests proved that the newly developed fire protection coating material has a wonderful fire protection ability. Also, the test showed that there exists an optimal thickness for fire protection from RABT fire loading. The addition of bottom ash, which has porous micro structure, improves fire protection ability. Moreover, the usage of industrial wastes from coal generated electric power plant such as bottom ash reduces the cost.

6．ACKNOWLEDGEMENT

This research was supported by Civil Division of Samsung Engineering & Construction Company research program, “Fire Protection of RC Tunnel Lining System Development”

7. REFERENCES

Tajima H., and Kishida M., Kanda T., Morita T. (2005). “Study on the deformation and load bearing capacity of TBM shield tunnel lining in fire”, Underground Space Use: Analysis of the Past and

2 CMT - 14


CMT -

Lessons for the Future, Proceedings of the International World Tunnel Congress and the 31st ITA General Assembly, pp. 793-799

Mai D. (2002). “Cement-based mineral-containing passive fire protection for underground structures.” Fourth International Symposium on Sprayed Concrete-2002

2 CMT - 15


CMT -

MECHANICAL PROPERTY OF STEEL-MAKING SLAG CONCRETE WITH JAPANESE SPECIFIED STRENGTH OF AN AIRPORT CONCRETE PAVEMENT

Ryosuke Takahashi1 and Hidenori Hamada2

ABSTRACT : A steel-making concrete slag is a recycled concrete materal consisted with co-productions. This material is expected to use as airport pavement because it has a good abrasion registance. However, there is a possibility of difference characteristics from normal concrete caused by using of the steel-making slag aggregate and the basic characteristics is not clearified in 5N/mm2 flexural strength range that is Japanese specificated strength of airport pavement conrete. Therefore, some tests and consideration are carried out in this study to clearify a basic mechanical properties.

KEYWORDS: steel-making slag concrete, compressive strength, flexural strength, young's modulus, fatigue, shrinkage, thermal expansion, water content.

1. INTRODUCTION Steel-making slag concrete is almost consisted with co-productions in the making of steel. The concrete use blast furnace slag powder and steel-making slag as a binder and an aggregate. An effective use of this material is expected because the use contributes natural environment by saving CO2 generation in cement production and natural resources. We expect to apply this to an airport pavement material because this material shows a good performance against an abrasion in a previous study. Heavy load by an air plane requires higher strength to the pavement. The Japanese standard strength of the airport pavement is over 5 N/mm2 flexural strength at 28th day (CDIT, 1999). However, basic characteristics of this concrete in such strength range are not clarified though a technical manual on this concrete (SCOPE, 1999) was published where characteristics in lower stress range were already clarified. The clarified characteristics shows almost similar mechanical properties to normal concrete. However, There is possiblity that the characters in lower strength doesn't agree with the charactreristics in all strength range because the characteristics of steel-making slag aggregate used is much different from a natural aggregate in normal concrete, for example, its structure, strength and boundary face property. Therefore, experiments and consideration are carried out in this study to clarify the basic mechanical property of steel-making slag concrete in 5N/mm2 flexural strength range.

2. OUTLINE OF EXPERIMENTS

2.1 Materials and Mixture Proportion Table 1 shows material properties of steel-making slag aggregate used in this study. A dephosphorylation steel-making slag co-produced in a hot metal pretreatment is used. The technical manual proposes specifications on the steel-making slag aggregate to prevent the concrete from the self expansion. All qualities in the table are satisfied with the specification. A blast-furnace slag used as binder is Type-4000 in JIS A 6206 and its density is 2.88 g/cm3. A fly ash used as an admixture is unprocessed one and equivalent to the second-class in Japanese Industrial Specification, JIS A-6201. A calcium hydroxide and normal portland cement are used as an alkali activator. The calcium hydroxide is special class in JIS R-9001 and its density is .2.23 g/cm3. The density of the cement is 3.16 g/cm3.

1 Researcher, Port and Airport Research Institute, Japan. 2 Associate professor, Kyushu University, Japan.

3 CMT - 16


CMT -

Table 2 shows mixture proportions. Mixture SNP and SCH are applied to compressive, flexure, shrinkage and thermal expansion test specimens. Mixture FNP is applied to fatigue test specimens. An alkali activator used in steel-making slag concrete is a calcium hydroxide in its principle. However practical use of this causes problems such as saving a special silo for this in a concrete plant. Therefore, use of the cement activator is regarded as basic here and difference in the mechanical property between these two activators is investigated with Mixture SNP and SCH.

2.2 Test method 2.2.1 Tests on Basic Mechanical Characteristics An axial compression test, flexural test, shrinkage test, thermal expansion test and flexural fatigue test on steel-making slag concrete are carried out in this study. φ 100 x 200 mm cylinder specimen is used in the compression test and thermal expansion test. 100 x 100 x 400 mm prism specimen is used in the other three tests. The compression and the flexural test method comply with JIS-A-1108 and JIS A-1106. The loading way is 2 pointed and a/d is 1.0 in flexural test. Change of distance between two tips is measured on a surface of the specimen with a contact-gauge in the shrinkage test and the thermal expansion test. This measurement method is based on JIS A-1129-2.Condition at shrinkage is 20 °C air temperature and 60% relative humidity after 7days underwater curing. The thermal expansion test specimens are submerged into 0, 10, 20, 30, 40 °C water after 28 days underwater curing. They are taken from water after enough time passage from start of submerging and the expansion is measured just after taking. The loading condition in the fatigue test is almost same as the flexural test. The long-term strength increment of the steel-making slag concrete is greater than normal concrete. Therefore, the specimens after over 91 days curing are used to except the influence of the increment. The specimens are coated with grease and lapped with polyvinylidene chloride film as soon as they are taken from water because the water content influences the fatigue strength. A waveform of loading is haversine and its cycle is

Table 2. Mixture ProportionUnit weight（kg/m3） Admixture（g/m3）

Alkali activator C

Desig-nation

Powder - water ratio

WFACBP )35.0( ++

Water

W

Blast furnace

slag powder

BP

Cement

NP

Calcium hydroxid

e

CH

Fri ash

FA

Steel- making

slag aggregate

SS

High-range water

reducing admixture

AE

SNP-1 2.00 244 300 139 0 141 1791 0 0 SNP-2 2.21 194 325 75 0 85 2053 3863 36 SNP-3 2.03 207 325 75 0 59 2048 2757 46 SNP-4 1.96 202 300 50 0 130 2013 3357 48 SCH-1 1.96 202 300 0 50 130 1996 5282 24 FNP 2.57 173 325 75 0 125 1916 2100 30

Table 1. Property of Steel Making Slag Aggregate Used

Oven-dry density

Saturated surface-dry

density Absorption Chemical composition

(% by mass) Percentage of solid

volume (%)

(g/cm3) (g/cm3) (%) MgO S

Fineness modules

-5 mm 5-20 mm

3.46 3.32 4.02 1.8 60.0 4.47 60.0 51.2

3 CMT - 17


CMT -

1Hz. Fluorine resin sheets are inserted between the specimen and a loading jig, the specimen and supports to except a lateral restriction. 2.2.2 Tests on Influence of Water Content The flexural strength of normal concrete is decreased by drying because tensile strain is caused by shrinkage distribution near the surface. (Shigenori, 1995) Therefore, compression and flexural strength tests under some drying conditions are also carried out to investigate its influence. Specimens with Mixture FNP after over 400 days underwater curing are used. Three conditions, saturated, air-dry and bone-dry are prepared (Table 3). Saturated condition specimens are taken from water just before the test. Bone-dry condition specimens are dried by oven at 110 °C till their weight change reaches 0. Air-dry condition specimens are dried in 20 °C and 60 % RH chamber till their weight change close to 0.1g/hour. A dryness degree of Air-dry condition in the table is defined as the ratio between weight loss in Air-dry condition and measured average water content of Bone-dry specimens.

3. TEST RESULTS AND CONSIDERATION

3.1 Compressive Strength Figure 1 shows relationship between material age and compressive strength. No influence of a kind of alkali activator and quantity of cement content are seen among test results. Figure 2 shows compressive strength normalized by the strength in the saturated condition at each dryness degree. The compressive strength doesn’t change by small increase of dryness degree. The strength decreases in bone-dry condition relatively. The decrease is about 20 %.

Table 3. Test Cases on Investigation of Water-content Influence

Designation Mixture Condition Shape Specimen amount

Averaged dryness degree (%)

Cylinder 3 CS Saturated

Prism 3 0.0

Cylinder 3 CB Bone-dry

Prism 2 100.0

Cylinder 3 7.8 CA

FNP

Air-dry Prism 3 8.8

0 50 100 15020

30

40

50

60 : SNP-1　 : SNP-4 : SNP-2　 : SCH-1 : SNP-3　 : FNP-1

Material age (day ）

Com

pres

sive

stre

ngth

(N/m

m2 )

Figure 1. Relationship between compressive strength and material age

0 50 1000.6

0.8

1

1.2

1.4

Dryness degree (%)

Nor

mal

ized

com

pres

sive

stre

ngth

Figure 2. Normalized compressive strength at

each dryness degree

3 CMT - 18


CMT -

3.2 Flexural Strength Figure 3 shows a relationship between flexural strength and material age. In comparison with a kind of alkali activator, it is shown in the figure that the cement is superior to calcium hydroxide in the growth of the flexural strength. There is a possibility that a kind of alkali activator influences growth of tensile strength in some way. However, no relation ship between quantity of cement content and growth of bending strength is seen. Detailed study on this is required to clarify the influence. Figure 4 shows the relationship between the flexural strength and the compressive strength. Test results in previous studies are also shown in the figure. The relationship of normal concrete calculated by the following equation proposed in JSCE standard specifications for concrete structure (JSCE, 1996) is also shown as a line in the figure.

3/2'43.0 cbd ff = (1) Where, fbd is flexural strength (N/mm2) and fc' is an axial compressive strength (N/mm2). A following equation is obtained from test results in this study with a regression analysis.

78.0'32.0 cbd ff = ( 89.0=γ ) (2) Where, γ is correlation coefficient. This equation shows the steel-making slag concrete has higher flexural strength than the strength of normal concrete at the same compressive strength. Surely an intension of higher flexural strength is seen around 5 N/mm2 flexural strength regions in the results. A photograph of failure surface seen in one of flexural test specimen is shown in Figure 5. No failure at boundary surface of aggregate is seen on the surface. Moreover, porous aggregates are seen also on

0 50 100 1503

4

5

6

7

8

Material age （day）

Flex

ural

stre

ngth

(N/m

m2 )

: SNP-1　 : SNP-4 : SNP-2　 : SCH-1 : SNP-3　 : FNP-1

Figure 3. Relationship between flexural strength and material age

10 20 30 40 50 60

2

4

6

8

0

Compressive strength (N/mm2)

Flex

ural

stre

ngth

(N/m

m2 )

P revious studies: Mix. H : Mix. I: Mix. J

This study: SNP-1: SNP-2: SNP-3: CS: CB

: Mix. K: Mix. L: Mix. M

: SNP-4: SCH-1: FNP: CA

: Experimental equation of normal concrete

: Assumed relationship under air-dry condition

Figure 4. Relationship between flexural strength and compressive strength

3 CMT - 19


CMT -

the surface. These facts mean that the boundary surface is strong and aggregate strength is weak. This kind of failure surface is not seen in normal concrete. This difference might influence to the difference of flexural strength between steel-making slag and normal concrete. Test results under several dry conditions are also shown in Figure 4. The flexural strength under air-dry condition and bone-dry condition is 80 % and 55 % of the strength under saturated condition. It is considered that the generation of tensile stress by dry shrinkage distribution decreases the flexural strength same as the normal concrete case. The compressive strength also decreases under the bone-dry condition. The boundary surface failure of the slag aggregate is seen in Figure 6 showing cracked surface of CA specimen. In addition to the shrinkage distribution, there is a possibility that the shrinkage of the aggregate is difference in the binder and the difference influences the failure. It is the subject for future study that the influence of this change of the failure to a flexural property under the actual condition. Almost specimens used in flexural tests of this study are considered to be under saturated condition because the tests are carried out within several hours after taken from water. Therefore, if it is assumed that the actual condition of all specimens in this study is same as the air-dry condition in this study, the relationship becomes a following equation.

3/2'40.0 cbd ff = ( 83.0=γ ) (3) This relationship is also shown in Figure 4 and it is clear that this becomes very close to normal concrete. There is a possibility that the condition in actual structure is very different in the assumed condition in this study. Therefore, detailed investigations with more kinds of the condition considering actual dry condition in the structure are required to clarify more aciculate flexural failure property of steel-making slag concrete.

3.3 Young's modulus Figure 7 shows a relationship between young's modulus and compressive strength. Test data in previous studies and recommended values in a case of normal concrete in JSCE standard (JSCE, 2003) are also shown in the figure. The steel-making slag concrete shows a strong correlation like normal concrete. However, its relationship is different from normal concrete. Previous study (SCOPE, 2003) showed that Young's modulus of the steel-making slag concrete is smaller than the modulus of normal concrete at the same compressive strength. Test results in this study, whose strength range is greater than the previous one, shows higher young's modulus than the normal concrete. However, we cannot conclude this intension as the characteristic of steel-making slag concrete because mixture-I in previous study, a part of whose strength range overlaps with the range of this study, shows another intension. It is said generally that steel-making slag aggregate quality has variety among producing ironworks. Therefore, this difference is supposed to be caused by its variety. Influence of dry condition is not seen between saturated and air-dry condition. However, the young's modulus under bond-dry condition shows much lower values than them. Its decrement rate is greater

10 20 30 40 50 60

10

20

30

40

0

Compressive strength (N/mm 2)

You

ng's

mod

ulus

(kN

/mm

2 )

P revious studies: Mix. H : Mix. I: Mix. J

: Mix. K: Mix. L: Mix. M

: Recommend values by JSCE standard

This study: SNP-1: SNP-2: SNP-3: CS: CB

: SNP-4: SCH-1: FNP: CA

Figure 7. Relationship between Young's modulus and compressive strength

3 CMT - 20


CMT -

than the rate of compressive strength. Considering intension of flexural test results, this decrement is caused by inner fracture in drying such as cracks at boundary surface of the aggregate.

3.4 Shrinkage Shrinkage strain of steel-making slag concrete from start of drying is shown in Figure 8. The strain at 100th day is less than 350 μ in all results. Though this value is slightly lower than general value of normal concrete, these results show almost same characteristic as normal concrete. In comparison among steel-making slag concrete, SNP-1 shows a different intension. Its quantity of shrinkage at 100th drying day and increasing rate after 20th day are very higher than other cases. Characteristic differences between SNP-1 and other cases are a unit quantity of water and cement. Generally the unit quantity of water is considered as one of main influencing factors to shrinkage. A unit quantity in SNP-1 (244 kg/m3) is relatively greater than the quantity in other cases (about 200kg/m3). Shrinkage strain equation of normal concrete proposed in JSCE standard (JSCE, 2003) considers its influence in a term of the ultimate shrinkage strain. If the equation is applied to steel-making slag concrete, the difference of ultimate strain between SNP-1 and other becomes tens of strains. Therefore, it is considered that an influencing way of water content to the shrinkage in steel-making slag concrete is different from normal concrete’s way or that a quantity of cement influences to its shrinkage.

3.5 Thermal expansion Figure 9 shows amount of length change from the length at 20 °C at each temperature. Table 4 shows thermal expansion coefficients gotten by applying a least-square method to the results. A range of the coefficient of normal concrete is generally considered from 7 to 13 x 10-6 / °C (JSEC, 2003).The average coefficient of the results (7.8 x 10-6 / °C.) is within the range of concrete and The variation range also overlaps the lower-side range of the normal concrete. The literature (Neville, 1979) said a thermal expansion coefficient of the blast furnace slag aggregate is lower than concrete and the value is about 8.0×10-6 / ℃. This value agrees with the average value in this study. Therefore, if it is assumed that the coefficient value doesn’t change greatly between binding matrix and aggregate of blast furnace slag, the coefficient of steel-making slag aggregate is

Table 4. Thermal coefficientMixture proportion SNP-1 SNP-2 SNP-3 SNP-4 SCH-1

Thermal coefficient (×10-6/degrees C) 8.0 6.6 7.5 9.6 7.1

-20 0 20 40 60

-200

-100

0

100

200

Temperture of test specimen (degrees C)

Stra

in c

hang

e m

akin

g 20

deg

. C o

rigin

(μ)

: SNP-1 : SNP-4 : SNP-2 : SCH-1 : SNP-3

Figure 9. Thermal expansion

5 10 50 1000

100

200

300

400

Days from start of drying (day)

Srin

kage

stra

in (μ

)

: SNP-1: SNP-2

: SNP-3: SNP-4

: SCH-1

Figure 8. Shrinkage strain

3 CMT - 21


CMT -

close to 8.0×10-6 / ℃ because main components of steel-making slag are blast furnace slag and steel-making slag.

3.6 Flexural fatigue strength In flexural fatigue test in this study, only maximum stress ratio S, which is a ratio of the maximum flexural stress to the flexural strength (= σmax / fbd), is considered as a parameter. Minimum stress ratio is not considered and that is 0 N/mm2. In the calculation of the maximum stress ratio, the flexural strength at start of loading is used. All test specimens are under saturated condition and have over 5 N/mm2 static flexural strengths. Table 5 shows test results. The relationship between S and fatigue life N is shown in Figure 10. A line in the figure shows S-N relationship calculated from a fatigue strength equation of normal concrete proposed in JSCE standard (JSCE, 2003). The equation is expressed as follows, when minimum stress ratio is 0.

)1(log SKN −= (3) Where, K is a material property coefficient (= 17: normal concrete, = 10: underwater and light weight concrete), S is ratio of design fatigue strength to design static strength and the maximum flexural stress and the flexural strength is used for them. From test results, coefficient K of steel-making slag concrete is calculated to 19 by using regression analysis. This shows flexural strength of steel-making slag concrete is slightly greater than that of normal

1 100 10000 10000000.4

0.5

0.6

0.7

0.8

0.9

1

Fatigue life, N (times)

Max

imum

stre

ss ra

tio, S

: Test results : Regression line (K=19) : S-N line of Normal concrete (K=17)

Figure 10. Relationship between maximum stress ratio and fatigue life

Table 5. Fatigue test results

Specimen Maximu- m stress （N/mm2）

Flexural Strength (N/mm2)

Maximu- m stress ratio, S

Fatigue life

(Cycle) FNP-90-1 4.93 5.43 0.91 296 FNP-90-2 4.93 5.45 0.91 2602 FNP-90-3 4.93 5.62 0.88 76 FNP-85 5.30 6.24 0.85 384 FNP-80-1 4.97 6.21 0.80 121413 FNP-75-1 4.38 5.75 0.76 42068 FNP-75-2 4.38 5.87 0.75 131561 FNP-75-3 4.38 5.90 0.74 422663 FNP-65-1 4.17 6.42 0.65 871837 FNP-60-1 3.65 6.08 0.60 636118

3 CMT - 22


CMT -

concrete about a region of 5N/mm2 flexural strength. Considering K of lightweight concrete and underwater concrete, flexural fatigue strength of steel-making slag concrete is almost equal to that of normal concrete. The strength of normal concrete with higher water content shows lower strength on the flexural fatigue strength. This is caused by water pressure. Therefore, it is said this result shows a safe side property. However, the influence of water content is considered to increase in higher stress ratio and investigation on the influence is required in future study.

4. CONCLUSION On basic mechanical property of steel-making slag concrete, followings are clarified: 1. No influence of a kind of alkali activator and quantity of cement content are seen in this study. 2. The flexural strength is equal to or higher than the strength of normal concrete at same compressive

strength. However, the strength test under several dry conditions considered the condition in the actual structure is required to confirm in a future study, because there is a possibility that the condition in actual structure is not same as the assumed conditions in this study.

3. The relationship between Young's modulus and compressive strength almost agrees with the relationship of normal concrete.

4. Characteristic on shrinkage is similar to normal concrete. However, there is a possibility that an Influence mechanism of water content to shrinkage is different from normal concrete or that a quantity of cement influences.

4. Thermal coefficient overlaps the lower-side rage of the normal concrete coefficient. 5. Flexural fatigue strength is almost equal to the strength of normal concrete.

5. REFERENCES A.M.Neville (1979). “Properties of concrete.” Gihodo Shuppan, Japan Coastal Development Institute of Technology (2003). “Technical Manual of Steel Making Slag

Concrete.”, Coastal Development Institute of Technology, Japan. Concrete Committee of JSCE (1996). “Standard Specifications For Concrete Structures-1996

"Design".” Japan Society of Civil Engineers, Japan Concrete Committee of JSCE (2003). “Standard Specifications For Concrete Structures-2002

"Structural Performance Verification".” Japan Society of Civil Engineers, Japan Shigenori, H., Junichiro, N., Tadaaki, T. (1995). “S Initial Stress Distribution and Flexural Strength of

Concrete Beams with Drying Shrinkage.” Proceedings of the Japan Concrete Institute, Vol. 17, No. 1, pp. 501-506.

Service Center of Port Engineering (1999). “Guideline for the Design of Airport Pavement.” The Service Center of Port Engineering, Japan, 22 pp.

3 CMT - 23


CMT -

MECHANICAL STRENGTH AND MICROSTRUCTURAL INVESTIGATIONS OF CIRCULATING FLUIDIZED BED COMBUSTION ASHES - GROUND VITRIFIED

BLAST FURNACE SLAG BLENDS

I Made Alit Karyawan Salain1

ABSTRACT : Paste of Circulating Fluidized Bed Combustion (CFBC) ashes - Ground Vitrified Blast Furnace (GVBF) slag blends has been investigated at different ages concerning their mechanical strength and microstructure. Each blend was made from a mixture of 15% CFBC ash and 85% GVBF slag by weight. The result of this investigation notes that all hardened pastes show good strength developing tendency. According to the type of CFBC ash, the flexural and compressive strength at 28 days can reach about 2.6-5.4 MPa and 30.7-54.0 MPa respectively. After 360 days, these values are about 5.6-7.0 MPa and 45.5-75.0 MPa respectively. This interesting development can be essentially attributed to the massive formation of C-S-H gel combined with certain quantity of ettringite, which produces a small amount of expansion. These products of hydration have been identified by X-ray diffraction and differential thermal analysis. KEYWORDS : CFBC ashes, GVBF slag, Strength, Thermal analysis, X-ray Diffraction

1. INTRODUCTION

Circulating Fluidized Bed Combustion (CFBC) represents one of the clean technologies for burning high sulfur combustibles in order to reduce the atmospheric pollution. This technology is particularly interesting because of its capacity to burn low quality combustibles and to reduce SO2 and NOx production. However, the addition of SO2 removing sorbents and the use of a lower combustion temperature (850-900°C) produces different kinds of ash compared to conventional coal combustion systems and creates a future management problem. The properties of CFBC ashes vary according to the nature and the quality of the coal burnt in the power plant. Generally, their specific gravity and specific surface are higher than for ashes produced from pulverized coal power plants that are usually used in the field of civil engineering. Due to desulphurising process in the power plant boiler, CFBC ashes contain more free CaO and SO3. Furthermore, there are fewer glass particles in this type of ash, which can be attributed to the lower combustion temperature used in this technology. Theses conditions render their characteristics quite different from those of ashes obtained from coal pulverized power plants. The hydraulic/pozzolanic property of CFBC ashes is possibly the most useful for civil engineers [1-4]. However, their use is still very limited because of the exothermic and expansive characteristics of certain CFBC ashes usually related to the presence of free lime (CaO) and anhydrite (CaSO4). Moreover, large variations in their physical, chemical and mineralogical properties increase the risk of using this ash. One author has proposed a process that allows for the selective hydration of quick lime content in fly ashes before use [5]. Another recommends the use of these ashes as raw materials for the synthesis of calcium sulphoaluminate cements [6]. In the present study, the blend of 3 types of CFBC ashes and GVBF slag has been studied from the point of view of strength and mineralogy. It is well known that both lime and anhydrite can be used respectively as the alkaline and sulfate activators of slag. This combined activator essentially stimulates the formation of C-S-H and ettringite, which will develop slag strength [7-10]. In the same way, the free lime and anhydrite present in these ashes will be used for activating slag hydration.

1 Department of Civil Engineering, University of Udayana, Bali- Indonesia

4 CMT - 24


CMT -

The characteristics of the materials studied are first described. Then the results drawn of the experimental study, carried out by hydrating blends of CFBC ash and GVBF slag, are presented and discussed.

2. MATERIALS AND EXPERIMENTAL METHOD

Some physical and chemical properties of two CFBC fly ashes (A1 and A2), a mix of 50% CFBC fly ash and 50% of CFBC bottom ash (A3) and GVBF slag used in this study are given in table 1 and 2 separately.

Table 1. Chemical Analysis of CFBC ashes and GVBF slag (% by Weight)

Materials Al2O3 CaO SiO2 Fe2O3 SO3 Free CaO LOI A1 16.28 15.34 39.36 6.67 7.40 4.98 10.30 A2 6.92 41.65 21.27 4.45 18.30 14.41 4.55 A3 4.56 50.85 10.45 3.57 25.49 20.49 2.40

GVBF slag 9.80 41.10 36.70 1.35 0.30 - 0.80 Each ash sample was used to prepare three blends, B1, B2 and B3 which respectively contain 15% A1, A2 and A3 and 85% GVBF slag by weight. Plain paste of each blend was used to cast 40x40x160 mm prisms. The water-cementitious materials ratio (w/cm) of each blend, given in table 3 was adjusted in order to produce a paste with a standard consistency according to European standard EN 196-3 [11]. These prisms were kept in molds at 20°C and 100% RH for 2 days, then cured in water at 20°C until used for the strength test. At 28, 90, 180 and 360 days of hydration, three prisms were used to measure flexural and compressive strength in accordance with European standard EN 196-1 [12].

Table 2. Physical properties of CFBC ashes and GVBF slag

Materials Fineness Blaine (cm2/g)

Specific gravity (g/cm3)

d50 (μm)

A1 6450 2.56 22.9 A2 6875 2.62 37.8 A3 - 2.78 50.0

GVBF slag 5650 2.93 6.9 The hydration products and the relative comparison of their progress over time were identified by X-ray diffraction (XRD) and Differential Thermal Analysis (DTA) after 1 day of hydration as well as after hydration and cured in water for 7, 28, 90, 180 and 360 days. The samples used for these analyses were taken from the central zone of the specimens, dried at 25°C and then crushed less than 100μm.

3. RESULTS AND DISCUSSION

3.1 Mechanical Strength Figure 1 illustrates the development of flexural and compressive strength of blends as a function of curing time. The development of both flexural and compressive strength is greatly influenced by the type of ash used. In general, the strength increases during the first 180 days and then it tends to stabilize or increases slowly. At 28 days, the flexural strength in the blends containing A1, A2 and A3 was 2.6, 3.5 and 5.4 MPa respectively and after 360 days it reached 5.8, 5.6 and 7.0 MPa. A strong increase was also observed in the development of the compressive strength. At the same period of hydration, the compressive strength in the blends using A1, A2 and A3 increased from 30.7, 47.0 and 54.0 MPa to 45.5, 68.5 and 75.0 MPa respectively.

4 CMT - 25


CMT -

Figure 1. Evolution of Flexural and Compressive Strength of Blends 3.2 Mineralogy 3.2.1 X-Ray Diffraction The results of XRD analysis indicate a quite similar phenomenon for all the samples. In general, while the anhydrite and the lime content in the blends disappears quickly, the formation of ettringite (3CaO Al2O3 3CaSO4 32H2O) was clearly identified. The formation of calcite (CaCO3) due to carbonation of the remaining lime can also be noted in each blend. On the other hand, quartz and merwinite are almost constant before and after hydration. Figure 2 shows an evolution of the X-ray diffractograms of the blend containing 15% A3 and 85% GVBF slag.

2 - T h e t a - S c a l e1 0 2 0 3 0 4 0

Figure 2. X-ray diffractograms of blend of 15% A3 and 85% GVBF slag A : Anhydrite, C : Free lime, Cc : Calcite, E : Ettringite, Mw : Merwinite, Q :Quartz

These diagrams show that the formation of ettringite in the blends increases essentially from the beginning until 7 days of hydration and tends to stabilize or slightly decrease in certain blends after this period. These diagrams also show that ettringite persists until 360 days of hydration. The intensity of ettringite increases according to the type and quantity of the CFBC ashes used in the blends. The blends containing more free lime and sulfate produced more ettringite than the others.

90 d

Mw

2-θ Scale

10 20 30 40

E

E

E

C A

E

E

Q

Q

Mw Mw

Mw

Cc

Cc

Anhydrous

1 d

7 d

28 d

180 d

360 d E

0

2

4

6

8

0 60 120 180 240 300 360

Time (day)

Flex

ural

Str

engt

h (M

Pa)

B1 B2 B3

0

20

40

60

80

0 60 120 180 240 300 360

Time (day)

Com

pres

sive

Str

engt

h (M

Pa)

B1 B2 B3

4 CMT - 26


CMT -

3.2.2 Differential Thermal Analysis Figure 3 shows DTA diagrams of the blend containing 15% A3 and 85% GVBF slag, obtained from 20°C up to 600°C, after 1, 7, 28, 90 and 180 days of hydration. Three principal endothermic peaks have been generally identified in these diagrams at the temperature ranges: 105-115°C, 120-130°C and 135-150°C.

100 200 300 400 500 600

Temperature (°C )

Figure 3. DTA diagrams of blend of 15% A3 and 85% GVBF slag The first peak indicates the dehydration of free water. During the long hydration process, it tends however to superimpose on the left side of the second peak, which is attributed to the water loss of, hydrated calcium silicate (C-S-H). The presence of this compound could not be observed by XRD analysis probably due to its amorphous structure. The C-S-H peak grows significantly during the curing time for all blends and tends to stabilize after 180 days. The third peak corresponds to the dehydration of ettringite. The presence of this compound observed by DTA confirms the previous result of XRD. The ettringite peak increases noticeably during the first seven days of hydration and tends to stabilize after that.

3.3 Discussion The mechanical performances presented by the blends are strongly influenced by the type of the CFBC ashes used. This can essentially be related to their different chemical compositions, especially their free CaO and CaSO4 content. The presence of these elements in the blend stimulates the slag hydration and the strength development of the blends. This is clearly indicated by the formation of ettringite and C-S-H in the blends as suggested by the XRD and DTA. The formation of ettringite combined with the massive formation of C-S-H develops the blend strengths. In fact, the presence of a certain quantity of ettringite reinforces paste solidification which produces a small amount of expansion [13]. This is also confirmed by other studies of slag cement

1 d

7 d

28 d

180 d

90 d

4 CMT - 27


CMT -

hydration [14-15]. However, it is also probable that the reactive silica and alumina existing in the CFBC ash play an important role in strength development on account of certain hydraulic property [4].

Thus, the presence of CFBC ash in the blends activates slag hydration attributed to the free CaO and CaSO4, which act as the alkaline and sulfate activators respectively. Moreover, it simultaneously reinforces the hardening process of the blend supplied by those activators and also the reactive silica and alumina existing in the CFBC ash.

4. CONCLUSION

• The mechanical performances presented by the blends are strongly influenced by the chemical

composition of the CFBC ashes used, especially on their free lime and sulfate contents. • By using 15% of CFBC ash, the flexural and compressive strength of the blends, at 28 days, can

reach about 2.6-5.4 MPa and 30.7-54.0 MPa respectively, according to the type of CFBC ash used. These strengths reach about 5.6-7.0 MPa and 45.5-75.0 MPa respectively after 360 days.

• This increased performance can be related to the combination of the massive formation of C-S-H

gel and a certain quantity of ettringite that does not produce any significant expansion in the blends.

5. REFERENCES

A. E. Bland (1991). “Ash management issues for Fluidized Bed Combustion technologies.” 1991 CANMET CFBC ash management seminar, E. J. Anthony and F. Preto (Ed), Nova Scotia, 207-276.

A. Carles-Gibergues, C. Delsol (1994). “CFBC ashes: prospects for use as mineral additions for concrete.” (in French), Annales de l’ITBTP 525, 47-59.

A. Tassart, J. Blondin, J.M. Siwak (1997). “Caractérisation de cendres sulfo-calciques provenant d’une chaudière à Lit Fluidisé Circulant : Comparaison des cendres avec et sans traitement sélectif de la chaux vive. “ Annales du Bâtiment et des Travaux Publics, 17-29.

I.M.A.K. Salain, P. Clastres (2000). “Valorisation des cendres de Lit Fluidisé Circulant.“ Communication au 3ème Congrès Universitaire du Génie Civil, Lyon, 369-375.

J. Blondin, E.J. Anthony (1995). “A selective hydration treatment to enhance the utilization of CFBC ash in concrete. “ Proc. 13th Int. Conf. on FBC, ASME, Orlando, Florida, pp. 1123-1127.

G. Bernardo, M. Marrocoli, F. Montagnaro and G.L. Valenti (2000). “Calcium sulphoaluminate cements made from fluidized bed combustion wastes.” Proceedings of the international conference on the science and engineering of recycling for environmental protection, G.R. Woolley, J.J.J.M. Goumans, P.J. Wainwright (ed), Wascon, pp. 750-758.

I.-A. Voinovitch, R. Dron (1976). “Action des différents activants sur l’hydratation du laitier granulé. “ Bull. Liaison Labo. P. et Ch., 83, 55-58.

R Dron (1984) “Nouveaux activants du laitier.“ Laitiers de Hauts Fourneaux, 57-3, 7-17. J. Alexandre, J.L. Sebileau (1988). “Le Laitier de Haut Fourneau: Elaboration - Traitement -

Propriétés – Emplois. “ Centre Technique et de Promotion des Laitiers. F.M. Lea (1970). “The Chemistry of Cement and Concrete.” third edition, Edward Arnold Ltd,

London. EN 196-3 (1994). “Methods of Testing Cement: Determination of setting time and soundness.”

European standard, Brussels. EN 196-1 (1994). “Methods of Testing Cement: Determination of strength.” European standard, CEN,

Brussels.

4 CMT - 28


CMT -

I.M.A.K. Salain (2001). “Réactivité des cendres de combustion en Lit Fluidisé Circulant et des mélanges cendres - laitier de haut fourneau.” Ph.D. Thesis INSA de Lyon, France.

M. Moranville-Regourd (1980). “Structure et comportement des hydrates des ciments au laitier.“ Proceedings of the 7th International Congress on the Chemistry of Cement, Paris, Volume I, III-2/9– III-2/26, Editions Septima Paris (France).

Li Dongxu, Wu Xuequan, Shen Jinlin, Wang Yujiang (2000). “The influence of compound admixtures on the properties of high-content slag cement.” Cement and Concrete Research 30, 45-50.

4 CMT - 29


CMT -

PRELIMINARY STUDY ON THE COMPACTION OF CONCRETE CONTAINING SLAG AS COARSE AGGERGATE BY USING HORIZONTAL VIBRATION

M.W. Tjaronge1 and Shinji Kawabe 2

ABSTRACT: One of the conventional methods for compaction concrete in manufacturing the precast concrete is the vibrating table. Vibrating table vibrates the mold vertically to compact fresh concrete in the mold. The vibrating table can attain the excellent compaction of concrete but during the compaction process, it generates the noise from 90 dB to 120 dB. This noise spoils the safety of the workers at the concrete plant and annoys the people who live nearby. When the horizontal long amplitude and low frequency are used to shake the mold, it is possible to control less noise level than 81 dB during the compaction process. This method is referred to as the horizontal vibration. Recently, researches and developments of the slag aggregate for concrete have been carried out. The purpose of this study is to examine the potential use of horizontal vibration to compact the concrete containing slag aggregate in the mold of 0.1 m x 0.1 m x 0.4 m. The electric arc furnace oxidizing slag was used as coarse aggregate and river sand was used as fine aggregate. The amplitude of horizontal vibration was 0.10 m. The frequencies of horizontal vibration were divided into fixed and variable condition. The fixed frequencies were 1.0Hz and 1.25Hz.These frequencies used only one frequency when fresh concrete was filled and consolidated in mold until covers the top of the mold. The variable frequencies were 1.0Hz-0.75Hz and 1.25Hz-0.75Hz. The first frequency was set until fresh concrete filled half of the mold then it was turned down to the second one until fresh concrete covers the top of the mold. Fresh concrete was divided and filled in two layers into the mold. Each layer was shaken horizontally in 45 seconds. This study focused on evaluating the result of compaction by using the horizontal vibration and the flexural strength of concrete. Also, the distribution of slag within the concrete was evaluated. Experimental results show at all frequencies tested, horizontal vibration can compact the concrete without honey comb. The average of flexural strength was 5.39 MPa. Distribution of slag aggregate was uniform without segregation. KEYWORDS: horizontal vibration, slag aggregate, compaction, flexural strength, distribution of slag aggregate

1. INTRODUCTION

One of the conventional methods for compaction concrete in manufacturing the precast concrete is the vibrating table. Vibrating table vibrates the mold vertically to compact fresh concrete in the mold. The vibrating table can attain the excellent compaction of concrete but during the compaction process, it generates the noise from 90 dB to 120 dB. This noise spoils the safety of the workers at the concrete plant and annoys the people who live nearby. When the horizontal long amplitude and low frequency are used to shake the mold, it is possible to control less noise level than 81 dB during the compaction process. This method is referred to as the horizontal vibration. Recently, researches and developments of the slag aggregate for concrete have been carried out. The purpose of this study is to examine the potential use of horizontal vibration to compact the concrete containing slag aggregate as coarse aggregate. This study focused on evaluating the result of compaction by using the horizontal vibration and the flexural strength of concrete. Also, the distribution of slag within the concrete was evaluated.

1 Lecturer, Department of Civil Engineering, Hasanuddin University, Makassar Indonesia 2 Professor, Department of Environmental Technology and Urban Planning, Nagoya Institute of Technology, Japan

5 CMT - 30


CMT -

2. EXPERIMENTAL

2.1 Horizontal Vibration

Figure 1 illustrates the movement of the horizontal vibration. Its properties are shown in Table 1. Horizontal vibration moves the mold horizontally with long amplitudes and low frequencies to compact the concrete. The amplitude of horizontal vibration was 0.10 m. The frequencies of horizontal vibration were divided into fixed and variable condition. The fixed frequencies were 1.0Hz and 1.25Hz. These frequencies used only one frequency when fresh concrete was filled and consolidated in mold until covers the top of the mold. The variable frequencies were 1.0Hz-0.75Hz and 1.25Hz-0.75Hz. The first frequency was set until fresh concrete filled half of the mold then it was turned down to the second one until fresh concrete covers the top of the mold. Fresh concrete was divided and filled in two layers into the mold. Each layer was shaken horizontally in 45 seconds.

2.2 Materials All mixes of SCC were prepared by mixing water, Type I Portland cement, superplasticizer, sand and slag. Figure 2 the electric arc furnace oxidizing slag. It was used as coarse aggregate and its chemical composition was shown in Table 2.

Figure 2. Slag (electric arc furnace oxidizing slag)

Roller

Mold

Horizontal direction

Figure 1. Horizontal vibration

Table 1. Properties of horizontal vibration

Properties Horizontal vibration Noise (dB) ≤ 81

Direction of vibration → ←

(horizontal) Amplitude (m) 0.05 to 0.15 Frequency (Hz) ≤ 1.5

5 CMT - 31


CMT -

Table 2. Chemical composition of slag

The tests for physical properties aggregates were designed in accordance with ASTM C33-03 and their physical properties aggregates are shown in Table 2. The sand was used in surface dry condition.

2.3 Mix Design of Concrete Table 3 shows mix design of concrete. The proportion was designed according to JIS A 5308-1996 (Ready mixed concrete) and JASS 5 (Japanese Standard Specification for Reinforced Concrete Work). The average of slump value of fresh concrete and the compressive strengths of cylindrical specimens (100 x 200 mm) at 28 days were 16.6 cm and 56.52 MPa, respectively.

No. Chemical composition (%)

1 CaO 20.69

2 SiO2 13.63

3 MnO 7.97

4 MgO 3.34

5 FeO 30.79

6 Fe2O3 5.67

7 Al2O3 10.32

8 Cr2O3 3.55

9 TiO2 0.47

10 P2O5 0.431

11 S 0.011

12 Total 96.87

13 T.Fe 27.90

Table 3. Mix design of concrete

Volume 1 m3

Aggregates

W/C

(%) Water

(kg)

Ordinary Portland cement

(kg) River sand

(kg)

Slag

(kg)

Superplasticizer

(kg)

40 166 414 721 1405 3.105

Table 2 Physical properties of aggregates Aggregate

Properties Slag River sand

Maximum size (mm) 20 2.50

Surface dry 3.66 2.45 Density (kg/l)

Oven dry 3.63 2.35

(24 hours) Water absorption (%) - 2.74

Fineness (%) - 2.50

5 CMT - 32


CMT -

2.4 Flexural Strength Test The flexural strength was measured at the age of 28 days, the test was conducted according to JIS 1106-1999. Each flexural strength value was the average of 3 beams test result.

2.5 Distribution Of Slag After flexural strength was done, each beam of slag concrete was split to detect the effect of horizontal vibration on the distribution of slag within the concrete. The split surface was divided in two layers, the bottom layer (5x5 cm2) and the top layer (5x5 cm2). The slag that appear on the split surface were counted and average slag densities (i.e., number of slag per layer) were determined for top layer and bottom layer. The densities of the top layer and the bottom layer were compared to evaluate the homogeneity of the slag distribution.

3. RESULTS AND DISCUSSIONS

3.1 Flexural Strength

The visual inspections of the compaction results show that the horizontal vibration can fully compact the slag concrete without honeycomb at the all sides of the specimens. Figure 4 shows the flexural strength of the slag concrete. The flexural strength ranged from 5.20 MPa to 5.52 MPa, with an average of 5.39 MPa. At the constant compaction time amplitude, the increase and decrease in frequency will not give a significant effect to the flexural strength of the concrete beams. It shows the flexural strength can be achieved when the compaction has reached the good results.

Figure 4. Flexural strength of the slag concrete

3.2 Distribution Of Slag

Figure 5 shows the slag densities within the top layer and bottom layer of specimens. The slag densities within the top layer and bottom ranged from 1.13 to 1.33 (slags/cm2) and from 1.14 to 1.47 (slags/cm2), respectively. No significant difference of slag density was shown between the top layer and bottom layer. This result reveals that horizontal vibration gave homogeneity of the slag distribution within the specimens.

0.001.00

2.003.00

4.005.006.00

7.008.00

9.0010.00

1.00Hz 1.25Hz 1.00Hz - 0.75Hz 1.25Hz - 0.75HzFrequencies of horizontal vibration

Flex

ural

stre

ngth

(MPa

)

5 CMT - 33


CMT -

Figure 5. The slag densities within the top layer and bottom layer of specimens

Observation of the split surface did not show any segregation of aggregates. Concrete with slag fully filled the beam without large voids occurred. Slag aggregates can be covered by a layer of cement paste. No pilling of slag, mortar or cement paste occurred at the lower end of the specimens. Figure 6 shows that an excellent bond was established between the slag and the matrix, resulting in a high flexural strength.

Figure 6. Observation of the split surface

4. CONCLUSIONS

1. Horizontal vibration can fully compact the slag concrete. 2. The distribution of slag within the specimens was practically homogenous. 3. Horizontal vibration gave an excellent bond between the slag and the matrix.

0.00

0.50

1.00

1.50

2.00

1.00Hz 1.25Hz 1.00Hz - 0.75Hz 1.25Hz - 0.75HzFrequencies of horizontal vibration

Slag

den

sity

(sla

gs/c

m2)

Top layerBottom layer

Top layer (5x5 cm2)

Bottom layer (5x5 cm2)

5 CMT - 34


CMT -

Placing hole of fresh concrete

Mold

Control system

Direction of horizontal vibration Turn plate

400

100

400

150

75

100

50

5. ACKNOWLEDGEMENTS

The authors are grateful for the support of this work by the Graduate School VBL Division, Techno-Innovation Center, Nagoya Institute of Technology.

6. REFERENCES

ACI Committee 309 (1981). “Behavior of Fresh Concrete During Vibration.” ACI Journal, Proceeding V.78, No.1, January-February, pp.36-53

JIS A 1132 (1999). “Standard Practice for Making and Curing Concrete Test Specimens in the Laboratory.”

JIS A 1106 (1999). “Method of Test for Flexural Strength of Concrete.” K.Kokubu and S.Yotsuya (2003). “The out line of the quality of electric arc furnace oxidizing slag

aggregate.” Concrete Journal, Vol. 41, No.8, pp.3-7 (in Japanese) S. Kawabe, T. Okajima, Tjaronge M.W and M. Muto (1999). “Proposal of horizontal vibration method

for manufacturing precast concrete.” Proceedings of the Japan Concrete Institute, Vol.21, No.2, pp.943-948 (in Japanese)

Tjaronge M.W., S. Kawabe, T. Okajima, M. Muto and S. Matsuoka (1999). “Compaction of fresh concrete with long amplitude and low frequency of vibration in manufacturing precast concrete.” Proceedings of the Second International Conference on Advanced Materials Development and Performance Evaluation and Application, pp.501-506, The University of Tokushima.

Tjaronge M.W., S. Kawabe, M. Muto and T. Okajima (2001). ”Study on the compaction of concrete by horizontal vibration with low frequency and large amplitude.” Journal of Structure and Construction Engineering-Transactions of AIJ, No.545, pp.7-11. (in Japanese)

6. APPENDIX

Figure A1 and A2 show the experimental equipment of horizontal vibration that used in the present work.

Figure A2. Detail of experimental equipment of the Horizontal Vibration (dimension: mm)

Figure A1. Experimental Equipment of the Horizontal Vibration

5 CMT - 35


CMT -

STUDY ON THE APPLICATION OF LOW- QUALITY RECYCLED COARSE AGGREGATE TO CONCRETE STRUCTURE

BY SURFACE MODIFICATION TREATMENT

Masato Tsujino1, Takafumi Noguchi2, Masaki Tamura3, Manabu Kanematsu4, Ippei Maruyama5 and Hironori Nagai6

ABSTRACT: This paper aims to establish a technique for recycling concrete easily in view of the essential solution toward establishing closed-loop recycling society. In establishing recycle-oriented concrete industry, the current recycle process, requiring a high-level modification technique such as heating and rubbing, still includes many problems, e.g., energy-reduction cost-saving and fine powder treatment. The recycling system of concrete structure is being significantly improved under the emphasis on more enhanced environmental consciousness. Japanese Industrial Standards (JIS) have been developed to put recycled concrete aggregate into practical use. The technique introduced in this study enables the improvement of recovery rate of original aggregate by enhancing a peeling-off effect of aggregate without damaging any mechanical properties. The enhanced peeling-off effect is realized by applying a surface improving agent to aggregate. In addition, high water absorption of recycled aggregate is also reduced. In this paper, material tests were conducted on recycled aggregates with low quality and middle quality. In the test, two types of surface improving agent, oil-type and silane-type, were used. The test results have shown that the recycled aggregate finished with silane-type improving agent was greatly improved in recovery rate but showed lowered strength. On the other hand, the recycled aggregate finished with oil-type improving agent is slightly improved in recovery rate comparing with that of non-finished aggregate. Reinforced concrete beams were made using the oil-finished aggregate, while the aggregate finished with silane-type improving agent was excluded because it was considered difficult to be used for structure. The reinforced concrete beams were subjected to the bending test. The bending test results have revealed that the flexural capacity of reinforced concrete beam with aggregate finished with oil-type improving agent was approximately equal to that of conventional reinforced concrete beam with non-finished crushed stone. The resistance to cracking caused by bending was within an allowable range of fully practical use. Consequently the possible applicability of recycled aggregate finished with oil-type surface improving agent was verified.

KEYWORDS: Concrete recycling, Aggregate recovery, Recycled coarse aggregate, Surface coating, Strength, Drying shrinkage, Creep, Bending behavior

1. INTRODUCTION

The recycle system of concrete is now being significantly improved under enhanced awareness of environment and vociferous request for recycling along with the JIS-standardization of high-quality recycled aggregate for widespread use. In establishment of recycle-oriented society, the reverse process, requiring high techniques like Heat-Treatment and Rubbing Method, still faces many subjects1), 2) such as energy-, cost-saving and treatment of such by-product as powder. The labor-saving of this reverse-process is essential to attain a closed loop easily. Paying particular attention to these technical points, this study aims to establish a technique for easy recycling of concrete. The technique introduced to Concrete with Easy-to-collect Aggregate3) in this study is easy process by applying surface improving agent to aggregate. This technique enhances the peeling effect of

1 Graduate Student, M. Eng., Dept. of Arch., The Univ. of Tokyo, Japan 2 Assoc. Professor, Dr. Eng., Dept. of Arch., The Univ. of Tokyo, Japan 3 Research Associate, Dr. Eng., Dept. of Arch., Tokyo Metro. Univ., Japan 4 Assist. Professor, Dr. Eng., Dept. of Arch., Tokyo Univ. of Science, Japan 5 Assoc. Professor, Dr. Eng., Dept. of Environmental Studies, Nagoya Univ., Japan 6 Research Associate, Dr. Eng., Dept. of Arch., The Univ. of Tokyo, Japan

6 CMT - 36


CMT -

Table 3. Test results original concrete of middle-quality recycled

coarse aggregate

aggregate from cement matrix without damaging the mechanical property and reduces the high water absorption of recycle aggregate. The objectives of the study are to clarify the influence of surface improving agent on the property of recycled coarse aggregate the peeling effect relating to the recovery of coarse aggregate as well as to investigate the hardening property of concrete and the flexural strength of reinforced concrete beams to discuss the usability of this technique as aggregate for highly recyclable concrete.

2. CHARACTERISTICS OF RECYCLED COARSE AGGREGATE

2.1 Quality The quality of recycled coarse aggregate used in this study is shown in Table 1. The recycled coarse aggregate referred to as middle-quality is screw-ground with a tertiary crushing after roughly crushing with a jaw/cone crusher and its particle size is so adjusted that the grain is ranged within the standard particle size specified in JIS. On the other hand, the recycled coarse aggregate referred to as low-quality is ground only without a screw grinding and a particle size adjustment. The particle size distribution curves of both recycled aggregates are shown in Fig. 1. For test materials, the middle-quality recycled aggregate was produced in the laboratory and its mix proportion was known. Waste concrete used for subgrade material, whose mix proportion was unknown, was purchased as the low-quality recycled aggregate. The mix proportion of the original concrete of the middle-quality recycled aggregate and the test results of the concrete are shown in Table 2 and Table 3, respectively.

Fine Coarse (N/mm2)

58.0 49.1 180 310 858 909 C×1.3% （AE & Water reducing agent） 18.0 5.5 21.4

Air(%)

Unit content(kg/m3)

w/c(%)

s/a(%) Water Cement

Compressive strengthat the time of crushAggregate Admixture

Slump(cm)

2.2 Content of paste and mortar in recycled coarse aggregate The fact has been reported that much more cement paste included in recycled aggregate would affect an adverse effect on high water absorption and property of hardened concrete. This is because cement paste is more porous than aggregate. Consequently, the content of cement paste is a typical index representing the quality of recycled aggregate. In this study, the constituents in recycled coarse aggregate were determined with a

Table 2. Mix proportion of original concrete of middle-quality recycled coarse aggregate

Table 1. Quality of recycled coarse aggregate Types of coarse

aggregate Code

Oven-dry density (g/cm3)

Surface-dry density (g/cm3)

Water absorption

(%)

Material passing 75μm sieve

(%)

Mass per unit volume

(kg/L)

Solid content in aggregate

(%)

Fineness modulus

Middle-quality M 2.36 2.47 4.80 0.64 1.51 64.1 6.51

Low-quality L 2.33 2.46 5.48 2.10 1.41 60.5 6.24

2.5 5 10 20 250

20

40

60

80

100

Sieve size (mm)

Per

cent

age

Pas

sing

(%)

Range of std sizeMiddle-qualityLow-quality

Figure 1. Grading curves of recycled coarse aggregate

46.1%

72.8%

100%

13.6%

17.5%

36.4%

13.6%

0

20

40

60

80

100

Originalcoarse aggregate

Middle-quality Low-quality

Mix

ture

wei

ght

(%)

Original coarse aggregate Fine aggregate Paste

Figure 2. Deposit rate of paste of recycled coarse aggregate

6 CMT - 37


CMT -

chlorine dissolution process. The results are shown in Fig. 2. The difference in cement paste content between middle- and low-quality recycled coarse aggregate is approximately three times, and that of mortar content is about two times. About 30% of total weight in middle-quality aggregate and more than 50% in low-quality aggregate are mortar. This fact suggests that the grinding effect of tertiary crushing has reduced the mortar deposit rate of middle-quality recycled coarse aggregate. By contrast, in the low-quality recycled coarse aggregate, the mortar deposit rate has not been reduced by rough crushing only.

3. CHARACTERISTICS OF COATED RECYCLED COARSE AGGREGATE

3.1 Types of surface improving agent The surface improving agents used in this study are oil- and silane-type agents, shown in Table 4, showing conventional application and main constituent4), 5). The schematic diagrams of the effect of surface improving agents are shown in Figures 3 and 4.

Table 4. Types of surface improving agents Type Oil (O) Silane (S)

ApplicationRelease agent which is used in

wooden form

Water repellant agent ofpenetration type to the concrete

surface

Mainconstituent

Mineral oil(Paraffin),Emulsifying agent,

Lanolin fatty acid salt, Water

Silicon analogue,Emulsifying agent,

WaterState Emulsion Emulsion

Saponification＆ Hydrolysis reaction

→ Alkali metal salt formationCa2+

Coating formation of alkali metal salt

Calcium ion

== Mineral oil

Application

Drying

RCOOCHRCOOCH22

RR’’COOCHCOOCH

RR””COOCHCOOCH22

RCOOCHRCOOCH22

RR’’COOCHCOOCH

RR””COOCHCOOCH22

Ca2+ Ca2+

RCOOCaRCOOCaRR’’COOCaCOOCaRR””COOCaCOOCa

RCOOCaRCOOCaRR’’COOCaCOOCaRR””COOCaCOOCa

Alkali metal salt film

Surfaced of aggregate



= Si-OR =Si-OH

1. FusionParticles are fused each otherfollowing the water evaporationfrom emulsion.

2. Dealcoholization by hydrolysisSilanol is formed by the reaction

of alkoxyl group, maintained inpolymers, with water.

3. BondingSilanol is reacted with hydroxylgroup of silicate contained in cementin order to bond to substrate.Condensation.Further, after water evaporation,the silanol reacts with another �silanol to form siloxane cross-link.

Application Mineral oil

Waterevaporation

↓Approach

between particles

Hydrolysisreaction

Cross-link formation



Reaction Process

Coating formation

Water-repellent coating is formed on the surface of aggregate.

Symbol used in this paper is as follows:

＊-＊(-＊) → Aggregate - Surface improving agent (-W/C) 3.2 Water absorption of coated recycled coarse aggregate The surface improving agent, which is dispersed in water with specified concentration, was applied by repetition of spraying and drying. The number of repetitions was four times to get stable coating. The water absorption test was conducted according to JIS A 1110. The test results of water absorption in accordance with the number of repetitions are shown

Figure 4. Schematic drawing of a silane-type surface improving agent6)

Figure 3. Schematic diagram of an oil-type agent for surface improvement

M-O3.78%

M-S 1.07%

M 4.80%

L-O3.53%

L-S 1.15%

L 5.48%

0.00

1.00

2.00

3.00

4.00

5.00

6.00

0 1 2 3 4Number of Applications

Wat

erAbs

orpt

ion

(%)

M-O

M-S

L-OL-S

Figure 5. Reduction of water absorption with improving agent

6 CMT - 38


CMT -

in Fig.5. Reduction of water absorption is not proportional to the number of repetitions. This might be due to uneven application over whole recycled coarse aggregate including paste parts of high water absorption. Ultimate water absorption is 3.5% for an oil-type agent and 1% for a silane-type agent, indicating that the reduction effect of water absorption of a silane-type is higher than that of the oil-type agent.

4. TESTS FOR RECYCLED CONCRETE USING COATED RECYCLED COARSE AGGREGATE

4.1 Types of surface improving agent Types of concrete and mix proportion are shown in Table 5. Two types of recycled coarse aggregate are used, i.e., middle-quality and low-quality. Three types of surface treatment are applied, i.e., no treatment (N), oil treatment (O), and a silane-type treatment (S). Concrete was made with two levels of water/cement ratio, i.e., 60% and 40%. Thus, twelve types of concrete were totally prepared. Ordinary Portland cement (density: 3.16g/ cm3) was used as cement. Oigawa River sand with surface dry density of 2.59g/cm3, water absorption of 0.59% and fineness modulus of 2.66 was used as fine aggregate. The mix proportion has been made in such a way that the recycled concrete using untreated coarse aggregate satisfies target properties in fresh state shown in Table 5. Supplemental air-entraining agent was used unless the targeted air volume was not obtained. The aggregates treated with surface improving agent were used at air-dry condition to refrain from separation of the agent under water, especially in the silane-type agent.

Table 5. Types of concrete and mix proportions

Water Cement Fine Coarse

Unit content(kg/m3)Slump

(cm)Air(%)

844 906

Surfaceimproving

agent

60.018±2

s/a(%)

47.0

42.0

C×0.7%（Superplasticizer）

250ml/c=100kg（AE & water reducing agent）

185 308 826 883

725 950

Admixture(g/m3)

Middle-quality (M)

Oil (O)

Silane (S)

No treatment (N) 4.0(±1) 47.0 175

Aggregatetype

W/C(%)

250ml/c=100kg（AE & Water reducing agent）

40.0 2.0(±1) 42.0 165 413 743 981

292

Low-quality (L)

60.018±2

4.0(±1)

40.0 2.0(±1)Oil (O)

No treatment (N)

C×0.7%（Superplasticizer）Silane (S) 175 438

4.2 Experiments on mechanical properties The compressive strength test and split tensile strength test were conducted at the age of 28 days according to JIS A 1108 and JIS A 1113. The test results are shown in Figures 6 and 7. Compressive strength slightly increased in the concrete with W/C of 60% made using recycled coarse aggregate treated with the oil-type agent. The increase might be due to the air-dry condition of the aggregate. The strength reduction in the concrete with W/C of 40% might be due to film forming, which caused the reduction of bond strength between aggregate andcement matrix. However the strength decrease is only 10% from the untreated specimens. Consequently, enough strength might be obtained in the application of the oil-type agent. On the contrary, in all specimens treated with the silane-type agent, considerable strength reduction was observed, whcih might be due to significantly weakened bonding properties. The above test results show that the aggregate treated with the oil-type agent can be used as conventional concrete aggregate, because no significant strength reduction is recognized and film forming can be found on the aggregage surface which may offer advantages in recovery of original aggregate. The mix proportion in the concrete made using recylced coarse aggregate treated with the oil-type agentcan be designed in the same manner as ordinary concrete. By contrast, the use of a silane-type agent decreases the strength considerably and the usual mix proportioning isimpossible. Further investigations are needed to clarify the relationship between the amount of application of the silane-type agent and the strength reduction.

6 CMT - 39


CMT -

0

10

20

30

40

50

M-60 L-60 M-40 L-40

Com

pre

ssiv

est

rengt

h

O : Oil S : SilaneN : No treatment

2in

28

days　

(N/m

m)

0

1

2

3

4

M-60 L-60 M-40 L-40

O : Oil S : SilaneN : No treatment

Split

tensi

lest

rengt

h 2in

28

days　

(N/m

m)

4.3 Experiments on peeling-off effect (Recovery of original aggregate) The importance of peeling-off effect exists in, as shown in Concrete with Easy-to-collect Aggregate3), recycling at low energy and keeping a size of aggregate. In this study, area of aggregate on the split surface of the specimen was measured by means of image analysis proposed by M. Tamura et al4),7) after split tensile strength test to evaluate the peeling-off effect of surface improving treatment and the recovery of original aggregate. The peeling-off effect was determined by the ratio of the aggregates which are peeled at the coated surface to total recycled aggregates. To clearly distinguish between the peeled aggregates and the crushed aggregates specimens were prepared by adding a red pigment of iron oxide of 3% of cement weight. The results of the image analysis in the concrete with W/C of 60% are shown in Fig.8. It is confirmed that almost 70% of untreated aggregate is crushed so that a size of aggregate cannot be maintained. Although the peeling-off effect is slightly improved in the aggregate treated with the oil-type agent comparing with that of untreated aggregate, the improvement does not satisfy the requirement. As the strength decreased in the concrete with W/C of 40%, the better peeling-off effect and the higher recovery rate are expected. By contrast, it is obvious that almost all aggregate treated with a silane-type agent is peeled off at the interface between aggregate and cement paste while the reduction of strength is significant. The future subject is to balance the trade-off relationship between the peeling-off effect and the mechanical properties so as to apply the surface-treated recycled aggregate to structure concrete. 4.4 Experiments on drying shrinkage A drying shrinkage test was conducted according to a dial gauge method specified in JIS A 1129-3. Specimens were removed from moulds at the age of one day and cured in water up to the age of nine days, followed by dry condition at 20℃ and 60% R.H. Figure 9 shows the experimental results of drying shrinkage and mass change and the predicted values of drying shrinkage up to the age of 726 days, which are calculated using AIJ equations8) for natural aggregate concrete (hereinafter referred to as ordinary concrete) with the same mix proportion.

Figure 7. Test results of split tensile strength Figure 6. Test results of compressive

Figure 8. Results of an image analysis on peeling-off effect in the concrete with

W/C=60%

＋＝

＋＝

＋＝

No treatment　　 Cracked surface 67% Peeled surface 33%

Oil　　　　 Cracked surface 57% Peeled surface 43%

Silane 　　　 Cracked surface 16% Peeled surface 84%

6 CMT - 40


CMT -

In case of W/C of 60%, the drying shrinkage of concrete containing recycled aggregate treated with the oil-type agent is smaller than others. The drying shrinkage of middle-quality aggregate is smaller by 10% comparing with that of untreated aggregate. The reduction of mass in concrete containing recycled aggregate treated with the oil-type agent is smaller than that of untreated aggregate and almost equal to that of ordinary concrete, which might demonstrate little problem in practical use. Very small effect of a silane-type agent was observed in drying shrinkage and mass change. In case of W/C of 40%, the drying shrinkage of concrete containing recycled aggregate treated with the oil-type agent is smaller than that of concrete with untreated aggregate. The oil-type agent is effective in the reduction of drying shrinkage regardless of the water/cement ratio, which gives concrete structures long service life and leads to sustainability. 4.5 Experiments on creep Figure 10 shows the experimental results and predicted values, which are calculated by AIJ equation for ordinary concrete with the same mix proportion, of the change of specific creep strain. The specific creep strain was so calculated that both the elastic strain at loading and the drying shrinkage strain were subtracted from the total strain. The specific creep strain in recycled aggregate concrete is greater than that in ordinary concrete regardless of water/cement ratio. This phenomenon might be due to the paste content in concrete made using recycled aggregate. Concrete with recycled aggregate treated with the oil-type agent shows nearly equal change in creep behavior to concrete with untreated aggregate, which proves no significant influence of the oil-type surface improving agent on creep. By contrast, the creep strain of concrete with recycled aggregate treated with a silane-type agent is very large. This phenomenon might be resulted from the decreased bond strength at an interface between aggregate and cement paste.

Experimental AIJ shrinkage strain prediction equation (naturalWeight rate of change

0

200

400

600

800

1000

1200

Dry

ing

shrinka

gest

rain

（μ

）91

92

93

94

95

96

97

Weig

ht

rate

of

chan

ge（％

）

M-N

-60

M-O

-60

M-S-60

L-N

-60

L-O

-60

L-S-60

M-N

-40

M-O

-40

M-S-40

L-N

-40

L-O

-40

L-S-40

aggregate concrete of the same formulation)value

Figure 9. Drying shrinkage strain at the age of 726 days

Figure 10. Specific creep strain

0

20

40

60

80

100

120

140

160

180

0 100 200 300 400 500Time since loading (days)

Specific

cre

ep

stra

in（×

10

/(N

/m

m)）

2-6



0 100 200 300 400 500 600Time since loading (days)

Middle　W/C=60％ Low　W/C=60％

Middle　W/C=40％ Low　W/C=40％

No treatment Oil Silane

AIJ creep strain prediction equation(natural aggregate concrete of the same formulation)

0

20

40

60

80

100

120

140

160

180

Specific

cre

ep

stra

in（×

10

/(N

/m

m)）

2-6

6 CMT - 41


CMT -

4.6 Experiments on flexural properties of reinforced concrete beams using aggregate treated with surface improving agent

4.6.1 Outline Flexural tests have been conducted in reinforced concrete beams made using concrete with aggregate treated with surface improving agent, except for concrete with aggregate treated with a silane-type agent, which was greatly reduced in strength in spite of an excellent recovery of aggregate and was considered to be difficult to be used for structure. The aim of the flexural test is to evaluate the strength and the cracking resistance of the concrete with aggregate treated with the oil-type agent, and to check the practical use comparing with ordinary concrete made using crushed virgin aggregate. The outline of loading is shown in Fig. 11. 4.6.2 Cracking moment Figure 12 shows experimental results and calculated values obtained by substituting the mechanical properties at loading age into the equation. As shown in Figure 12, cracking moment in recycled coarse aggregate concrete is slightly smaller than that in ordinary concrete. However, no singular point is recognized in the concrete with recycled coarse aggregate treated with the oil-type agent and the experimental results are nearly equal to the calculated values, therefore a conventional equation for designing can be used to predict a bending moment causing cracks. 4.6.3 Cracking behavior To investigate properties of cracks against the long-term allowable stress, Figures 13 and 14 show the results of cracking properties due to serviceable load when, assuming a RC section, the force applied on the main reinforcement attained the long-term allowable stress, 215N/mm2. Figure15 shows the stress when cracks occurred. The maximum crack width did not increase by the use of the oil-type agents. This might be considered that the effect of bond of recycled coarse aggregate is smaller than that of ordinarfy aggregate. However, there would be no problem in practical use, because for all the levels, the extent of cracks is significantly smaller than 0.3 mm, i.e., allowable crack width to ensure the resistance to degradation in general environment specified in a Recommendations for Practice of Crack Control in Reinforced Concrete Buildings (Design and Construction)8) published by Architectural Institute of Japan. In addition, although the maximum carck spacing of recycled coarse aggregate concrete tends to be small compared with those of ordinary concrete, no significant difference is recognized due to the use of the oil-type agent. Although the recycled coarse aggregate produces larger deflection in RC beam, the extent is not so significant for the practical use and the low water/cement ratio may overcome the increase of deflection. Consequently, in this experiment, as significant degradation of crack behavior due to a surface improving treatment is not recognized, significant attention is not needed to durability related cracking

200 600 200 200 600 200

2000

D6@100(SD345)

D13(SD345)

20030 140 30

230

30

170

30

Displacement gaugePin・Roller bearing

Testing bench

Test piece

Load cell

Loading plate

Pressing head

【Cross-section】

Figure 11. Schematic diagram of loading in the RC bending test (Unit : mm)

6 CMT - 42


CMT -

Figure 17. Toughness ratio

behavior. Recycled concrete containing the oil-type agent may be applicable to a structural use in combination with a surface finishing material which can restrain water penetration and follow the movement of crack. The use of the oil-type agent providing recovery mechanism due to peeling-off effect is considered to be valuable in practical use.

0

1

2

3

M-60

L-60

M-40

L-40

stone

M-N

-60

M-O

-60

L-N

-60

L-O

-60

M-N

-40

M-O

-40

L-N

-40

L-O

-40

No treatment

OilCalculated values

0

5

10

15

Bendin

gm

om

ent

cau

sing

cra

cks

（kN

・m

）

0.0

0.1

0.2

0.3

Max

imum

cra

ckw

idth

（m

m）

0

100

200

300

Max

imum

cra

ck

inte

rval

（m

m）

Deflection

（m

m）

M-60

L-60

M-40

L-40

M-60

L-60

M-40

L-40

Cru

shed

-60

stone

Cru

shed

-40

stone

Cru

shed

-60

stone

Cru

shed

-40

stone

Cru

shed

-60

stone

Cru

shed

-40

stone

Cru

shed

-60

stone

Cru

shed

-40

No treatment

OilNo treatment

Oil

No treatment

Oil

4.6.4 Plastic behavior Figure 16 shows load-deflection curves of RC beams. All specimens collapsed due to the failure of concrete at the ultimate compression fiber. No significant difference in yielding moment and ultimate moment is recognized between recycled coarse aggrecate concrete and ordinary concrete. Load-deflection curves are similar to those obtained by Mukai9) and Sato10) as well as to that of ordinary concrete and no influence of the oil-type agent is recognized in all specimens. In addition, in this study, the observed concrete strains at ultimate conpression fiber were constant, nearly 3500 μ, for all specimens even though the oil-type agent was used. Accordingly, a beam theory can be applied up to an ultimate point for all specimens. The ultimate strength might be predicted by assuming that the concrete strain at ultimate compression fiber is 3500μ. The ductility factor, an ultimate deflection divided by an yield deflection, as shown in Figure 17, is determined as a deformation property after yielding. The ductility factor of RC beam containing recycled aggregate treated with an oil-type agent is almost the same as that of RC beam with untreated recycled aggregate and a significant difference is not observed in the ductility comparing with that of ordinary RC beam. In addition, no serious problem would occur and plastic deformation ability would be enough, because the ductility factor is almost 5. In this experiment, many cracks occurred in the range from yielding point to ultimate point in the concrete containing aggregate treated with surface improving agent. The cracking after yielding point might be effective from the viewpoint of aggregate recovery at the demolision of structure. The high recovery rate of aggregate and the high resistant property against shear should be discussed based on the future study. Reinfoced concrete beams containg recycled aggregate with the oil-type agent has a comparable propeties in both the strength and the resistance to cracking with those containing ordinary aggregate and threfore the oil-type agent has no problems in practical use. It could be concluded that flexural capacity of reinforced concrete whose aggregate is modified by the oil type agent will not become an issue for practical application of determinate structure under the condition that anchorage performance is sufficient and the tensile reinforcing bars yield before failure of concrete in compression.

Figure 12. Cracking moment Figure 13. Maximum Figure 14. Maximum Figure 15. Deflection crack width crack spacing

0

1

2

3

4

5

6

7

Tough

ness

ratio

（δ

u/δ

y）

M-60

L-60

M-40

L-40

No treatment

Oil

stone

Cru

shed

-60

stone

Cru

shed

-40

6 CMT - 43


CMT -

0

20

40

60

80

100

120

Load

(kN

)

0 5 10 15 20 25 0 5 10 15 20 25

Ｍ-Ｎ-60

Ｍ-Ｏ-60

Crushed stone

0 5 10 15 20 25

Deflexion (mm)

0 5 10 15 20 25 30

-60Ｌ-Ｎ-60

Ｌ-Ｏ-60

Crushed stone-60

Ｍ-Ｎ-40

Ｍ-Ｏ-40

Crushed stone-40Ｌ-Ｎ-40

Ｌ-Ｏ-40

Crushed stone-40

30

0

20

40

60

80

100

120

Load

(kN

)

Deflexion (mm) Figure 16. Load-deflection curves

5. CONCLUSIONS

Following concluding remarks were obtained through the experiments. (1) A surface improving agent reduced water absorption of low- and middle-quality recycled

aggregate. (2) As the oil-type agent does not decrease the strength of concrete significantly, it can be used for the

surface improvement of recycled aggregate in structural concrete. By contrast, as a silane-type aggregate decreased the strength significantly, how to use a silane-type agent should be discussed by investigating the relationship between the amount of application and the strength.

(3) As the peeling-off effect of aggregate treated with a silane-type agent is extremely higher than that of untreated aggregate, a silane-type agent is very effective from the view point of recovery of aggregate. The peeling-off effect of the oil-type agent is slightly recognized and further improvement can be expected by investigating the relation between the amount of application and the peeling-off effect.

(4) The oil-type agent possibly reduces drying shrinkage of recycled aggregate concrete with W/C of 60% to that of ordinary concrete.

(5) The creep deformation of recycled aggregate concrete treated with the oil-type agent is not significantly different from that of concrete with untreated aggregate.

(6) Although the recycled aggregate treated with the oil-type agent reduces the resistance to bending cracking, the crack width never overcomes 0.3 mm which is a threshold limit value from the viewpoint of durability, which proves that recycled aggregate concrete treated with the oil-type agent may be applicable to a structural use in combination with a surface finishing material which can restrain water penetration and follow the movement of crack.

(7) The bending strength of RC beam made using recycled aggregate treated with the oil-type agent is nearly equal to that of ordinary concrete and the load at which crack occurs might be estimated. Ultimate strength also can be estimated assuming that the concrete strain at ultimate compression fiber is 3500μ.

(8) Flexural capacity of reinforced concrete whose aggregate is modified by the oil-type agent will not become an issue for practical application to determinate structure under the condition that anchorage performance is sufficient and the tensile reinforcing bars yield before failure of concrete in compression.

6 CMT - 44


CMT -

6. ACKNOWLEDGEMENT

This study was supported by the scientific research grant on the waste disposal sponsored by the Ministry of Environment for fiscal years of 2004-2005, “The development of the next-generation recycling technology of demolished concrete” (Research representative is Dr. Takafumi NOGUCHI). A grateful appreciation is given for those who may concern.

7. REFERENCES

H. Shima et al. (2001). “LCA evaluation of high quality aggregate collection from concrete mass by Heat-Treatment and Rubbing Method.” Proceedings of the Japan Concrete Institute, Vol.23, No.2, pp.67-72.

A proposal toward the spread of concrete recycling systems (Activity reports about advanced use of recycle concrete) (2005). Japan Concrete Institute, 2005.9

M. Tamura et al. (1997). “Recycle-oriented concrete with easy-to-collect aggregate.” Cement Science and Concrete Technology, No.51, pp.494-499.

D. Tsuji et al. (2002). “Study on application of improved low-quality recycled aggregate for concrete structure.” Proceedings of the Japan Concrete Institute, Vol.24, No.1, pp.1251-1256.

Wang Cheng Hung (2003). “The study of the surface improvement of the low quality recycled aggregate for structure concrete.” The University of Tokyo master's thesis.

M. Hasegawa (1999), “Aqueous silicon analogue coating agent "Silas".” TOAGOSEI study annual report TREND1999, No.2, pp.45-49

M. Tamura et al. (2002). “Fundamental properties of recyclable concrete with easy-to-collect aggregate.” Proceedings of the Japan Concrete Institute, Vol.24, No.1, pp.1353-1358.

Recommendations for Practice of Crack Control in Reinforced Concrete Buildings (Design and Construction) (2006), Architectural Institute of Japan.

T. Mukai et al. (1979). “Fundamental examination about use to a structure member of Recycled aggregate concrete.” Annual Meeting of Japan Cement Association, NO33, PP.208-211.

R. Sato et al. (1999). “Mechanical properties of reinforced concrete members made of low quality recycled coarse and fine aggregates containing large amount of mortar or cement paste.” Cement Science and Concrete Technology, NO.53, PP.573-580.

6 CMT - 45


CMT -

APPLICATION OF RECYCLED AGGREGATE TO PRECAST PLAIN CONCRETE

Norihiro Mihara1, Fumio Taguchi2, Yasuhiko Sato3 and Katsuyuki Konno4

ABSTRACT : It is necessary to use the recycled aggregate produced from concrete structures for preserving natural aggregate. Most of recycled aggregate are currently used as roadbed materials, but road structure is coming not to be constructed in Japan. Concrete using recycled aggregate, however, has problems in terms of its strength and durability. In cold, snowy regions like Hokkaido in Japan, in particular, deterioration of concrete caused by repeated freezing and thawing has been a problem. In this study, compressive strength, freeze-thaw and other tests were conducted for plain concrete without reinforcing bar to use recycled aggregate. As a result of these tests, it was found that freeze-thaw durability could be improved by mixing recycled and ordinary aggregate. It was also established that freeze-thaw durability could be evaluated by controlling the absorption and water cement ratio by laboratory test. The Quality Standards for Recycled Coarse Aggregate Used in Precast Plain Concrete (draft) were prepared based on these results, to provide regulations on the absorption and other specifications of recycled coarse aggregate. To verify the durability in actual construction, a factory product was used for repairs on a national highway’s sidewalk as part of a test field project in 2002. As a result of continuous research on the performance of the on-site specimens, it was confirmed that they had sufficient freeze-thaw durability. As the result that actual constructed curbstones were examined when three years passed, it is confirmed there is little difference between concrete using recycled aggregates and that using ordinary aggregate.

KEYWORDS: recycled aggregate, absorption, freeze thaw durability, drying shrinkage,

1. INTRODUCTION

While approximately 90% of recycled materials produced from demolished concrete are currently used as roadbed materials, it is necessary to expand the use of the recycled aggregate produced from concrete structures since the construction of new roads is decreasing. The amount of demolished concrete is expected to further increase in the future as many concrete structures constructed during the period of high economic growth in Japan will reach the end of their design lifetime, and it is also desirable to facilitate the use of recycled aggregate from the viewpoint of preserving natural aggregate. Concrete using recycled aggregate, however, has problems in terms of its strength and durability. In cold, snowy regions like Hokkaido, Japan, in particular, deterioration of concrete caused by repeated freezing and thawing has been a problem.

1 Materials Research Team, Civil Engineering Research Institute for Cold Region, Japan 2 Materials Research Team Leader, Civil Engineering Research Institute for Cold Region, Japan 3 Division of Built Environment, Hokkaido University, Japan 4 Division of Civil Engineering, Hokkaido Institute of Technology, Japan

7 CMT - 46


CMT -

Table 1. Experimental items Experimental

itemExperim- ental factor

Freeze- thaw test

Length variation

test Compressive strength test

Absorption of coarse

aggregate ○ ○ ○

Water-cement ratio

(35,45,55％) ○ ○ ○

Test method JSCE-G501-1999

JISA1129-1993

JISA1108-1993

In this study, compressive strength, freeze-thaw and other tests were conducted for plain concrete without reinforcing bars for the purpose of expanding the use of recycled aggregate. To verify its durability in actual construction, a secondary product of plain concrete was used for repairs on a national highway’s sidewalk as part of a test field project in 2002. This paper presents the results of laboratory tests of concrete using recycled aggregate and continuous research on the performance of the on-site samples.

2. SUMMARY OF THE EXPERIMENT

2.1. Experimental items and methods The experimental items are as listed in Table 1. Properties of concrete using recycled aggregate are affected greatly by the quality of recycled aggregate, especially the absorption. Concrete samples were therefore prepared using recycled aggregate, whose absorption was changed within the range of 2.5 to 2.6% by mixing ordinary aggregate, to conduct freeze-thaw, drying shrinkage and compressive strength tests. The same tests were also conducted for concrete using highly absorbent recycled coarse aggregate, by changing its water-cement ratio to 35, 45 and 55%. Samples used for the freeze-thaw and drying shrinkage tests were 10×10×40 cm in size, and the material age at the beginning of those tests was 28 and 7 days, respectively. In the freeze-thaw test, one cycle was comprised of the process of completely freezing water-bearing concrete samples in the water by changing the temperature from +5 to –18℃, and thawing them by raising the temperature back to +5℃ again. This cycle, which took 3 to 4 hours, was repeated up to 300 times. In the test, the relative dynamic modulus of elasticity and rate of mass decrease were measured for every set of 30 cycles to determine the freeze-thaw durability. Concerning drying shrinkage, concrete using a shrinkage reducing admixture was prepared to verify its effect on recycled aggregate. The drying shrinkage test was conducted under the environmental conditions of 20℃in temperature and 60% in relative humidity, and the changes in the length of the samples were measured by the dial gauge method. The size of samples for the compressive strength test wasφ15×30 cm, and the test was conducted at the material age of 7, 28 and 91 days.

Table 2. Physical properties of coarse aggregate

Coarse aggregate Maximum size (mm)

Specific gravity in saturated surface-dry

condition Absorption

(％) Loss in stability test (％)

Mix ratio (Recycled: Ordinary)

Ordinary 2.69 1.8 1.0 0：10 Recycled A 2.38 6.7 32.5 10：0 Mixed A1 － 2.7 10.2 3：7 Mixed A2 － 3.8 16.6 5：5

Recycled B 2.45 6.1 51.0 10：0 Mixed B1 2.57 4.0 35.3 5：5 Mixed B2 2.52 4.8 36.2 7：3

Recycled C

40

2.44 6.0 51.1 10：0

7 CMT - 47


CMT -

2.2. Materials used Test samples were prepared using ordinary Portland cement and sea sand produced in Tomakomai Japan as fine aggregate. Table 2 displays the physical properties of the coarse aggregate that was used. The absorption here is that obtained after mixing recycled and ordinary aggregate. As ordinary coarse aggregate, crushed stone produced in Otaru Japan was used. Recycled coarse aggregate was obtained from different recycled aggregate manufacturing plants around Sapporo, Hokkaido, Japan, and the absorption was improved by mixing it with ordinary coarse aggregate in proportions shown in the table. The above recycled coarse aggregate was manufactured mainly using jaw crushers, while a combination of cone and impact crushers was used in few cases. An AE water reducing agent mainly containing lignin sulfonic acid-polyol complex and an air-entraining agent mainly containing a modified rosin acid compound were used as admixtures. A shrinkage reducing admixture mainly containing an alkylene oxide adduct, which is a lower alcohol, was also used as a measure against drying shrinkage. 2.3. Mix proportion of concrete Table 3 lists the typical mix proportions of concrete. The target slump was 8±1 cm and the target air content was 4.5±1％. Concrete was mixed using a pan-type forced mixer with a capacity of 100 liters and the mixing time was 3 minutes after all the materials were loaded. The mixing and curing temperature of the concrete was 20℃. Moisture curing was conducted for a day from the time the concrete was placed until it was deformed. This was followed by water curing conducted until the test material age was reached.

3. EXPERIMENT RESULTS AND DISCUSSION

3.1. Freeze-thaw durability

3.1.1. Effect of the absorption

Figures 1 and 2 illustrate the relationship of the relative dynamic modulus of elasticity to the rate of mass decrease and absorption, respectively, in the 300-cycle freeze-thaw test conducted for concrete with a water-cement ratio of 45%. The relative dynamic modulus of elasticity was closely related to the absorption of coarse aggregate, and it varied greatly when the absorption exceeded 4.5%. The relative dynamic modulus of elasticity required for concrete of “general structural members used in an environment without contact with water” in the Standard Specifications for Concrete is 60% or higher, while 70% or higher is required in an “environment saturated with water.” In this experiment, 80% or higher relative dynamic modulus

Table 3. Mix proportions of concrete Unit volume（kg/m3） Water-cement

ratio (％)

s/a (％) Water Cement Fine

aggregate Coarse

aggregate 35 40 140 400 748 1,004

45 40 132 293 794 1,064

55 44 134 244 889 1,014

7 CMT - 48


CMT -

0 60 120 180 240 300

サイクル（回）

50

60

70

80

90

100

相対動弾性係数（%）

W /C=45%（普通骨材）

W /C =35%

W/C=45%

W/C=55%

Rela

tive

dyna

mic

mod

ulus

of

ela

stic

ity (%

)

Cycle (no.)

Ordinary aggregate

Figure 3.W/C and relative dynamic modulus

of elasticity (absorption: 6.0％)

40

50

60

70

80

90

100

2 3 4 5 6 7吸水率（％）

相対動弾性係

数（％）

Rel

ativ

e dy

nam

ic m

odul

us

of e

last

icity

(%

)

Absorption (%)

Figure 1. Absorption and relative dynamic modulusof elasticity (W/C＝45％)

0

0.5

1

1.5

2

2.5

3

3.5

2 3 4 5 6 7吸水率（％）

質量

減少率

（％）

Rat

e of

mas

s dec

reas

e (%

) Absorption (%)

Figure 2. Absorption and rate of mass decrease (W/C＝45％)

of elasticity was achieved by improving the absorption of recycled coarse aggregate with an absorption of around 6%, which was used for this experiment, to 4.5% or lower by the method of mixing ordinary aggregate. Regarding the rate of mass decrease, variations in the data tended to increase when the absorption exceeded 4.5%, in the same way as in the case of the relative dynamic modulus of elasticity.

3.1.2. Effect of the water-cement ratio Figure 3 displays the changes in relative dynamic modulus of elasticity in the freeze-thaw test. The absorption rate of recycled aggregate was 6.0%. The modulus decreased with an increase in the water-cement ratio. In the case where the water-cement ratio was extremely low (35%), however, the relative dynamic modulus of elasticity was almost the same as in the case of using ordinary aggregate. It was thus considered possible to produce concrete using recycled aggregate with satisfactory frost-damage durability by setting an appropriate water-cement ratio.

0100200300400500600700800900

0 100 200 300 400経過日数（日）

乾燥

収縮ひず

み（μ）

4.0% 4.8%

6.1% 1.8%（普通骨材）

4.8%（収縮低減剤）

Ordinary aggregate

Dry

ing

shrin

kage

stra

in (μ

)

No. of days elapsed (days)

Shrinkage reduction admixture

Figure 4. Absorption and drying shrinkage (W/C＝45％)

7 CMT - 49


CMT -

10

15

20

25

30

35

40

45

50

55

0 20 40 60 80 100

圧縮

強度

(N/m

m2 )

1.8%（普通骨材）

2.7%4.0%4.8%6.0%

Ordinary aggregate

Com

pres

sive

stre

ngth

(N/m

m2 )

Test material age (days)

Figure 6. Absorption and compressive strength (W/C＝45％)

3.2. Drying shrinkage

Figure 4 displays the relationship between drying shrinkage strain in concrete with a water-cement ratio of 45%. The result was compared with the test results of concrete using ordinary aggregate and a shrinkage reducing admixture to control drying shrinkage strain. In the case where recycled coarse aggregate with an absorption rate of 6.1% was used, the amount of strain after one year was more than 1.5 times as high as that of ordinary aggregate. It was decreased to around 1.3 times by reducing the absorption to 4%, indicating the quality improvement effect of mixing with ordinary aggregate. While the amount of strain was almost the same as that of ordinary aggregate until around the material age of 90 days in the case where a shrinkage reducing admixture was added, shrinkage continued after that and the amount of strain became almost the same as that of concrete without an admixture at the material age of approximately 200 days. The effect of the shrinkage reducing admixture did not last over a long period of time. 3.3. Dynamic properties

Figure 5 displays the relationship between the compressive strength and water-cement ratio. It was compared with the test result of concrete using ordinary aggregate with a water-cement ratio of 45%. Since the absorption rate of the recycled coarse aggregate that was used was as high as 6.7%, its strength was much lower than that of concrete using ordinary aggregate even when the water-cement ratio was 35%. However, the compressive strength tended to decrease with an increase in the water-cement ratio, in the same way as ordinary concrete.

Figure 6 presents the relationship between the compressive strength and absorption in concrete with a water-cement ratio of 45%. There was a correlation between the compressive strength and absorption. When aggregate with a lower absorption was used, the compressive strength of the concrete became greater. The increase in strength from 7 to 28 days in material age slowed slightly when the absorption was high, although no significant difference was found in the increase from 28 to 91 days. The same level of strength as that of concrete using ordinary aggregate was obtained when using recycled aggregate with an absorption of 2.7%, whose quality was improved to the level of Type 1, as presented in the “Provisional Quality Standards for Recycling of Concrete

10

15

2025

30

35

40

45

50

55

0 20 40 60 80 100試験材齢（日）

圧縮強

度(N

/mm

2 )

W/C=35％W/C=45％W/C=55％W/C=45% 普通骨材

Ordinary aggregate Co

mpr

essi

ve st

reng

th (N

/mm

2 )

Test material age (days)

Figure 5. W/C and Compressive strength (absorption: 6.7％)

7 CMT - 50


CMT -

Byproducts According to Application (draft).” 5)

4. REFLECTION IN THE STANDARDS

The above result was reflected in the “Quality Standards for Recycled Coarse Aggregate Used in Precast Plain Concrete (draft) – April 2002” (“Standards (draft)”). The absorption of recycled aggregate was set at 3% in consideration of the fact that whether or not the base concrete was AE had not been determined.

5. TEST CONSTRUCTION

5.1. Summary of construction Secondary products of plain concrete using recycled aggregate were used for repairs on the sidewalk of National Route 12 in Iwamizawa, as part of a “test field project.” This section presents an evaluation of their performance as products, on-site construction conditions and the results of continuous research.

The types and quantities of the secondary concrete products used for this test construction are as described below. The shape of curbs for planting zones was the same as that of ordinary curbs.

① Types of products … curbs for boundaries between sidewalks and roadways (Type I), curbs for planting zones, ordinary curbs

② Construction length … boundaries between sidewalks and roadways: 12 m×2 sections, planting zones: 9 m×2 sections, pavement: 9 m×2 sections

5.2. Types of aggregate and their physical properties Secondary concrete products for testing were prepared using two types of recycled coarse aggregate (Recycled A and B), whose absorption rates were adjusted to be 3% or lower in accordance with the Standards (draft) taking the frost damage durability into consideration. In order to make a comparison, the same tests were conducted for a product using ordinary aggregate. Table 4 lists the physical properties of aggregate.

Table 4. Types and physical properties of aggregate

Mark Coarse aggregate used Absorption Specific gravity in absolute dry condition Remarks

Recycled A Recycled aggregate 100％ 2.27％ 2.60 High-quality recycled coarse

aggregate

Recycled B Recycled : ordinary ＝50 : 50 2.40％ 2.59 Absorption of recycled coarse

aggregate only: 3.46％

Ordinary Ordinary aggregate 100％ 2.34％ 2.54

5.3. Mix proportions of concrete

Table 5 presents mix proportions of concrete. The target slump of the concrete was 8±2.5 cm and target air content was 5.0±1.5%. Steam curing was conducted from 2 hours after placement. The temperature was raised from 15℃ to 65℃in 2.5 hours, maintained at that level for another 2 hours and reduced to 15℃ again over the following 12 hours. After that,

7 CMT - 51


CMT -

tests were conducted by storing the products in the air, samples for the freeze-thaw test in the water and samples for the compressive strength test at a constant temperature and high humidity condition (temperature: 20℃, humidity: 60%).

Table 5. Mix proportions of concrete

W/C (％)

Cement (kg/m3)

Water (kg/m3)

Recycled coarse

aggregate (kg/m3)

Ordinary coarse

aggregate (kg/m3)

Fine aggregate (kg/m3)

Slump (cm)

Air content

(％)

Recycled A 50 310 155 1056 0 804 10.5 5.5 Recycled B 〃〃〃 528 528 〃 8.5 4.6 Ordinary 〃〃〃 0 1028 〃 10.0 4.5

5.4. Test items and methods The freeze-thaw test is a very severe test, in which rapid freezing and thawing is repeated. One cycle consists of the process of completely freezing water-bearing concrete samples in the water by lowering the temperature from +5 to –18℃ and then thawing them by raising the temperature to +5℃ again. The cycle, which took 3 to 4 hours, was repeated up to 300 times. In the test, the relative dynamic modulus of elasticity and the rate of mass decrease were measured for every set of 30 cycles to study the freeze-thaw durability.

Table 6. Test items Test item Content Test method Remarks

Compressive strength test

Measurement of changes in compressive strength by material age usingφ10 cm samples and cores collected from products

JIS A 1108 JIS A 1107

・φ10cm (material age: 1, 7, 28, 90, 180, 360 days, constant-temperature/high humidity curing)

・Cores (1 and 3 years later, collected on site)

Freeze-thaw test

Test using 10×10×40 cm samples

JIS A 1148 ・300 cycles of freezing/thawing

On-site visual inspection

Observation of deterioration conditions from the appearance

Visual observation

・3 year after construction

6. TEST RESULTS AND DISCUSSION

6.1. Compressive strength Figure 9 illustrates the increase in compressive strength by material age. In the case of concrete using recycled aggregate, it exceeded 30 N/mm2

at the material age of 28 days, indicating the development of sufficient strength compared with concrete using ordinary aggregate. While Recycled B showed the same level of compressive strength as that of the case using ordinary aggregate, the strength of Recycled A was slightly lower than those of others.

6.2. Freeze-thaw durability

Figure 10 displays the results of the freeze-thaw test. While the values of Recycled A were slightly smaller than those of others, those of Recycled B were almost the same as those in the case using ordinary aggregate. Since the absorption of recycled coarse aggregate used for this

7 CMT - 52


CMT -

0

10

20

30

40

50

0 7 14 21 28

材齢（日）

圧縮強度(N/mm

２)

再生 A

再生 B

普通

Com

pres

sive

stre

ngth

(N/m

m2 )

Material age (days)

Recycled A

Recycled B

Ordinary

Figure 9. Result of the compressive strength test

60

70

80

90

100

0 100 200 300

凍結融解サイクル（回）

相対動

弾性係

数（

％）

再生A

再生B

普通

Rela

tive

dyna

mic

mod

ulus

of

elas

ticity

(%)

Freeze-thaw cycle (no.)

Recycled A

Recycled B

Ordinary

Figure 10. Result of the freeze-thaw

test

Photo 1. Appearance of the samples

test was 3% or lower in all cases, it was confirmed that sufficient frost damage durability was maintained even when the water-cement ratio was slightly higher (50%). The relative dynamic modulus of elasticity is required more than 60% for concrete of “general structural members used in an environment without contact with water” in the Standard Specifications for Concrete

Photo 1 shows the appearance of the samples after the freeze-thaw test. While slight scaling deterioration is observed on the surface of concrete using Recycled A, it is considered very minor.

6.3. Results of continuous research on on-site samples Photo 2 presents the concrete curbs using recycled and ordinary aggregate, which were constructed 3 years earlier in downtown Iwamizawa, Hokkaido, Japan. It can be seen that there was no difference in the deterioration of the curbs using recycled aggregate and those using ordinary aggregate. While scaling and popouts were found at the corners, they are also found in ordinary aggregate, and it was presumed that it was not caused by the effect of water contained in the mortar portions of recycled aggregate. It was probably due to the weakening of the structure caused by the absence of cement paste at the corners, since the construction method was different from usual. It is therefore considered that concrete using recycled aggregate has frost damage durability equivalent to that of ordinary concrete.

7 CMT - 53


CMT -

Recycled aggregate (mix proportion A)

Recycled aggregate (mix proportion B)

Ordinary aggregate

Photo 2. Curbs after three years

7. CONCLUSION

(1) By reducing the absorption rate of concrete using recycled aggregate by mixing it with ordinary aggregate, the relative dynamic modulus of elasticity and frost damage durability can be improved.

(2) By setting an appropriate water-cement ratio, it is possible to produce concrete using recycled aggregate with frost damage durability.

(3) Correlation was found between the compressive strength and absorption rate of recycled aggregate.

(4) After conducting a number of laboratory tests as those mentioned above, the Quality Standard for Recycled Coarse Aggregate Used in Precast Plain Concrete (draft) were prepared.

(5) Concrete using recycled aggregate, whose quality was improved by mixing 50% of ordinary aggregate, had the same level of compressive strength as that of ordinary aggregate. The product using 100% high-quality recycled coarse aggregate had sufficient strength compared with that using ordinary aggregate with the same mix proportion.

(6) As a result of the freeze-thaw test, it was confirmed that the required frost damage durability was satisfied in both concrete using high-quality recycled coarse aggregate and that using recycled coarse aggregate, whose quality was improved by mixing with ordinary aggregate.

(7) In terms of on-site construction conditions, deterioration of concrete using recycled aggregate was equivalent to that of concrete using ordinary aggregate even after 3 years, indicating sufficient frost damage durability of recycled aggregate.

8. AFTERWORD

Sufficient frost damage durability of precast plain concrete could be proved through laboratory tests and test construction. With the use of recycled aggregate in reinforced concrete, however, it is necessary to clarify the effect of the salt content of recycled aggregate on reinforcement corrosion. The authors are currently considering establishing a simple evaluation method and quality standards for evaluation of the water content of recycled aggregate.

7 CMT - 54


CMT -

9. REFERENCES

Shinichi Sasaki, Fumio Taguchi and Hisatoshi Shimada (2002). “Properties of Concrete Using Recycled Coarse Aggregate.” 45th Research Presentation of the Hokkaido Regional Development Bureau, Kyo-18.

Hokkaido Branch, Japan Concrete Institute (2002). “Report of the Research Committee on Recycling.”

Masashi Kobamatsu, Takumi Hoshi and Shinichi Sasaki (2003). “Test Construction of Precast Plain Concrete Products Using Recycled Aggregate.” 46th Research Presentation of the Hokkaido Regional Development Bureau.

7 CMT - 55


CMT -

CHARACTERISTICS OF SELF-COMPACTING CONCRETE IN FRESH STATE WITH NEW-TYPE ARTIFICIAL LIGHT–WEIGHT AGGREGATE

Koichi Kobayashi1

ABSTRACT : Recently, a new type of artificial light-weight aggregate has been developed. This aggregate has lower water absorption ratio than ordinary light-weight aggregate because of its tight surface structure, and can be used for concrete mixing without pre-wetting procedure. Another advantage of this aggregate is its spherical shape that is expected to increase the fluidity of concrete. In this study, this new light-weight aggregate was applied to self-compacting concrete to increase its deformability, and the self-campactability of this concrete was investigated and discussed. The results show that self-compacting concrete with this aggregate has higher self-compactability than that with crushed stone. This concrete can have high self-compactability even if its slump-flow is very large or very small. Segregation between the aggregate and mortar, however, tends to be large because of larger difference of specific gravity between them than in the case of ordinary self-compacting concrete with crushed stone. Increase of unit mass of the light-weight aggregate does not affect so much on self-compactability of concrete. A drop of self-compactability of self-compacting concrete with the light-weight aggregate with the lapse of time after mixing is less than that with crushed stone.

KEYWORDS: Artificial light-weight aggregate, Self-compacting concrete, Self-compactability, Slump flow,

1. INTRODUCTION

Conventional light-weight aggregate has a very high water absorption rate due to its continuous voids, and it needs undergoing a pre-wetting process so as to avoid slump loss and to keep its high pumpability. However, this eventually results in a high water cement ratio because the water in the aggregate is released into cement paste after mixing, which decreases the strength and durability of hardened concrete. Therefore, despite its great advantage that it can decrease dead load of concrete structure, the amount of production of light-weight aggregate concrete is only 800,000m3 per year in Japan (Okamoto, 1998). Under these circumstances, a new type of light-weight aggregate has been developed, using granulated and fired perlite (Okamoto, 1999). While this new aggregate contains voids formed by foaming during the firing, it has a low water absorption rate because of its tight surface. Therefore, the pre-wetting process is not necessary for this aggregate. Another characteristic of this aggregate is its spherical shape (Figure 1). It is expected that the use of this aggregate will decrease mutual interference and interlocking of aggregates in fresh concrete and increase self-compactability of self-compacting concrete. In this study, the self-compactability of self compacting concrete using this new light weight aggregate 1 Associate Professor, Dr. Eng, Chubu University, Japan

8 CMT - 56


CMT -

was investigated, with the aim to clarify the applicability of this aggregate as an material for self compacting concrete.

2. SUMMARY OF EXPERIMENT

Table 1 shows materials employed in this study. In addition to ordinary Portland cement, fly ash was used as binder to improve the fluidity of concrete. Another advantage of using fly ash is its small specific gravity, which decreases the specific gravity difference between mortar and light-weight aggregate, and reduces segregation of concrete.

Table 1. Materials

Self-compacting concrete with ordinary crushed stone, which has the same maximum size as that of light-weight aggregate, was also prepared for comparison with the light-weight self-compacting concrete. The light-weight aggregate was used in an absolutely dry condition, while crushed stone was used in a surface dry condition. Therefore, the unit mass of light-weight aggregate in tables of mixture design in this study is the mass in the absolutely dry condition. The mixing procedure is shown in Figure 2. Concrete was mixed in the laboratory under room temperature.

Figure 3. Box test apparatus for passability test

Figure 1. Light-weight aggregate

8 CMT - 57


CMT -

Figure 2. Mixing procedure

Slump flow and the flow time until 500mm were investigated. Passability of concrete through spaces was investigated according to JSCE-F511-1999 (JSCE, 1999) with a box-shaped apparatus (Figure 3). After the passability test, after three minutes after the flow of concrete stopped, 1.3 little of concrete was sampled from the top part of concrete in both room A and room B, the former being the throw-in side of the box apparatus and the latter being the flow-out side. The concrete samples underwent wet screening with a 5mm sieve to take out only coarse aggregates, which were then surface-dried and weighed to determine the coarse aggregate density in concrete. Mortar having the same mixture design as the mortar part in concrete was prepared using a mortar mixer and the yield value and viscosity of the mortar were measured using a rotational viscometer.


3.1 Mixture design Table 2 shows the mixture design of trial concrete mixtures with light-weight aggregate (about 20). Figure 4 shows the relationship between slump flow and fill height obtained through a passability test. Apart from the mixture with a water mass of 144kg/m3 and a slump flow of 530mm, all the mixtures had a fill height of more than 300mm. It is therefore concluded, at least from the results of the test using an ordinary passability test apparatus, that almost all mixtures have good self-compactability.

Table 2. Trial mixtures

Figure 4. Relationship between slump flow and fill height in box apparatus

Some concrete mixtures had a much higher slump flow (over 800mm) than the usual 600 to 700mm, which is an indication of good self-compactability. These mixtures showed signs of segregation between mortar and coarse aggregate at the edge of slump flow. On the other hand, some concrete

8 CMT - 58


CMT -

mixtures with a slump flow of under 600mm had a fill height of more than 300mm. This high self-compactability is probably attributed to the round shape of coarse aggregates, which decreases mutual interlocking among the aggregates when the concrete flows through rebar obstacles. It should be noted, however, that the maximum size of the light-weight aggregate used in this study was 15mm, which is smaller than the ordinary 20 to 25mm, and that further investigation may be necessary on the effect of the maximum size of coarse aggregate 3.2 Coarse aggregate density after the passability test Ordinary self-compacting concrete would have a low fill height and low self-compactability, because segregation causes settlement of coarse aggregate in room A (see Figure 3) in the passability test box, resulting in interlocking of aggregate at the obstacles. On the other hand, self-compacting concrete with light-weight aggregate may have a large fill height because the coarse aggregate lifts up in room A due to segregation, allowing the mortar rich concrete to flow easily through the obstacles. Therefore, the segregation and self-compactability of concrete cannot be precisely determined only on the basis of the fill height. For this reason, the coarse aggregate densities of concrete in room A and room B were investigated after the passability test as mentioned in Section 2. Figure 5 shows the relationships between the volumetric density of coarse aggregate taken out from the box apparatus for the passability test and the slump flow and the slump flow time until 500mm. Data points that plot the data of the same concrete are connected with lines. Average densities are theoretical values calculated from the mixture design.

Figure 5. Relationship between aggregate density and slump flow or flow time until 500mm

Even though the fill height is more than 300 mm, the concrete with light-weight aggregate is prone to segregation and lift-up of aggregate because of its small specific gravity and therefore the coarse aggregate density was more than 50% in some concrete samples taken out from the box apparatus in comparison to the average density of 30% or more. Generally, it is considered that self-compacting concrete should have a slump flow of less than 700mm

8 CMT - 59


CMT -

so as to ensure that there is no segregation and to achieve good self-compactability. The passability test results are consistent with this understanding and the coarse aggregate density of the self-compacting concrete samples with light-weight aggregate was very high when the slump flow was more than 700mm. On the other hand, when the slump flow was less than 700mm, while the aggregate density in room A of the box apparatus was larger than the theoretical value by 3~8 %, the aggregate density in room B was very close to the theoretical value. Therefore, the slump flow of 700mm can be regarded as a threshold value for avoiding segregation. It is reported that, with the use of artificial light-weight aggregate having a specific gravity of 0.94, and with the same water cement ratio, the coarse aggregate density ranging from 0.27 to 0.4m3/m3 does not affect the strength and Young’s modulus of self-compacting concrete. Therefore, it is considered that the coarse aggregate density has no bearing on the mechanical properties of the self-compacting concrete with a slump flow of less than 700mm. Table 3 shows a standard mixture of concrete designed such that it has good self-compactability with little segregation based on the above test results.

Table 3. Standard mixture of concrete with light-weight aggregate

The slump flow of all the concretes used hereafter was less than 700mm. The coarse aggregate densities of the samples taken out from room A and room B differed from the theoretical values only by less than 7% and 5%, respectively. 3.3 Comparison with concrete with crushed stone Figure 6 shows the effect of unit mass of water on the properties of crushed stone self-compacting concrete having the same unit mass of cement, fly ash, chemical admixture, and unit volume of coarse aggregate as the light-weight self-compacting concrete mixture shown in Table 3. As shown in this figure, the crushed stone concrete can achieve the same level of flowability and self-compactability as the light-weight concrete by increasing the unit mass of water by 1 or 2 kg/m3. The higher flowability of concrete with a smaller unit mass of water is attributable to less interaction and interlocking of coarse aggregate due to the spherical shape of light-weight aggregate.

Figure 6. Effects of coarse aggregate type on the fresh concrete properties

8 CMT - 60


CMT -

Figure 7 shows the relationship between slump flow time until 500 mm and plastic viscosity of mortar. As mentioned above, the mortar was mixed aside from the concrete. Since the slump flow range is very large, the 500mm slump flow time was corrected, being multiplied by the square of the quotient of slump flow divided by 500. As shown in this figure, there is no clear relationship between the plastic viscosity of mortar and the flow rate of concrete. However, the flow rate of light-weight concrete, which is closely related to the apparent plastic viscosity of concrete, is smaller than that of crushed stone concrete, even though the plastic viscosities of mortar in both crashed stone concrete and light-weight concrete are more or less the same. This is attributable to the small specific gravity of light-weight concrete, i.e., self-compacting concrete flows by the shear stress caused by its own weight, and therefore, the small weight resulted in the small flow rate.

Figure 7. Relationship between 500mm flow time of concrete and plastic viscosity of mortar

3.4 Effect of unit mass of coarse aggregate As demonstrated above, the shape and the specific gravity of coarse aggregate seem to have significant bearing on the flowability and the self-compactability of self-compacting concrete. In designing a mixture of self-compacting concrete with crushed stone, it is generally understood that the volumetric density of coarse aggregate must be less than half of the solid content volume of coarse aggregate so as to avoid interlocking at the obstacles in the passability apparatus (Okamura, 1993). Therefore, we investigated the effects of coarse aggregate density on the properties of light-weight self-compacting concrete. The concrete mixture used in this investigation (Table 4) is based on the standard mixture of Table 3, with the unit mass of light-weight coarse aggregate being increased from 354kg/m3, to 424kg/m3 and the unit mass of sand being decreased. According to this mixture design, the volumetric density of coarse aggregate in concrete will range from 30.8 to 36.9%. Figure 8 shows the effects of unit mass of light-weight coarse aggregate on the properties of fresh concrete.

Table 4. Mixture used in the investigation of the effect of unit mass of light-weight aggregate

Fa*: Fly ash

8 CMT - 61


CMT -

While the 500mm slump flow time was increased in the range of 25 to 40 sec. due to the increase of the unit mass of coarse aggregate, there were no changes in slump flow and fill height. The highest coarse aggregate density here was 36.9%, which is far more than the half of solid volume content of coarse aggregate (67.0%), and nevertheless all the mixtures had good self-compactability. It is reported that mutual interference among coarse aggregates can be decreased by increasing plastic viscosity of moratar. However, it is clear from Figure 7 that the plastic viscosity of mortar was almost the same in crushed stone concrete and light-weight aggregate concrete. Therefore, the round shape of coarse aggregate is considered to have reduced interference among the aggregates during the flow and resulted in the good self-compactability.

Figure 8. Effects of unit mass of light-weight aggregate on the characteristics of fresh concrete

3.5 Slump flow loss and fill height loss One of the advantages of the light-weight aggregate used in this study over ordinary light weight aggregate is that it can be used without a pre-wetting process due to its low water absorption rate. In actual application, this light-weight aggregate will most likely be used in air-dried condition, in view of the difficulty in providing a separate processing site in concrete plants exclusively for the light-weight aggregate. A possible risk is that water absorption in the aggregate after mixing may cause slump flow loss. Therefore, we investigated changes with time in the slump flow, 500mm flow time, and fill height after mixing, using the light-weight self-compacting concrete with the mixture design of Table 3 and the crushed stone self-compacting concrete with a unit weight of water of 150kg/m3. The results are shown in Figure 9.

Figure 9. Effects of time after mixing on the characteristics of fresh concrete

8 CMT - 62


CMT -

The slump flow decreased with time in both concrete mixtures and there was no difference between the two types of aggregates. On the other hand, the slump flow time of light-weight aggregate concrete was larger than that of crushed stone concrete, and the difference between the two types of concrete increased with time after mixing. The flow time until 500mm of the light-weigh aggregate mixture was more than 60 sec after 90 minutes have passed, and yet its fill height remained more than 300 mm until after 120 min. after mixing. It is expected that water absorption in light-weight aggregate may adversely affect the fresh concrete performance as mentioned before, and actually, the slump flow rate was decreased more in the concrete mixture with the light-weight aggregate that was in an absolutely dry condition. However, the self-compactability, which is the most important property of self-compacting concrete, did not deteriorate with time after mixing. It is generally considered that self-compacting concrete should have a high deformation rate so as to avoid interlocking among coarse aggregates. On the contrary, the self-compacting concrete can maintain its high self-compactability even if the deformation rate is decreased, because the spherical shape of the aggregate prevents interlocking. It should be noted, however, that concrete using the light-weight aggregate having such a low flow rate will have poor pumpability and construction efficiency, and, the workability of this concrete on site should further be investigated.

4. CONCLUSIONS

In this study, the properties of fresh self-compacting concrete with perlite light-weight aggregate were investigated. The significant characteristics of this aggregate are a low water absorption rate and the spherical shape. The characteristic features of the self-compacting concrete with this aggregate can be summarized as follows. 1. It has high self-compactability for a wide range of slump flow. 2. Segregation and lift up of aggregate may occur irrespective of good passability test results. 3. It has better self-compactability than crushed stone concrete if the unit mass of water is the same. 4. An increase of unit mass of coarse aggregate does not affect the self-compactability.

5. REFERENCES

Okamoto, T., Hayano, H., Shibata, T. (1998). “Super Light Weight Concrete.” JCI Concrete journal, Vol. 36, No. 1, Jan., pp. 48-52

Okamoto, T., Ishikawa, Y., Tochigi, T., Sasajima, M. (1998). “High Performance Lightweight Concrete.” JCI Concrete journal, Vol. 37, No. 4, April, pp. 12-18

Okamura, H., Maekawa,K., Ozawa, K. (1993). “High Performance Concrete.” Gihodo JSCE Research subcommittee on Recommendation of Self-Compacting Concrete. (1999).

“Recommendation of Self-Compacting Concrete.”

8 CMT - 63


CMT -

STUDY ON SELF COMPACTING CONCRETE CONTAINING MARBLE SLUDGE AS PARTIAL REPLACEMENT FOR ORDINARY PORTLAND CEMENT

Rita Irmawaty 1 and M.W. Tjaronge2

ABSTRACT: The creation of durable concrete structures requires adequate compaction by skilled workers but the numbers of them in the construction industry are not enough. This factor leads to reduction in the quality of construction work. One solution for the achievement of durable concrete structures independent on the quality of construction work is the employment of Self Compacting Concrete (SCC). SCC is a concrete that can be placed by pump or skip and flow under its own weight with maintaining its homogeneity to completely fill the mold of any shape, even with congested reinforcement. There are some marble mining and factories in South Sulawesi, Indonesia. Marble sludge is a disposal material from the marble production which must be disposed of in the landfills. The continuous disposing of sludge marble can destroy the environment and disturb the ecosystem around the disposing area. Recently, the natural conservation is a global issue and becomes one of the prominent aspects in the development program of many countries such as Indonesia. It is important to reduce the disposal of marble sludge in order to prevent the environmental damage. Marble sludge contains CaO of about 50-60% therefore it could be used as a replacement of cement. The incorporating of marble sludge in concrete production can reduce the disposal of marble sludge and also contribute to environmental protection. This paper is a preliminary study on the using of marble sludge as partial replacement for ordinary Portland cement to produce SCC. Replacements levels of cement by marble sludge were 0, 5, 10, 15 and 20 % by weight of cement. A water cement ratio of 0.34 was kept invariant. This research investigated its effect on the slump flow, compressive strength, flexural strength and splitting tensile of SCC. A combination of qualitative observation and quantitative measurements was applied to evaluate the slump flow of SCC. The measurement of slump flow indicates the free deformability of mixtures, which can also indicate segregation resistance of SCC. KEYWORDS : Self Compacting Concrete (SCC), marble sludge, slump flow, compressive strength, flexural strength , splitting tensile strength.

1. INTRODUCTION

The creation of durable concrete structures requires adequate compaction by skilled workers but the numbers of them in the construction industry are not enough. This factor leads to reduction in the quality of construction work. One solution for the achievement of durable concrete structures independent on the quality of construction work is the employment of Self Compacting Concrete (SCC). SCC is a concrete that can be placed by pump or skip and flow under its own weight with maintaining its homogeneity to completely fill the mold of any shape, even with congested reinforcement [4,9]. There are some marble mining and factories in South Sulawesi, Indonesia. Marble sludge is a disposal material from the marble production which must be disposed of in the landfills. The continuous disposing of marble sludge can destroy the environment and disturb the ecosystem around the disposing area. Recently, the natural conservation is a global issue and becomes one of the prominent aspects in the development program of many countries such as Indonesia. It is important to reduce the disposal of marble sludge in order to prevent the environmental damage. Marble sludge contains CaO of about 50-60% therefore it could be used as a replacement of cement. The incorporating of marble sludge in concrete production can reduce the disposal of marble sludge and also contribute to environmental protection. This paper is a preliminary study on the using of

1 Lecturer, Department of Civil Engineering, Hasanuddin University, Makassar Indonesia 2 Lecturer, Department of Civil Engineering, Hasanuddin University, Makassar Indonesia

9 CMT - 64


CMT -

marble sludge as partial replacement for ordinary Portland cement (Type I Portland cement) to produce SCC.

2. EXPERIMENTAL

2.1 Materials

All mixes of SCC were prepared by mixing water, Type I Portland cement, superplasticizer,sand, crushed stone and with addition of marble sludge. Cement is from Indonesia Cement Corporation and complies with ASTM C150. A commercial naphthalene sulfonate-based superplasticizer was used to enhance the deformability of SCC. Its dosage was kept constant for all mixes. It should be noted that the content of superlasticizer was taken into consideration in calculation of the total and the effective water of the mixes. The chemical composition of marble sludge is shown in Table 1. It has an average particle size of 150 μm.

Table 1. Chemical composition of marble sludge

The tests for physical properties aggregates were designed in accordance with ASTM C33-03 and their physical properties aggregates are shown in Table 2. The aggregates were used in surface dry condition.

Table 2. Physical properties of aggregates

Aggregate Properties

Crushed stones River sand

Maximum size (mm) 20 5.0

Surface dry 2.71 2.45 Density (kg/l) Oven dry 2.63 2.35

(24 hours) Water absorption (%) 0.86 2.23

Fineness (%) - 2.17

Abrasion (%) 25.2 -

No. Chemical composition (%)

1 SiO2 (Silica) 0.13

2 Al2O3 (Alumina) 0.31

3 Fe2O3 (Iron) 0.04

4 CaO (Lime) 55.07

5 MgO (Magnesia) 0.01

6 K2O (Potassa) 0.14

7 SO3 (Sulfuric anhydride) 0.08

8 LoI 44

9 CMT - 65


CMT -

2.2 Mixes Design of SCC

Table 3 shows mixes design of SCC. The water/binder (W/B) ratio was kept constant at 0.34. Mix 1 was a control mix which contained marble sludge of 5 % by weigth of cement as a filler. Replacement levels of cement by marble sludge for SCC were 0, 5, 10, 15 and 30 % by weigth of cement.

Table 3. Mixes design of SCC (1 m3 )

Unit weight (kg) Material

Mix 1 Mix 2 Mix 3 Mix 4 Mix 4

Water 199.29 199.29 199.29 199.29 199.29 Cement 554.80 525.60 496.40 467.20 438.00 Sand 782.31 782.31 782.31 782.31 782.31 Crushed stone 786.31 786.31 786.31 786.31 786.31 Superplasticizer 5.11 5.11 5.11 5.11 5.11 Sludge marble 29.10 58.40 87.60 116.80 146.00

2.3 Tests for Deformability and Self Compactibility

A combination of qualitative observations and quantitative measurements was applied to characterize the fresh condition of SCC. The measurement of the slump flow indicates the free deformability of SCC which can also indicate segregation resistance of SCC. Slump flow test with T50 test were done according to JIS A 1150. The slump flow is the average of horizontal flow (the largest diameter and the orthogonal to this) after lifting Abram`s cone. T50 is the time taken for the SCC to reach a spread diameter of 50 cm from the moment when the Abrams` cone lifted up. SCC was placed in the molds in accordance with JSCE-F515-1999. The cylindrical specimen was used for compressive and splitting tensile strength. The beam specimen was used for flexural strength. The dimension of cylindrical specimen was 15 (diameter) by 30 cm and dimension of beam specimen was 10 by 10 by 40 cm. After demoulding, the sides, bottom and ends of the cylindrical and beam concrete specimens should be checked by visualization method to evaluate the self compactibility of SCC.

2.4 Tests for Compressive, Flexural and Splitting Tensile Strength The specimens were cured in water at 20 ± 30 C after demoulding in accordance to ASTM C192/C192M-02. Compressive strength, flexural strength and splitting tensile strength were measured in accordance to ASTM C39/C39M-01, supplement of JIS A 1106 -1999 and ASTM C496-96, respectively. All strength tests were conducted at the age of 28 days. Each compressive strength and tensile splitting value was the average of four cylindrical specimens. Each flexural strength was the average of four beam specimens


3.1 Deformability and Self Compactibility of SCC

Figure 1 shows the time for SCC diameter to reach 50 cm (T50). Less time of T50 pointed at a higher content of marble sludge. The experimentally obtained results concerning the influence of cement replacement by marble sludge on slump flow are given in Figure 2. The slump flow of SCC were ranging from 55 to 67 cm. Figure 3 shows a part of slump flow test. Two observations indicated that all mixes flow homogeneously. First, it was observed that the flow of all mixes was circular. This indicates that the flow had been acted in all directions. Second, for all mixes it was observed that the coarse aggregates did not remain at the centre of the flow spread after removal the Abram’s cone. The coarse aggregates can flow with the other materials together and material segregation did not occur.

9 CMT - 66


CMT -

Figure 1. Time for SCC diameter to reach 50cm (T50)

Figure 2. Influence of cement replacement by marble sludge on slump flow

Figure 3 A part of slump flow test

Figure 3. A part of slump flow test

00.5

11.5

22.5

33.5

44.5

5

0 2 4 6 8 10 12 14 16 18 20

Replacement levels of cement by marble sludge (%)

T 5

0 (s

econ

d)

01020

30405060

7080

0 5 10 15 20

Replacement levels of cement by marble sludge (% )

Slum

p flo

w (c

m)

9 CMT - 67


CMT -

Result of compaction and density are summarized in Table 4. The result of compaction shows that the honey comb did not occur on the cylindrical and beam specimens, the surface of the hardened concrete was smooth and the large void did not appear. All mixes of SCC can flow under its own weight to completely fill the mould without assistance of mechanical vibration. SCC had densities ranging from 2212 to 2334 kg/m3, which can be category as normal weight concrete.

3.2 Test for Compressive, Flexural and Splitting Tensile Strength The influence of cement replacement by marble sludge on the compressive strength of SCC are shown in Figure 4. A 5-20% cement replacement by marble sludge decreased the compressive strength of SCC, with respect to the control mix, of about 2.7% -28.6%.

Figure 4. Influence of cement replacement by marble sludge on the compressive strength

Table 4 Result of compaction and density of SCC

No. Replacement

levels (%)

Density

(kg/m3)

Copaction result

Mix 1 0 2212 Excellent compaction, no honey combs , smooth surface without

large voids


large voids


large voids


large voids


large voids

0.0010.0020.0030.0040.0050.0060.00

0 5 10 15 20Replacement levels of cement by marble sludge (%)

Com

p St

reng

th (M

Pa)

9 CMT - 68


CMT -

The influence of cement replacement by marble sludge on the flexural strength and splitting tensile strength of SCC are shown in Figure 5 and Figure 6, respectively. The flexural strength and splitting tensile strength of SCC ranged from 4.79 MPa to 5.38 MPa and from 3.48 to 4.39 MPa, respectively. The increase in cement replacement by marble sludge did not give a significant effect to the flexural strength and splitting tensile strength. It shows that the flexural strength and splitting tensile strength can be achieved when the compaction has reached the good results.

Figure 5. Influence of cement replacement by marble sludge on the flexural strength

Figure 6 Influence of cement replacement by marble sludge on the splitting tensile strength

4. CONCLUSIONS

The following conclusions can be drawn from the experimental program conducted to evaluate the effect of cement replacement by marble sludge on the deformability and strength of SCC. SCC containing marble sludge can flow homogenously and completely fill the mould without assistance of mechanical vibration. The compressive strength was found to decrease as the cement replacement by marble sludge

0.00

2.00

4.00

6.00

8.00

10.00

0 5 10 15 20


Flex

ural

stre

ngth

(MPa

)

0.00

2.00

4.00

6.00

8.00

10.00

0 5 10 15 20


Split

ting

Ten

sile

Str

engt

h (M

Pa)

9 CMT - 69


CMT -

increased, but it did not give a significant effect to the flexural strength and splitting tensile strength.

5. ACKNOWLEDGEMENTS

The tests were conducted at Material and Construction Laboratory, Department of Civil Engineering, Hasanuddin University- Makassar, Indonesia.

6. REFERENCES

ASTM C192/C192M-02, “Standard practice for making and curing concrete test specimens in the laboratory.”

ASTM C39/C39M-01, “Standard test method for compressive strength of cylindrical concrete specimens.”

ASTM C496-96, “Standard Test Method for Splitting Tensile Strength of Cylindrical Concrete Specimens.”

Hajime Okamura and Masahiro Ouchi, (2003). “Self compacting concrete.” Journal of Advanced Technology, Vol.1, No.5

JIS A 1150 (2001). “Standard Test Method for Slump Flow of Concrete.” Japanese Standards Association, Tokyo.

JSCE-F515 (1999) (proposed). “Standard Practice for Making Test Specimens of High Fluidity Concrete.” Japan Society of Civil Engineering, Tokyo.

JIS A 1106 (1999). “Method of Test for Flexural Strength of Concrete- Supplement : Using Simple Beam With Center-Point Loading.” Japanese Standards Association, Tokyo.

Neville, A.M. (1995). “Properties of Concrete.” 4th ed, Prentice Hall. Steffen Grunewald and Joost C.Walraven (2001). “Parameter-study on the influence of steel fibers and

coarse aggregate contents on the fresh properties of self compacting concrete.” Cement and Concrete Research, Vol.31,No.12,pp.1793-1798

Tjaronge M.W. (2004). “Study on the Strength of Self Compacting Concrete Containing Marmoreal Gravel as Coarse Aggregate.” Proceeding of the First International Conference of Asian Concrete Federation, acf-016, Chiang Mai –Thailand.

9 CMT - 70


CMT -

STUDY ON PROPERTIES OF CONCRETE USING LOW QUALITY COARSE AGGREGATE FROM CIRCUM-PACIFIC REGION

Takahiro Nishida1, Nobuaki Otsuki2, Jyunji Yokokura3, Pitiwat Wattanachai4, Wanchai Yodsudjai5 and Ryosuke Onitsuka6

ABSTRACT : Since coarse aggregate occupies more than 40% of the volume in concrete, coarse aggregate phase largely influences on properties of concrete, particularly in the case of highly porous and low strength aggregates. It is provided with a vast supply of low-quality coarse aggregates from Circum-Pacific Region, i.e. volcanic aggregate or coral aggregate due to its geographical location surrounded by many volcanoes and oceans. From the construction viewpoint, utilization of material resources found around or near the construction site is deemed in order to optimize resources effectively. However, there were no researches that examine the properties of concrete, especially strength and chloride ion diffusivity, using these low-quality coarse aggregates. Therefore, this study attempted to examine the properties of concrete using low-quality coarse aggregates from Circum-Pacific Region. Firstly, the strength and chloride ion diffusivity of low-quality coarse aggregate were examined using minute property tests developed by the authors. Then the influences of the low-quality coarse aggregate on the properties of concrete were examined. Lastly, the quality improvement methods, coating of coarse aggregate and combined with the normal aggregate, for concrete using the low-quality coarse aggregates were proposed. As a result, it could be concluded that the strength and chloride ion diffusivity of concrete with low-quality coarse aggregate could be clarified by the properties of aggregates. In addition, the improvement methods proposed in this study could be successfully implemented for improving the strength and chloride ion diffusivity of concrete using low-quality coarse aggregate.

KEYWORDS: low quality coarse aggregate, concrete, strength, chloride ion diffusivity.

1. INTRODUCTION Since coarse aggregate in concrete occupies 40% of the volume, the coarse aggregate phase predominantly influences on characteristics of the concrete, particularly for a case of highly porous and weak aggregates. In Circum-Pacific region, where Japan also belongs to, it is provided with a vast supply of low quality coarse aggregate, i.e. volcanic aggregate and coral aggregate due to its geographical location surrounded by many of volcanoes and oceans. From the construction viewpoint, utilization of material resources found around or near the construction site is deemed in order to optimize resources effectively. However, there were no researches undertaken at the moment, that examine the strength and durability of concrete using these low quality coarse aggregates gathered from the Circum-Pacific region. Therefore, this study attempted to examine the strength and the durability of the concrete using the low quality coarse aggregates from the Circum-Pacific region. First, the strength and the diffusivity of the low quality coarse aggregates were investigated. Secondry subsequently the influences of the low quality coarse aggregate on the concrete characteristics were investigated. Lastly, the quality improvement methods for concrete using low the quality coarse aggregate were proposed.

1 Assistant Professor, Tokyo Institute of Technology, Japan

2 Professor, Tokyo Institute of Technology, Japan 3 Japan Green Resource Agency, Japan 4 Doctoral Course Student, Tokyo Institute of Technology, Japan 5 Lecturer, Kasetsart University, Thailand 6 Engineer, JR Kyusyu, Japan

10 CMT - 71


CMT -

2. EXPERIMENTAL PROCEDURE

2.1 Outline of Low Quality Coarse Aggregate Four types of the low quality coarse aggregate from the Circum-Pacific region are examined: Philippines aggregate (Volcanic), Kiribati aggregate (Coral), Nicaragua aggregate (Igneous Rock) and Okinawa aggregate (Lime Stone). In addition, crushed stone from Oume-made (Tokyo) is used as a normal coarse aggregate serving as a control variable for the comparative study. The physical properties of the coarse aggregates are shown in Table 1. Figure 1 shows the void structure of each low quality coarse aggregate taken from the microscope. It can be observed that the aggregates coming from the Philippines and Kiribati show continuous voids, and that from Nicaragua have separated voids. It is interesting to note that out of the four types of aggregates observed only the aggregates from Okinawa did not contain any void spaces, however, the results from porosity test shows that the aggregate from Okinawa contains many of very small voids.

Table 1. Physical Properties of Coarse Aggregate Name of

Aggregates Materials Density (kg/l)

Absorption Ratio (%)

Fineness Modulus

Solid Ratio (%)

Void Ratio (%)

Philippines Volcanic 0.95 35.0 6.47 65.6 35.80 Kiribati Coral 1.73 13.7 7.12 59.5 35.90

Nicaragua Igneous Rock 2.69 1.8 6.88 55.6 7.93 Okinawa Lime Stone 2.67 0.8 6.92 57.3 6.98 Normal Crushed Stone 2.62 0.9 6.48 59.8 0.93

200μm 200μm 200μm 100μm

Philippines aggregate Kiribati aggregate Nicaragua Aggregate Okinawa Aggregate Figure 1. Voids Structure of Low Quality Coarse Aggregate

Table 2. Mixing Proportions

(kg/m3) (C×%) Kind of Aggregates

W/B (%)

s/a (%) W C SF S G Water Reducing

Agent AE Agent Superplasticizer

Philippines 40 155 558 62 657 489 - - 0.9 Kiribati 40 155 558 62 657 734 - - 1.2

Nicaragua 40 155 558 62 657 1030 - - 2.0 Okinawa 40 155 558 62 657 1012 - - 2.0 Normal

25

40 155 558 62 657 997 - - 2.0 Philippines 42.5 170 425 - 729 489 1.00 1.20 -

Kiribati 42.5 170 425 - 729 734 1.00 0.60 - Nicaragua 42.5 170 425 - 729 1030 1.50 1.00 - Okinawa 42.5 170 425 - 729 1012 1.00 1.00 - Normal

40

42.5 170 425 - 729 997 2.00 1.00 - Philippines 45 175 318 - 806 489 0.50 1.00 -

Kiribati 45 175 318 - 806 734 0.00 0.60 - Nicaragua 45 175 318 - 806 1030 1.00 1.00 - Okinawa 45 175 318 - 806 1012 0.25 1.00 - Normal

55

45 175 318 - 806 997 0.30 0.60 - Philippines 46 180 257 - 840 489 0.00 0.25 -

Kiribati 46 180 257 - 840 734 0.00 0.25 - Nicaragua 46 180 257 - 840 1030 0.00 1.00 - Okinawa 46 180 257 - 840 1012 0.00 0.50 - Normal

70

46 180 257 - 840 997 0.00 1.00 -

10 CMT - 72


CMT -

2.2 Other Materials Ordinary Portland cement with 3.16 g/cm3 of density is used. Fine aggregate is river sand with 2.62 g/cm3 of density and 2.87 fineness modulus.

2.3 Mixing Proportions and Mixing Procedures Mixing Proportions are shown in Table 2. For concrete mixing procedure, the cement is sandwiched between two layers of sand in the mixer, and then the mixer rotates for 30 seconds. Water is added as the mixer rotates for 30 seconds. The mixing is stopped as soon as all water is already poured into the mixer. Finally, coarse aggregate is added and the concrete is mixed for 120 seconds.

2.4 Testing Procedures for Evaluation of Low Quality Coarse Aggregate (1) Minute Compressive Measurement A minute compressive measurement is used for evaluation a compressive strength of the low quality coarse aggregate. This method is developed originally by the authors (Yamane, 2001). The low quality coarse aggregate is cut into a cube of 3 x 3 x 3 mm as can be seen in Figure 2. The compressive measurement is performed by using the universal test machine with 5 kN maximum test load. The loading of the test piece can be seen in Figure 3. In addition, this minute compressive measurement is applied to use for evaluation the mortar that taken from the concrete specimens. (2) Minute Diffusion Measurement A minute diffusion measurement is used for evaluation a chloride diffusion coefficient of the low quality coarse aggregates. The test piece of 5 x 5 x 1 mm is cut from the coarse aggregate and set in the epoxy resin (φ30 x 1 mm) as can be seen in Figure 4. The minute-diffusion cell is set-up as can be seen in Figure 5. The 15 ml of 3% NaCl solution is filled one side of its container while on another side is filled with 15 ml saturated Ca(OH)2 solution. To obtain the chloride ion diffusion coefficient of the low quality coarse aggregate, it can be calculated using the Fick’s first law of diffusion (Cranck, 1996) by determining the change of concentration of the chloride ion on the saturated Ca(OH)2 side.

Test Piece

5mm5mm

Figure 2. Test Piece for Figure 3. Minute Compressive Measurement

Minute Compressive Measurement

Aggregate

Epoxy Resin Matrix

5mm

Test Piece setting in Epoxy

3% NaClSolu t ion

Sa tura ted Ca(OH)2Solu t ion

Figure 4. Test Piece for Figure 5. Minute Diffusion Cell

Minute Diffusion Measurement

10 CMT - 73


CMT -

2.5 Testing Procedures for Evaluation Concrete (1) Compressive Strength After 28 days curing in water at 20±2 ｡C, the compressive strength test is determined with φ100 x 200 mm cylindrical specimen. Three specimens are used in each case of the experiment. (2) Chloride Diffusion Coefficient The diffusion coefficient ion in concrete is investigated through an electro-migration test (Otsuki, 1998).


3.1 Strength and Chloride Diffusion Coefficient of Low Quality Coarse Aggregate Figure 6 shows the strength of the low quality coarse aggregate from the Circum-Pacific region. It can be seen that their compressive strength is lower as compared to the normal aggregate. Particularly, the strength of the Philippines aggregate and the Kiribati aggregate is less than 10% of the normal aggregate. Figure 7 shows the chloride ion diffusion coefficient of the low quality coarse aggregate from the Circum-Pacific region. It can be seen that the chloride ion diffusion coefficient of all low quality coarse aggregates are higher than that of the normal coarse aggregate from Tokyo. Particularly, the chloride ion diffusion coefficient of the Philippines aggregate and the Kiribati aggregate are around 100 times higher than that of the normal aggregate. As a result, it can be confirmed that the low quality coarse aggregates from the Circum-Pacific region have lower strength and higher chloride ion diffusion coefficient than that of the normal coarse aggregate from Tokyo.

3.2 Strength and Chloride Ion Diffusion Coefficient of Concrete Using Low Quality Coarse

Aggregates from Circum-Pacific Region Figure 8 shows the strength of concrete using the low quality coarse aggregates from the Circum-Pacific region. It can be seen that when the cement-water (C/W) ratio is low the type of coarse aggregate does not much influence on the strength of concrete. On the other hand, the difference of concrete strength can be clearly seen when the cement-water ratio is high, i.e. the type of coarse aggregate influences on the strength of concrete. Figure 9 shows the chloride ion diffusion coefficient of the concrete using the low quality coarse aggregate from the Circum-Pacific region. It can be seen that the coarse aggregate type does not much influence on the chloride ion diffusion coefficient of concrete in the case of 0.25 water-cement (W/C) ratio. On the other hand, for the case of 0.40, 0.55 and 0.70 water-cement ratios, the coarse aggregate type influences on the chloride ion diffusion

0

50

100

150

200

250

300

Philippines(Volcanic)

Kiribati(Coral)

Okinawa(Limestone)

Nicaragua(Lgneous Rock) Normal

100%

2% 9%

38%

58%

Kinds of AggregatesCom

pres

sive

stre

ngth

of A

ggre

gate

s (M

Pa)

0

2

4

6

8

10

Philippines(Volcanic)

Kiribati(Coral)

Okinawa(Limestone)

Nicaragua(Lgneous Rock) Normal

Kinds of Aggregates

94 times97 times

18 times17 times

1 timesChl

orid

e Io

n D

iffus

ion

Coe

ffici

ent

of A

ggre

gate

(x

10-8

cm2 /s

ec)

Figure 6. Compressive Strength of Coarse Aggregate Figure 7. Cl- Diffusivity of Coarse Aggregate

10 CMT - 74


CMT -

coefficient of concrete, i.e. the concrete using the low quality coarse aggregates have higher chloride ion diffusion coefficient than that of the normal aggregate.

3.3 Influence of Low Quality Coarse Aggregate on Concrete Strength and Chloride Diffusion

Coefficient Figure 10 shows the influence of the strength of the coarse aggregate on the strength of the concrete. It can be seen that the strength of the concrete increases as the strength of the coarse aggregate increases, however, the strength of the concrete trends to be steady after reach the peak for all cases of water-cement ratios. The reason can be explained from Figure 11 that when the strength of the coarse aggregate is lower than that of the mortar the strength of coarse aggregate influences on the strength of concrete, i.e. the strength of concrete increases as the strength of coarse aggregate increases because the cracks develops in both mortar and coarse aggregate. Figure 12(a) shows the fracture surface of the concrete in the case of the strength of the coarse aggregate is lower than that of the mortar (Philippines aggregate). It can be seen that the cracks propagate not only on the mortar portions but also throughout the coarse aggregate. On the other hand when the strength of coarse aggregate is higher than that of the mortar, the strength of coarse aggregate does not influence on the strength of concrete, i.e. the strength of the concrete is steady even the strength of the coarse aggregate increases. It can be confirmed by the cracks pattern as can be seen in Figure 12(b). Figure12(b) shows the fracture surface of the concrete in the case of the strength of the coarse aggregate is higher than that of the mortar (Normal aggregate). It can be seen that the cracks propagates only in the mortar portion. Figure 13 shows the influence of the chloride ion diffusion coefficient of the coarse aggregate on the chloride ion

0

20

40

60

80

100

0 1 2 3 4 5

PhilippinesKiribatiNicaraguaOkinawaNormal

Cement-Water Ratio C/W

Com

pres

sive

stre

ngth

ofC

oncr

ete

(MPa

)

0

10

20

30

40

50

60

70

80

0.25 0.40 0.55 0.70

PhikippineKiribatiNicaraguaOkinawaNormal

Chl

orid

e Io

n D

iffus

ion

Coe

ffici

ent

of C

oncr

ete

(x 1

0-8cm

2 /sec

)

Water Cement Ratio W/C Figure 8. Strength of Concrete Figure 9. Cl- Diffusivity of Concrete Using Low Quality Coarse Aggregate Using Low Quality Aggregate

0

20

40

60

80

100

120

0 50 100 150 200 250 300

W/C = 0.25W/C = 0.40W/C = 0.55W/C = 0.70

0

10

20

30

40

50

0 50 100 150 200 250 300

Philippines AggregateKiribati AggregateOkinawa AggregateNicaragua AggregateNormal Aggregate

Lightweight Aggregate

Philippines AggregateKiribati AggregateOkinawa AggregateNicaragua AggregateNormal Aggregate

Lightweight Aggregate

Strength of Mortar

W/C of Concrete = 0.40

Compressive Strength of Coarse Aggregate (MPa)Com

pres

sive

stre

ngth

ofC

oncr

ete

(MPa

)

Compressive Strength of Coarse Aggregate (MPa)

Com

pres

sive

stre

ngth

ofC

oncr

ete

(MP

a)

Figure 10. Relationship between Compressive Figure 11. Compressive Strength of Concrete Strengths of Coarse Aggregate and Concrete Using Different Aggregate Strengths

10 CMT - 75


CMT -

diffusion coefficient of concrete. It can be seen that the chloride ion diffusion coefficient of the concrete increases as the chloride ion diffusion coefficient of the coarse aggregate increases. This can be explained that when the high chloride ion diffusion coefficient coarse aggregate is used the chloride ion penetrates not only into the mortar portion but also into the coarse aggregate portion. On the other hand, when the low chloride ion diffusion coefficient coarse aggregate is used the chloride ion can penetrates into only the mortar portion.

3.4 Improvement Methods for Concrete Using Low Quality Coarse Aggregates The concrete using the low quality coarse aggregates from the Circum-Pacific region have almost lower strength and higher chloride ion diffusion coefficient than that of the concrete using the normal coarse aggregate as can be seen in Table 3. If the designed strength of concrete is less than the result shown in Figure 8, these low quality coarse aggregates can be used without any improvement. However, in order to utilize these low quality coarse aggregates in the better quality concrete, it is necessary to find the appropriate improvement methods. One example is the Nicaragua aggregate, the

b) St rength of Coarse Aggregate is h igher than tha t of Mor ta r(Only mor tar pa r t is broken .)

2cm2cm

Aggregate

Mortar

FractureSurface

a) St rength of Coarse Aggregate is Lower than tha t of Morta r(Both of mor ta r pa r t an d coarse aggrega te par t a re broken .)

2cm2cm

Aggregate

Mortar

FractureSurface

Figure 12. Fracture Surface of Concrete

0

20

40

60

80

0 2 4 6 8 10Cl- Diffusion Coefficient of Concrete (x 10-8 cm2/sec)

Cl-

Diff

usio

n C

oeffi

cien

tof C

oncr

ete

(x 1

0-8cm

2 /sec

)

W/C=0.70

W/C=0.55

W/C=0.40

W/C=0.25

Figure 13. Relationship between Cl- Diffusivity of Concrete And Aggregate

10 CMT - 76


CMT -

concrete using the Nicaragua aggregate have the strength almost equivalent to the concrete using the normal coarse aggregate. However, the chloride ion diffusion coefficient of the concrete using the Nicaragua aggregate is higher than that of the concrete using the normal coarse aggregate. Therefore, the coating of coarse aggregate (Nathaniel, 1999) is thought as the method for improving the concrete using the Nicaragua aggregate. The Nicaragua aggregate is coated with the cement paste of 0.25 water-cement ratio prior to concrete mixing. The coating of coarse aggregate method leads the concrete using the Nicaragua aggregate more permeable because of the chloride ion can penetrate into only the mortar portion but cannot penetrate into the coarse aggregate due to the protection of coated cement paste. As can be seen in Figure 14, the coating of coarse aggregate method can lower the chloride ion diffusion coefficient of the concrete using the Nicaragua aggregate around 45%. In addition, it can be seen that the coating of the coarse aggregate can also improve the strength of the concrete using the Nicaragua aggregate around 50%. For other types of the low quality coarse aggregates, the improvement methods and the results after the improvement are shown in Table 4. It can be seen that the improvements in strength and chloride ion diffusion coefficient of the concrete using the low quality coarse aggregates from the Circum-Pacific region can be achieved by using the proposed improvement methods.

Table 3. Comparison between Concrete Using Low Quality And Normal Coarse Aggregate

0

10

20

30

40

50

Com

pres

sive

Stre

ngth

（ＭＰａ）

0

5

10

15

20

25

30

Chl

orid

e Io

n D

iffus

ion

Coe

ffic

ient

(×10

-8cm

2 /sec

)

BeforeImprovement

AfterImprovement

BeforeImprovement

AfterImprovement

50% Up45% Down

Figure 14. Improvement of Strength and Cl- Diffusivity of Concrete Using Nicaragua Aggregate by

Coating of Coarse Aggregate

Compressive Strength Chloride Ion Diffusion Coefficient

Water-Cement Ratio Water-Cement Ratio Aggregate Type

0.25 0.40 0.55 0.70 0.25 0.40 0.55 0.70

Philippines Aggregate X X X X ○ X X X

Kiribati Aggregate X X X X ○ X X X

Nicaragua Aggregate ○ ○ ○ ○ ○ X X X

Okinawa Aggregate X ○ ◎ ◎ ○ X X X

◎ = Better than Normal Aggregate ○ = Same as Normal Aggregate X = Worse than Normal Aggregate

10 CMT - 77


CMT -

.

4. CONCLUSIONS This study aims to examine the strength and the durability of the concrete using the low quality coarse aggregates from Circum-Pacific region. The conclutions obtained by this study can be summarised as follows; (1) The minute compressive measurement and minute diffusion measurement, which are

developed by authors, can be successfully applied to the evaluation of properties of low quality aggregates from Circum-Pacific regions.

(2) The influence of the coarse aggregate on the strength and the chloride ion diffusion coefficient

of concrete can be clarified as follow.

•When the strength of the coarse aggregate is lower than that of the mortar, the strength of the coarse aggregate influences on the strength of the concrete. On the other hand, when the strength of the coarse aggregate is higher than that of mortar, the strength of the coarse aggregate does not influence on the strength of the concrete.

•The chloride ion diffusion coefficient of the coarse aggregate greatly influence on the

chloride ion diffusion coefficient of the concrete. (3) The improvement in strength and chloride ion diffusion coefficient of the concrete using the

low quality coarse aggregates form the Circum-Pacific region can be achieved by using the proposed improvement methods.

5. REFERENCES Cranck, J. (1996). “The Mathematics of Diffusion” 2nd ed., Clarendon Press, New York. Nathaniel, B.D., Otsuki, N. and Miyazato, S. (1999). “The Use of Coated Coarse Aggregate and

Double Mixing in Making Concrete” Proc. of the 1st International Summer Symposium, JSCE, Tokyo, Japan pp.275-278.

Otsuki, N., Hisada, M., Otani, M. and Maruyama T. (1998). “Theoretical Assessment of Diffusion Coefficient of Chloride Ion in Mortar by Electro-Migration Method” Journal of Materials, Concrete Structure and Pavements, No.592/V.39, pp. 97-105 (in Japanese).

Yamane, H., Otsuki, N., Nishida, T. and Yodsudjai, W. (2002). “Development of Measuring Methods to Evaluate Tensile and Compressive Strength of Minute Regions in Concrete” Proc. of the 4th International Summer Symposium, JSCE, Tokyo, Japan, pp.291-294.

Table 4. Improvement Method for Using Low Quality Coarse Aggregate in Concrete

Aggregate Types Improvement Methods Improvement of

Compressive Strength

Improvement of Chloride Ion Diffusion

Coefficient

Nicaragua Aggregate Coating of Coarse Aggregate 50% Improvement 45% Reduction

Okinawa Aggregate Coating of Coarse Aggregate 15% Improvement 70% Reduction

Philippines Aggregate Combine with Normal Aggregate 45% Improvement 40% Reduction

10 CMT - 78


CMT -

TENSILE STRENGTH BEHAVIOR OF POLYMER-MODIFIED CONCRETE MADE WITH PPMM AS AN ADDITIVE

Abdullah Saand1, Mohammad Ismail2 and Salihuddin Radin Sumadi3

ABSTRACT: This paper reports the results of tensile strength behavior as a part of the ongoing research on possible development of polymer-modified concrete (PMC) by introducing prepackaged polymer-modified mortar (PPMM) as an additive to ordinary concrete. Polymer-modified concrete was manufactured with two different approaches; I: whatever %age of PPMM was added, same %age of hydraulic cement was reduced, II: both hydraulic cement and fine sand were reduced as per their manufacturing %age content in PPMM. The approach II shows significantly better results in terms of tensile strength and workability, with all the dosages (%ages) of PPMM. For this experimental study, 54 specimen as cylinders (100mmφ, 200mm height) were prepared and tested on 28 day. This study demonstrates that tensile strength of PMC is improved considerably by addition of various %ages of PPMM. Addition of PPMM as 5%–20% resulted upto 34 % higher tensile strength than that of ordinary concrete and further increase in dosage of PPMM gave less strength. Polymer-modified concrete mixes were also found workable and cohesive even with slump range of 8mm–50mm.

KEYWORDS: Polymer-Modified Concrete (PMC), Prepackaged Polymer-Modified Mortar (PPMM), Additive, Tensile Strength

1. INTRODUCTION AND BACKGROUND

Concrete is one of the most widely used construction materials worldwide due to its simplicity, superior properties and low cost. However, there are few major flaws in the properties of concrete, one amongst is low tensile strength. Polymers are often used to improve the strength, durability and workability of the mortar and concretes. On the word of Ohama (1995), polymer-modified or polymer cement mortar and polymer cement concrete are materials made by partially replacing the cement hydrate binders of conventional cement mortar or concrete with polymer. A number of thermoplastic or thermosetting polymers are used in modifying mortar and concrete; used in various forms, like; liquid resin, latexes, redispersible polymer powders and water-soluble homo-polymers or copolymers (Ohama, Y. 1998 and Okada, K. & Ohama, Y. 1987). Mostly, polymers in the form of liquid or powders are mixed in fresh cement mortar and concrete, during mixing process. However at the jobsite problems are some times faced in preparing mixes because of complex mix calculations, resulting batch to batch variation in the properties. The advent of powdered emulsions (powdered cement modifiers) with improved qualities made possible the production of PPMM and settle down that tedious job and uncertainties in the mixes, in case of mortar. The ready-mixed products as polymer-modified mortars have being launched in various countries. Prepackaged polymer-modified mortars (PPMMs) are highly helpful in improving the handling procedures and to avoid mixing errors (Afridi, et. al. 2001). Only water has to be added in the dry blends on the construction site prior to application (Hackel, E., et. al, 1987). 1 Ph. D Researcher, Faculty of Civil Engineering, Universiti Teknologi Malaysia, 81310 UTM Skudai, Johor, Malaysia Corresponding Author’s Email: [email protected] or [email protected] Fax No. +607-5566157 2 Associate Professor, Head of Material & Structures Dept., Faculty of Civil Engineering, Universiti Teknologi Malaysia. 3 Professor, Faculty of Civil Engineering, Universiti Teknologi Malaysia, 81310 UTM Skudai, Johor, Malaysia

11 CMT - 79


CMT -

But still same problem exists for polymer-modified concrete production, to undertake the such complexity, possible usage of those strictly controlled manufactured products; PPMMs as an additive, can get rid of that and has not been studied. In Japan, polymer-modified mortar (PMM) is most widely used for finishing and repair works, but polymer-modified concrete (PMC) is seldom employed because of poor cost-performance balance (Ohama, 1997). However, PMC is widely used for bridge deck overlays and patching work in the United States (Fowler, D.W. 1990). The primary limitation and one of impediments to the use of PMC is its cost (Fowler, 1999 & Kardon, 1997). The polymer-modified concrete (PMC) can possibly be developed as cost- befitting and effective concrete repair material as well, by utilizing PPMM, as an additive. The usage of prepackaged polymer-modified mortars as additive is not so well documented, no any research data is available on use of PPMM as additive; except few successful case-histories (Afridi, et al., 2003, and Afridi, et al., 2006) of repair and renovation and recent initiated research by author. The compressive strength of polymer-modified concrete (PMC) composites, developed by introducing PPMM as an additive, improves significantly (Saand, A. et. al. 2006a). The pre-damaged beams repaired with polymer-modified mortar (PMM) prepared by using PPMM as additive; shows significantly improved flexural behaviour, as damaged to reinstate, for operational use (Saand, A. et. al 2006b). This study presented here, as tensile strength behaviour of polymer-modified concrete (PMC), is a part of author’s ongoing behavioural research on concrete modified with prepackaged polymer-modified mortar (PPMM) to produce PMC.

2. EXPERIMENTAL WORK

2.1 Materials Used Cement: Ordinary Portland Cement (OPC) of “Seladang” brand from Tenggara Cement Manufacturing Sdn. Bhd. complying with the Type I Portland Cement as in ASTM C150: 1992 and BS 12: 1991 which is same as Malaysian Standard MS 522: part I: 2003. Prepackaged Polymer-Modified Mortar (PPMM): A commercial prepackaged formulation developed in Pakistan as Hi-Bond (Universal) - Prepackaged Polymer-Modified Mortar (PPMM); a ready-mixed product, of M.H. DadaBhoy Group, Karachi, Pakistan. That ready-mixed product, possibly be used as an additive to normal concrete to develop suitable matrixes of Polymer-Modified Concrete, complying essential criterion of repair materials. Fine Aggregate & Coarse Aggregate: The fine aggregate and coarse aggregate used were crushed granite type, complying the requirement of ASTM C778-91. The sand used was sieved; to remove foreign and rubbish material and to get size less than 2.36 mm. The percentage passing the 600 µm sieve of 55 % is used in concrete mix design and the coarse aggregate of nominal size 10 mm is used in this study. The coarse aggregates are being washed, to remove the dust, dirt, debris or any other rubbish material and air dried, before use. Mixing Water: Tap water; suitable for drinking; was used for manufacturing of concrete.

2.2 Ordinary Concrete Mix Design and Trials The characteristic compressive strength of normal concrete (reference concrete) used in this study was taken as 30 N/mm2 at 28 days. To get correct proportion of constituent materials, the concrete mix design was done on the basis of DoE’s ‘Design of Normal Concrete Mixes’, BRE, 1997. The designed proportioning is shown in Table 1.

11 CMT - 80


CMT -

Table 1. Proportioning of Ordinary Concrete Mix To satisfy the mix design for normal concrete of targeted characteristic strength, trial mixing was done, as per ASTM C 192- 02, and the testing of trial mix cubes, as per ASTM C109.

2.3 Preparation of Specimens as Polymer-Modified Concrete (PMC) In this study basic mix proportion of PMC was: Ordinary Concrete (OC) : %age PPMM : w/b ratio; (by weight). Different %ages of PPMM (5%, 10%, 15%, 20%, 25%, 30%, & 40%) as additive were added to Ordinary Concrete (proportioning shown in Table 1). Two w/b ratios; 0.5 & 0.45 were selected to use, after evaluating different w/b ratios in terms of workability. The curing of PMC specimen was done as per Japanese Industrial Standard (JIS); 48 hour demolds, 5 days wet curing & 21 days dry curing. Polymer-modified concrete specimens were fabricated with two different approaches of addition & subtraction. Approach I: whatever %age of PPMM was added, same %age of hydraulic cement was reduced, and approach II: both hydraulic cement and fine sand were reduced as per their manufacturing %age content in PPMM. Altogether 54 cylinders (100mmφ, 200mm height) were fabricated as per specifications of JIS A 1171: 2000 and ASTM C192, to study material behaviour in terms of tensile strength and slump.

2.4 Testing Considering the relative importance of tensile strength, in some application for restoration of deteriorated concrete in tension zones, this property was investigated. The testing for tensile strength and slump was carried out with modified and unmodified concretes. Cylinder-splitting testing for indirect tensile strength on 28 days, as per JIS 1171: 2000 & ASTM C109 using DARTEC universal testing machine with loading rate 2KN/s (Figure 1 & Figure 2).

Figure 1. Specimen under indirect tensile stresses

Ordinary Cement 426 (Kg/m3) Fine Aggregates 722 (Kg/m3) Coarse Aggregates 997 (Kg/m3) Water 230 (Kg/m3) W/C Ratio 0.54 Slump 30-60 (mm)

11 CMT - 81


CMT -

Figure 2. Two-halves of cylinder after cylinder-splitting test The slump of each individual batch was carried out and recorded. The details of all test results are precisely incorporated in Table.2


Table 2, presents thoroughly compiled test data of 54 specimens as; tensile strength, and slump of polymer-modified concrete (PMC) mixtures. During all the casting and testing for slump, significant cohesiveness was witnessed in all samples under observation, even with slump value 8mm-50mm behaviour was as true slump, no any single test shown shear or collapse. With lowest optimized w/b (water-binder) ratio; 0.45, slump was measured as 8mm (lowest) and 32mm (highest) but the mixture found be cohesive and workable. %age PPMM, w/c ratio and slump relation is illustrated in Figure 3 & Table 2.

0

10

20

30

40

50

60

70

0 5 10 15 20 25 30 40

% age PPMM

Slum

p (m

m)

PMC (approach I) PMC (approach II)

Figure 3. Slump of PMC with Different approaches and %age dosage of PPMM

11 CMT - 82


CMT -

0.00

0.50

1.001.50

2.00

2.50

3.00

3.504.00

4.50

5.00

0 5 10 15 20 25 30 40

% PPMM

Ten

sile

Str

egth

(N/m

m2)

PMC (approach I) PMC (approach II)

Figure 4. Tensile Strength of Polymer-Modified Concrete Vs Ordinary Concrete

Figure 4, shows the behaviour of different dosages (%age) of PPMM in PMC made with different approaches, in terms of strength gain. From data in Table 2 & Figure 4 it is obvious that introduction of PPMM as an additive by replacing hydraulic cement and fine aggregate as per manufacturing contents of PPMM improves the tensile strength significantly. It is apparent from the data as shown in Figure 4 & the data in Table 2, that the tensile strength is improved by the use of PPMM as an additive; 5% to 30%age dosage of PPMM to ordinary concrete, resulted increased strength than characteristic strength of ordinary concrete. However, 5% to 20%age shows higher strength than the trial mix on ordinary concrete. PPMM more than 30%age gave less strength than the characteristic strength.

Table 2. Details of tensile strength and slump of PMC trial mixes (48 specimens)

Ordinary Concrete

%age of PPMM

Average* Ultimate

Load (KN)

Tensile Strength (N/mm2)

Slump (mm)

Average* Ultimate

Load (KN)

Tensile Strength (N/mm2)

Slump (mm)

* Average; means average value of 3 specimen

139.60 4.45

113.50

86.60 2.76

100.65 3.21

97.37 3.10

142.85 4.55

146.05 4.6510 117.10

15 94.70

40 77.74

20 86.10

25 79.67

30 79.07

PMCw/c

Ratio

Approach I

Grade 30

0 0.54 113.10

5

0.45

141.00

13

15

16

Approach II

58

8

10

108.70 3.46

3.61

18

55

8

10

15

20

24

30

32

12

2.52

2.48

3.60

4.49

3.73

3.02

2.74

2.54

11 CMT - 83


CMT -

The produced PMC composites containing PPMM as additive, shows improved tensile strength; as PPMM contains, hydraulic cement, fine aggregates, a combination of synthetic polymers including redispersible polymer-powder, anti-foaming, water-repelling and shrinkage compensating agents. Hence, in such composite system, the cement as the inorganic binder is responsible for compressive strength, and the redispersible powder, as the organic binder, acts as reinforcement and is responsible for most of internal tensile strength and adhesion bond strength at interfaces (Schulze, J., 1993). Hopefully, further going-on study on investigation of other essential repair material criterion will show improved behaviour, at least to acceptable extent. That will certainly lead a new trend of utilization of Prepackaged Polymer-Modified Mortar (PPMM) as additive and production of low-cost Polymer-Modified Concrete (PMC) with reasonable properties, as repair material.

4. CONCLUSION

Tensile strength of developed polymer-modified concrete composites is improved significantly and mixtures found cohesive and workable. From the previous and present study, it is obvious that the utilization of PPMM as additive for development of PMC will positively work, resulting more effective and cost benefiting repair and strengthening for restoration of deteriorated reinforced concrete structures. The PMC composites will only be formulated and proposed on the basis of outcomes of on-going in-depth evaluation of PMC trial mixes in-terms other mandatory properties / criterion of repair materials, as; flexural strength, bond strength, adhesion with substrate, permeability, and drying shrinkage.

5. ACKNOWLEDGEMENTS

The PPMM for this research was supplied by M.H. DadaBhoy Group & Construction Technologies (Pvt.) Ltd., Pakistan. The facilities for carrying out this research were provided by Structure and Material Laboratory, Faculty of Civil Engineering, Universiti Teknologi Malaysia (UTM).

6. REFERENCES

Afridi, M.U.K., Rizwan, S. A. and Memon, J.I., (2001). “A Revolutionary Prepackaged Polymer-Modified Mortar with Field Applications and Economic Advantages”, In PEC Proceedings, Vol. 68, pp. 373-381.

Afridi, M.U.K., Khan, A. A., Rizwan, S. A, and Khaskheli, G.B., (2003). “Hi-Bond; Polymer–Modified Mortar with filed application and economic advantages”, Proc. IQRIP 2003, International Conference on Investment & Quality in Roads & Infrastructure Projects, Lahore, Pakistan, January 21-23.

Afridi, M.U.K., Abbasi, A.G., Khaskheli, G.B. & Saand, A. (2006). “Polymer-Modified Cement Systems roved Successful in Resisting Earthquake Jolts.” 5th Asian Symposium on Polymers in Concrete, SERC, Chennai, India. (Accepted for presentation).

Fowler, D.W., (1990). “In Polymers in Concrete.” Proceedings of 6th International Conference on Polymers in Concrete, Huang, Y., Wu, K., Chen, Z., (Eds.). International Academic Publishers; Beijing: p.10-27

Fowler, D.W., (1999). “Polymers in Concrete: a vision for 21st century”, Cement and Concrete Composites, Vol. 21: pp. 439-452

Hackel, E., Beug, P. & Horler, S., (1987). “The use of redispersible polymer powders in concrete restoration”, B.W. Stynes (Ed.). Proceedings of the 5th International Congress on Polymers in Concrete, Brighton UK, pp. 306

Kardon, J.B., (1997). “Polymer-modified concrete: Review.” Journal of Materials in Civil Engineering, Vol.9, No.2, May: pp. 85-92

11 CMT - 84


CMT -

Okada, K. & Ohama, Y. (1987). “Recent Research and Application of concrete-polymer composites in Japan.” In Proceeding o f the 5th International Congress on Polymers in Concrete, Brighton UK, September, 13-21.

Ohama, Y. (1995). “Properties and Process Technology, Hand Book of Polymer Modified Concrete and Mortars.” Noyes Publications, USA.

Ohama, Y., (1997). “Recent Progress in Concrete-Polymer Composites”, Advanced Cement Based Materials, Vol. 5, pp. 31-40

Ohama, Y., (1998). “Polymer based Admixtures.” Cement and Concrete composites, Vol. 20, pp. 189-212.

Saand, A., Ismail, M., Sumadi, S.R, & Afridi, M.U.K. (2006a). “Strength Behaviour of Polymer-Modified Concrete using Prepackaged Polymer-Modified Mortar as an Additive.” 5th Asian Symposium on Polymers in Concrete, SERC, 10-11 September, Chennai, India. (Accepted for presentation).

Saand, A., Ismail, M., Sumadi, S.R. & Afridi, M.U.K., (2006b). “Behaviour of Pre-Damaged RC Beams Repaired with Polymer-Modified Mortar”, In Proc. Tenth East Asia Pacific Conference on Structural Engineering and Construction (EASEC-10). 3-5 August, Bangkok.

Schulze, J., Jodlbauer, F, & Adler, K., (1993). “Polymer-modified mortars for the Renovation and Rehabilitation of Concrete Structures”, In: Y. Tezuka, G.W. de Puy (Eds.) International Congress on Polymers in Concrete, Bahia Brasil.

11 CMT - 85


CMT -

INFLUENCE OF FIBER VOLUME FRACTION ON THE COMPRESSIVE STRENGTH OF A FIBER-REINFORCED, LIGHTWEIGHT AGGREGATE,

CELLULAR CONCRETE

H.K. Lee1 and S.Y. Song2

ABSTRACT : This paper summarizes the results of an experimental study (Lee and Song, 2006) conducted to investigate the effect of volume fraction of fibers on the compressive strength of a fiber-reinforced, lightweight aggregate cellular concrete (FRLACC). A series of compressive strength tests on FRLACC specimens with various fiber volume fractions (0 %, 0.10 %, 0.25 %, 0.50 %) were carried out. The FRLACC that was produced without an autoclave process enhanced the compressive strength of cellular concrete by adding lightweight aggregates and polypropylene fibers. It was observed from the tests that a 0.25 % volume fraction of fibers maximized the increase in the strength of FRLACC and the crack propagation in FRLACC was hindered by the presence of fibers in the cement matrix. KEYWORDS: lightweight aggregate, cellular concrete, optimal volume fraction of fibers,

compression strength tests

1. INTRODUCTION

Autoclaved lightweight concrete has been used to reduce the self weight of commercial and residential buildings, and to minimize the risk of earthquake damages to the structures (e.g., Yasar et al., 2003). However, the autoclave curing process adopted into the conventional casting process of lightweight concrete to increase its strength is very costly and has limitations on the size of the products. On the other hand, cellular concrete reinforced with polypropylene fibers has been emerged as a possible viable construction material due to its excellent thermal insulation and freeze-thaw resistance and high resistance to cracking (El-Aryan, 1995; Zollo and Heys, 1998). However, its low compressive strength ranging from 1.7 to 2.1 MPa (Zollo and Heys, 1998) has greatly hindered the progress of the cellular concrete as a mainstream of construction materials. This experimental study aimed at investigating the influence of fiber volume fraction on the compressive strength of a fiber-reinforced, lightweight aggregate, cellular concrete (FRLACC) that was produced without an autoclave process. A foam agent mixed with a polycarboxylate polymer based superplasticizer was used to produce air bubbles in the cement matrix. Multifilament shape polypropylene fibers having a length of 19 mm were added in the cement matrix. A series of compression strength tests on FRLACC specimens with various fiber volume fractions (0 %, 0.10 %, 0.25 %, 0.50 %) were conducted to investigate the effect of volume fraction of fibers on the compressive strength of FRLACC and to determine the optimal volume fraction of fibers.

2. SPECIMEN PREPARATION

The concrete mix used in this test was composed of Type I Portland cement, lightweight fine and coarse aggregates made with expanded shale, a foam agent mixed with a polycarboxylate polymer based superplasticizer, and multifilament shape polypropylene fibers. The foaming agent used in this experiment was a proprietary system developed by Thermoflex, Inc. The mix proportion adopted to prepare the specimens is elaborated in Table 1. 1 Department of Civil and Environmental Engineering, Korea Advanced Institute of Science and Technology, Guseong-dong, Yuseong-gu, Daejeon 305-701, South Korea 2 Department of Civil and Environmental Engineering, Korea Advanced Institute of Science and Technology, Guseong-dong, Yuseong-gu, Daejeon 305-701, South Korea

12 CMT - 86


CMT -

Table 1. Mix proportion (see also Lee and Song, 2006)

Mix no. Water

(kgf/m3)

Cement

(kgf/m3)

Fine Agg.

(kgf/m3)

Coarse Agg.

(kgf/m3)

Fibers

(kgf/m3)

Admixture

(kgf/m3)

F1 144.8 420 275 220 - 0.880

F2 144.8 420 275 220 0.910 (0.10 %) 0.880

F3 144.8 420 275 220 2.275 (0.25 %) 0.880

F4 144.8 420 275 220 4.550 (0.50 %) 0.880

In this study, twenty cylindrical specimens of size 150 mm (diameter) × 300 mm (height) were prepared from FRLACC mix. In order to prepare the FRLACC mix, firstly lightweight fine and coarse aggregate, and water were pre-mixed for about one minute in a mechanical mixer machine, and then the cement was poured slowly. The admixture prepared with foaming agent and superplasticizer was added at the time when a half of the cement was poured and then rest of the cement was added into the mixer. Finally, the polypropylene fibers were added and mixed for about five minutes until a uniform and flowing mixture was obtained. The FRP cylindrical molds were filled with fresh concrete mix and compaction of concrete within molds was carried out by using a portable vibrator. The specimens were removed from the molds after 24 hours and tested at the age of ten days curing.

3. COMPRESSIVE STRENGTH TEST

3.1 Test Description The testing method, material testing system and the relevant apparatus utilized for this test are based upon ASTM C 39 (Ref. 5) and ASTM C 496 (Ref. 6). The test was conducted under the displacement control at a constant head-loading rate of 0.005 mm/sec. Deflections, strains and loads were automatically recorded after every 3 seconds using a laptop computer connected with the UCAM-20PC data logger until the specimen reached failure. Failure was determined by monitoring the major drop in the load–time curve.

3.2 Results The measured density, compressive strength, modulus of elasticity, Poisson’s ratio of the specimens is summarized in Table 2. It is observed from the table that the compressive strength of the specimens ranges from 5.27 to 10.82 MPa, depending on the fiber volume fraction. It is seen from the table that the compressive strength is increased as the volume fraction of fibers continues to increase up to 0.25%, whereas it is decreased beyond the volume fraction of fibers of 0.25%. Figure 1 shows a typical fracture pattern of the compression test specimens. As shown in the figure, the specimens were cracked near the top edge and then the cracks propagated vertically. As the majority of air bubbles are distributed on the top portion of the specimens, the top edge was almost crushed before the vertical cracks reached the bottom. The scanning electron micrograph (SEM) images of cracked surface of FRLACC are shown in Figure 2. It is observed from the figure that the fibers connect two separated segments, illustrating the fiber reinforcing effect. It is concluded that the crack propagation in FRLACC is hindered by the presence of fibers in the cement matrix.

12 CMT - 87


CMT -

Table 2. The measured density, compressive strength, modulus of elasticity, and Poisson’s ratio of the FRLACC specimens (see also Lee and Song, 2006)

Mix no. Specimen no. Density

(kgf/m3)

Compressive strength

(MPa)

Modulus of elasticity

(MPa) Poisson's ratio

1 1050 5.01 5456 0.19

2 1110 4.08 5754 0.20

3 1090 6.01 6017 0.21

4 1110 5.74 5799 0.20

5 1070 5.52 5915 0.20

F1

Average 1086 5.27 5788 0.20

1 - - - -

2 1100 6.26 4946 0.18

3 1070 5.90 4528 0.21

4 1100 5.98 6579 0.23

5 1110 6.40 5012 0.19

F2

Average 1095 6.14 5266 0.20

1 1110 11.47 6475 0.21

2 1090 9.79 5843 0.18

3 1120 11.68 6345 0.20

4 1080 11.18 6001 0.18

5 1050 9.98 5963 0.19

F3

Average 1090 10.82 6125 0.19

1 1060 6.05 4735 0.17

2 1060 6.08 4466 0.17

3 1060 5.96 4492 0.17

4 1070 5.50 5816 0.24

5 1070 6.51 4730 0.13

F4

Average 1064 6.02 4848 0.18

12 CMT - 88


CMT -

Figure 1. Failed specimen after the compression test (Lee and Song, 2006)

Figure 2.SEM images of cracked surface of FRLACC (Lee and Song, 2006)

4. CONCLUDING REMARKS

This study determined the effect of fiber on the compression strength and cracking behavior of cellular light weight concrete (FRLACC) produced without an autoclave curing process. A special foaming agent was used to produce air bubbles in the cement matrix. The specimens were prepared with various percentages of fiber volume fractions (0 %, 0.10 %, 0.25 %, 0.50 %) to investigate the effect of volume fraction of fibers on the compression strength of FRLACC and to determine the optimal volume fraction. A series of compression tests on FRLACC specimens were carried out to investigate the influence of fiber volume fraction on compression strength of FRLACC. The experimental results indicated that fibers and lightweight aggregates are capable of substantially increasing the compressive strength of FRLACC. The details of tests and the results can be found in Lee and Song (2006). The findings emerged from this experimental study can be grouped as follows: 1. Mixing of fibers substantially increased the compressive strength of lightweight cellular concrete. 2. The optimal fiber volume fraction was found to be 0.25 % which maximized the rise in the

12 CMT - 89


CMT -

compressive strength of the FRLACC. 3. It was also found that fibers control the crack propagation of FRLACC as shown in the SEM

images. Significant research work will be conducted along this line of research to enhance the performance of FRLACC: Investigations on the effect of other types of fibers, fine silica fume, fly-ash, and curing conditions on the mechanical properties of FRLACC.

5. ACKNOWLEDGEMENT

The authors gratefully acknowledge the technical support from Thermoflex Inc., FL and the cooperation of Mr. B.R. Kim, Mr. S.K. Ha and Mr. Muhammad Afzal at the KAIST. This research was supported by start-up funds provided to Dr. Lee from the Korea Advanced Institute of Science and Technology.

6. REFERENCES

El-Aryan, E.(1995). “An investigation of selected material properties of fiber reinforced cellular concrete.” Master Thesis, University of Miami, Florida, U.S.A., 1995.

Lee H.K., Song S.Y., “Effect of volume fraction of fibers on the mechanical behavior of a lightweight aggregate concrete reinforced with polypropylene fibers.” Journal of Korea Concrete Institute, submitted.

E. Yasar, C.D. Atis, A. Kilic, H. Gulsen (2003). “Strength properties of lightweight concrete made with basaltic pumice and fly ash.” Materials Letters, 57 (15), pp 2267-2270.

Zollo R.F., Heys, C.D. (1998). “Engineering material properties of a fiber-reinforced cellular concrete.” ACI Materials Journal, 95 (5) (1998) 631-635

American Society for Testing and Materials (2004). “Standard Test Method for Compressive Strength of Cylindrical Concrete Specimens, C 39/C 39M – 04a.” IHS Specs & Standards.

American Society for Testing and Materials (2002). “Standard Test Method for Static Modulus of Elasticity and Poisson’s Ratio of Concrete in Compression, C 469 – 02.” IHS Specs & Standards.

12 CMT - 90


CMT -

STUDY ON PORONG MUD-BASED GEOPOLYMER CONCRETE

Januarti Jaya Ekaputri1 and Triwulan2

ABSTRACT: This paper describes the efforts of how to use the solid form of the hot mud in Sidoarjo, Indonesia, as a base material of geopolymer concrete. The experimental study shows that the mud mixed with fly ash in certain composition is potential to become a binder in concrete, instead of Portland cement. It also shows that the mud mixed with Portland cement and lime does not have any compression strength although it is mixed with fly ash and alkaline activator. Dry mud has similar chemical composition to fly ash, but the compound of dry mud is not amorphous. Wet mud can only be used as a binder paste when mixed with a fly ash and alkaline activator solution. Test results show that the compressive strength of the binder does not vary with age and curing, but the concrete does. The longer the curing period and the higher ratio of the binder : aggregates will result in the higher compressive strength of geopolymer concrete. Furthermore, the size of coarse aggregate and the molarity of alkaline activator play very important role in the improvement of compressive strength. KEYWORDS: Geopolymer concrete, mud, fly ash, alkaline activator, binder

1. INTRODUCTION

Since May 29, 2006, Porong mud materials have become a big major issue from the bowels of the earth as a side effect of the missed gas mining process in Porong, Sidoarjo. The quantity of hot mud is equal to 50.000 m3 per day and became greater to 126.000 m3 since September 2006. It is flowing continuously, oozing out and spreading so that it has now covered more than 400 Ha of productive land and has completely immersed many villages around the mining area as shown in figure 1 (a) and 1 (b).

(a) (b)

Figure 1. (a). The mud has hampered productive land in Porong City. (b) The mud has immersed many villages

1 Master in Civil Engineering, Concrete and Construction Materials Laboratory, Lecturer in Civil Engineering Department, Faculty of Civil Engineering and Planning, Institute Technology of Sepuluh Nopember, Surabaya, Indonesia (e_mail: [email protected]) 2 Professor in Material Mechanics, Concrete and Construction Materials Laboratory, Lecturer in Civil Engineering Department, Faculty of Civil Engineering and Planning, Institute Technology of Sepuluh Nopember, Surabaya, Indonesia (e_mail: [email protected])

13 CMT - 91


CMT -

The spreads of this mud also threatened the ecosystem of thousands fishponds, hampered the infrastructure, economical, education and social life.Many efforts have been applied to minimize its volume and to stop its flow. And some geologists have proposed a hypothesis about it, that a mud volcano is oozing out and that the efforts to stop the outpouring of mud is impossible. An alternative effort to spillway mud to the sea has become a controversial issue. Another alternative is using the mud on a large scale to use it for construction material. One of the efforts is to use its solid form as a base material for geopolymer concrete. Some real efforts have been done so far, which is creating bricks and paving blocks. To make use of this material as a simple concrete, it is mixed with a polymer liquid and ordinary portland cement. Chemical analysis of this mud shows that it consists of materials such as alumina and silicate, which could be used as a binder in concrete if it is mixed by activator solution. Concrete without Portland cement is known as geopolymer concrete. This concrete usually uses fly ash as a base material. Research has shown that the performance of geopolymer concrete produces high compressive strength and greater durability than ordinary concrete. It is also well known that geopolymers possess excellent mechanical properties, as well as fire and acid resistance (Ta-Wui Cheng, 2003) and (Yodmunee, Sarawut and Yodsudjai, Wanchai, 2006). This geopolymer paste, induced by high-alkaline solutions, can be used as a binder to produce concrete instead of the ordinary cement paste. The aims of this research are to analyse certain chemical composition of this mud in order to use it for construction materials, mix mud with fly ash and activator solution in an appropriate composition to get the binder and make geopolymer concrete base on the selected binder

2. PREVIOUS RESEARCH

Very limited research data is currently available in published literature. Most of the past research on the behavior of geopolymeric material was based on the binder paste formed from fly ash using only small laboratory scales and samples. The chemical composition of geoplymer is amorphous (Davidovits, J, 1999). The silicone and the aluminium atoms in the source materials are induced by alkaline solutions to dissolve and form a gel to be a binder. The gel binds the aggregates and other un-reacted materials together to form geopolymer concrete. Some test results show that the compressive strength of geopolymer concrete does not vary with age. Longer curing, molar H2O-to Na2O and water content in the concrete play vary important role in the improvement of higher compressive strength (Hardjito, D., Wallah, S. E. and Rangan, B. V., 2004) and (Hardjito, D., Wallah, S. E. and Rangan, B. V., 2005 Earlier work by the author (Ekaputri, J.E and Triwulan, 2006) reported the effect of various parameters such as curing time, sodium silicate to sodium hydroxide ratio and diameters of coarse aggregate. Some preliminary research has been done on a small scale to assess the mechanical properties of geopolymer concrete based on wet, dry, and calcified mud mixed with a fly ash-alkaline activator solution. Some suitable compositions of the binders have been created and show good potential. Wet mud mixed with an alkaline activator does not have any potential as a binder: it can only be used as a binder paste when mixed with a fly ash and alkaline activator solution. Good performance has been demonstrated by wet, dry and calcified mud mixed with a fly ash-alkaline activator solution, but wet mud mixed with ordinary Portland cement and lime has very poor performance.

13 CMT - 92


CMT -

3. EXPERIMENTAL WORK

3.1 Materials

In the experimental work, the solid mud form was obtained from Porong, Sidoarjo, Indonesia. It was used as the filler in the binder composition. Fly ash, as a base material was obtained from Jawa Power Paiton, Indonesia. Tabel 1 shows the chemical composition of dry mud and fly ash. Figure 2 shows the samples of wet mud and fly ash.

Table 1. Composition of Dry Mud and Fly Ash as Determined by XRD (% mass) MATERIALS SiO2 Al2O3 Fe2O3 CaO Na2O K2O TiO2 MgO P2O5 SO3 SO2 LOI Microstructure

Mud 53.08 18.27 5.6 2.07 2.97 1.44 0.57 2.89 - - 2.96 10.15 Crystallin

Fly Ash 52.24 38.58 2.94 0.69 0.52 0.44 2.42 0.49 0.13 1.21 - 1.39 Amorphous Data was taken from “Balai Besar Keramik”, Bandung, Indonesia

Figure 2. (a). Strained Solid Mud (b) Fly Ash

Majority compound of the dry mud consists of Kaolin and Feldspar, so it can be used as ceramics materials. Alkaline activator contains of sodium silicate solution (Na2O = 8.5%, SiO2 = 28.5% and H2O = 63 %) and flake form of sodium hydroxide.

3.2 Binder

Experimental work has been done to find the binder which is used to make geopolymer concrete. Some compositions of fly ash to wet, dry or calcified mud ratio by mass and the molarity of alkaline activator have been tried in small size of sample. The binder consists of fly ash mixed with alkaline activator, as the main binder, and wet or dry mud as the filler since the microstructure of the mud is crystallin. All of the samples are compared to the base binder: pure fly ash mixed with alkaline activator. Each mixture needs 223 grams of fly ash. The activator liquid-to-fly ash ratio was kept constantly at 0.35.Wet or dry mud is added in vary ratio of mass fly ash. Mud and fly ash were mixed drily in a bowl mixer for 2 minutes. The alkaline activator was added to the paste, and mixed together for another 2 minutes. In order to avoid the hardening process, which happened too fast, the paste should not be mixed in the long time. The paste was then prepared to cast in the 2 x 4 cm cylinder fiber moulds and left in ambient temperature. After 24 hours, the samples were then removed

13 CMT - 93


CMT -

START

ALL MATERIALS ARE WEIGHTED THE AGGREGATES TO BINDER RATIO BY MASS = 7:3 or 3:1 BINDER: 1. FLY ASH 2. STRAINED SOLID MUD. MUD MASS = 0.5 x FLY ASH MASS 3. ALKALINE ACTIVATOR AGGREGATE IN SATURATED SURFACE DRY CONDITION 1. 0.5 -1 cm OR 1 - 2 cm COARSE AGGREGATE 2. FINE AGGREGATE

MATERIALS ARE MIXED IN SEQUENCE: 1. MUD AND COARSE AGGREGATE WERE MIXED TOGETHER 2. FLY ASH 3. ALKALINE ACTIVATOR 4.FINE AGGREGATE

CAST THE PASTE 1. THE MIXTURE WAS CAST IN 10 X 20 cm CYLINDER MOULDS IN THREE LAYER 2. EACH LAYER RECEIVED 10 MANUAL STROKES

1. THE SPECIMENS WERE LEFT IN 24 HOURS 2. THE SPECIMENS WERE REMOVED FROM THE MOULD AND CURED IN ROOM TEMPERATURE FOR 4 -5 DAYS

THE SPECIMENS WERE LEFT TO DRY AIR UNTIL LOADED IN COMPPRESSION AT THE SPECIFIED AGE

FINISH

from the moulds and then cured in a waterproof cover for 2 days. At the end of the curing period, the specimens were left at room temperature until loaded in compression at the specified age. The best performance of the binders created, would be chosen to find a fixed paste composition for the production of mud-based concrete.

3.3 Geopolymer Concrete

The binder chosen was then mixed with coarse and fine aggregate to make geopolymer concrete. Two types of locally aggregate were used and compared: 0.5 – 1 cm mixed with fine sand and 1 – 2 cm aggregate mixed with sand. Some compositions of mass ratio binder to aggregate have been tried in 10 x 20 cm cylinder samples. All of the samples are compared to the base mixing: pure fly ash mixed with alkaline activator and aggregates. Due to its substantial fast process, the chemical reaction of the binder gel, the mixture should not be mixed in more than 3 minutes. Figure 3 shows the sequence of geopolymer concrete casting and figure 4 shows the specimens.

Figure 3. Process casting in creating mud-based geopolymer concrete

13 CMT - 94


CMT -

(a)

(b)

Figure 4. (a) Geopolymer Concrete Mixture in the Steel Moulds (b) Concrete Specimens

3.4 TEST RESULTS

In this paper, the effect of various parameters on the compressive strength of both the binder and the geopolymer concrete are reported. Each of the test data points plotted in various graph corresponds to the mean value of the compressive strength of two test cylinder specimens.

3.4.1 Compressive Strength of The Binders at Different Ages

Details of the mixture composition used in the binders are given in Table 2. and Table 3 shows the other mixtures used in the binder.

Table 2. Detail of the Binder Mixture Composition

BINDER CODE

Molarity of NaOH

Solution

Sodium Silicate/NaOH

Solution by Mass

Fly Ash/Mud by Mass

Additional Materials

D 14 M 2.5 1:1 - E 14 M 2.5 1:1 5 % Lime

F 14 M 2.5 1:1 5 % Lime + 5% Portland cement

G 14 M 2.5 - Pure calcified mud with 60 mL distilled water

I 14 M 2.5 0.75:0.75

Table 3. Detail of the Binder Mixture Composition

BINDER CODE

Molarity of NaOH Solution


Solution by Mass

Fly Ash/Mud by Mass

Additional Materials

A 14 M 2.5 Pure Fly Ash - B 14 M 2.5 1:1 - H 14 M 2.5 0.5:0.5 Dry Mud 30 mL water K 8 M 1.5 1:0.5 -

13 CMT - 95


CMT -

Figure 5 and 6 show the compressive strength of The Binders in different Ages. In Figure 5, it is found that wet mud mixed with flay ash and ordinary Portland cement or lime has very poor performance and mass ratio fly ash: wet mud = 1: 1 likewise. Only calcified mud mixed with alkaline activator given by Binder G shows a good potential. Figure 6 shows that good performance has been demonstrated by wet and dry mud mixed with a fly ash-alkaline activator solution. The compressive strength of the binder does not vary with age. Fly ash to dry mud ratio by mass equal to 0.5:0.5 shown as Binder C gives good performance. And fly ash to wet mud ratio by mass equal to 1:0.5 shown as Binder B is sufficient to achieve satisfactory results. Thus, Binder B as suitable compositions was then chosen as a binder to make geopolymer concrete.

0 MPa

5 MPa

10 MPa

0 days 10 days 20 days 30 days

Age

Com

pres

sive

Stre

ngth

(MPa

)

BINDER D

BINDER E

BINDER F

BINDER G

BINDER I

Figure 5. Compressive Strength of The Binder at Different Ages

13 CMT - 96


CMT -

0 MPa

5 MPa

10 MPa

15 MPa

20 MPa

25 MPa

30 MPa

0 days 10 days 20 days 30 days

Age

Com

pres

sive

Stre

ngth

(MPa

)

BINDER A

BINDER B

BINDER H

BINDER K

Figure 6. Compressive Strength of Binder at Different Ages

3.4.2 Compressive Strength of Geopolymer Concrete at Different Ages

The details of the mixture composition used in the geopolymer concrete are given in Table 4.

Table 4. Detail of the Concrete Mixture Composition

CONCRETE CODE Molarity of

NaOH Solution


Solution by Mass

Diameter of Aggregate

Curing Time

Binder/Aggregate by Mass

CONCRETE A 14 M 2.5 1 – 2 cm - 1:3 CONCRETE C 14 M 2.5 1 – 2 cm 4 days 1:3

CONCRETE D (Pure Fly Ash) 14 M 2.5 1 – 2 cm 4 days 1:3 CONCRETE E 14 M 2.5 0.5 – 1 cm 4 days 1:3 CONCRETE F 14 M 2.5 0.5 – 1 cm 5 days 3:7 CONCRETE G 8 M 1.5 0.5 – 1 cm 5 days 3:7 CONCRETE H 8 M 2.5 0.5 – 1 cm 5 days 3:7

Figure 7 shows some effects of the size of aggregate, curing time, molarity of alkaline activator and the binder to aggregate ratio by mass on the compressive strength. The compressive strength of Concrete E, which has smaller diameter of aggregate, is higher than Concrete C. The addition of wet mud equal to 50% mass of fly ash, the workability was considered as desirable. It is still very hard to cast the mixture in the short time and it takes time to harden and dry it. At the early age of concrete, the more curing time is done, the higher the compressive strength improved. The effect of the binder to aggregate ratio by mass

13 CMT - 97


CMT -

and molarity of alkaline activator plays an important role in the concrete compressive strength, so does the age. Concrete F which has higher molarity of alkaline activator and binder to aggregate ratio by mass, shows the best performance, besides Concrete D, which contains pure fly ash mixed with alkaline activator in 14 Molar.

0

5

10

15

20

25

30

35

0 10 20 30 40Age (Days)

Com

pres

sive

Stre

ngth

(MPa

)

CONCRETE A

CONCRETE C

CONCRETE D

CONCRETE E

CONCRETE F

CONCRETE G

CONCRETE H

Figure 7. Compressive Strength of Geopolymer Concrete at Different Ages

4. CONCLUSIONS

Based on the experimental results reported in this paper, the following conclusions are drawn: 1. Wet and dry mud are not amorphous materials so that the mixture in geoplymer concrete needs

higher alkalinity (in molarity) to bind un-reactive materials so that they can only be used as a filler.

2. The compressive strength of the binder does not vary with the age of mixture. It depends on the addition of wet mud and the other material, such as lime and Portland cement. The addition of lime or Portland cement to the mixture gives very poor performance.

3. Longer curing time improves the polymerization process resulting in higher compressive strength of geopolymer concrete, because of the addition of wet mud equal to 50% mass of fly ash, it takes time for concrete to be hardened or dry. Thus the age influences concrete performance. It is still difficult to carry out the research result in the field.

4. An increase in binder to aggregate ratio increases the concrete compressive strength In order to optimise the usage of the mud for the community it is essential to carry out research and analysis both in the laboratory as well as in the field. The use of the research findings would be beneficial not only for the development of research of geopolymers in Indonesia, but also for the community in Porong Sidoarjo, who are the victims of the hot

13 CMT - 98


CMT -

mud disaster, as it would create opportunities for employment using the hot mud as a base building material.

5. REFERENCES

Davidovits, J.(1999). “Chemistry of Geopolymeric System, Terminology.” Proceeding of Geopolymer ’99 International Conference, France

Ekaputri, J.E and Triwulan (2006). “The Usage of Porong Mud as a Construction Materials.” National Conference, Surabaya

Hardjito, D., Wallah, S. E. and Rangan, B. V. (2004). “On The Development of Fly Ash Based Geopolymer Concrete.” ACI Material Journal 101, 467-472

Hardjito, D., Wallah, S. E. and Rangan, B. V. (2005). “Factors Influencing The Compressive Strength of Fly Ash-Based Geopolymer Concrete.” Dimensi Teknik Sipil

Ta-Wui Cheng (2003). “Fire Resistant Geopolymer Produced by Waste Serpentine Cutting.” Proceedings of the 7th International Symposium on East Asian Resources Recycling Technology, Taiwan, 2003

Yodmunee, Sarawut and Yodsudjai, Wanchai (2006). “Study on Corrosion of Steel Bar in Fly Ash-Based Geopolymer Concrete.” International Conference on Pozzolan, Concrete and Geopolymer, Thailand, 2006

13 CMT - 99


CMT -

A STUDY OF COMPOSITION OF C-S-H GEL IN CEMENT PASTE

Pipat Termkhajornkit1 and Toyoharu Nawa2

ABSTRACT : C-S-H gel is the main product from the hydration reaction of cement. It is necessary to study about its compositions. In this study, the cement paste compositions during the hydration were estimated by the XRD and analyzed by the Rietveld analysis. The combined water was measured by ignited samples at 950 °C. A CaO/SiO2 molar ratio and bonding water in the C-S-H gel were calculated. The results showed that at the early age the CaO/SiO2 ratio increased with time and then became constant at later age while the H2O/SiO2 ratio in the C-S-H gel was almost constant. When the water to cement ratio decreased from 0.30 to 0.26, there is little change of the composition of C-S-H gel. However, when the water to cement ratio increased from 0.30 to 0.46, the composition of C-S-H gel changed. KEYWORDS: C-S-H gel, Ca(OH)2, CaO/SiO2, H2O/SiO2, water to cement ratio

1. INTRODUCTION

C-S-H gel is the main component from the reaction of cement. It is also the main component from the reaction between the calcium hydroxide solution and pozzolanic materials such as silica fume, fly ash, slag, and metakaolin. The C-S-H gel plays very important role in concrete. The density of C-S-H gel is lower than that of cement so that porosity decreases as the reaction of cement progresses. It also acts like a glue to bind aggregates together. As a consequence, concrete develops strength and permeability decreases. C-S-H gel has been noticed more than a century. Nevertheless, the structures and the compositions of C-S-H gels are still an interesting issue for cement and concrete researchers. The present study was designed to investigate the effect of water to cement ratio on the composition of C-S-H gel. The compositions of cement pastes during hydration were estimated by the XRD and analyzed by the Rietveld analysis. The combined water was measured by ignited sample at 950 °C. CaO/SiO2 and H2O/SiO2 molar ratio in C-S-H gel were calculated.

2. EXPERIMENTAL

2.1 Materials and preparing conditions Ordinary Portland cement was used. The chemical compositions of cement and the phase compositions were determined by the XRF and the XRD and further analyzed the Rietveld analysis respectively. The chemical and physical properties of cement are shown in Table 1. The cement was mixed with water within a 2-liter pan mixer at low speed for 90 seconds and further mixed at high speed for 90 seconds. Flow value was measured according to JASS15 M-103. The flow diameter was controlled to be 200 to 250 mm by adjusting the amount of Polycarboxylate-based superplasticizer. As for the hydration before 3 days, specimens were prepared with water to cement (w/c) ratio 0.30 by weight and cured with a seal condition. The specimens were analyzed at 6, 12, 18, 24, 48 and 60 hours after mixing.

1 Post doctor, Graduate School of Engineering, Hokkaido University, Japan, [email protected] 2 Professor, Graduate School of Engineering, Hokkaido University, Japan, [email protected]

14 CMT - 100


CMT -

Table 1. The chemical and physical properties of cement

As for the hydration from 3 to 182 days, specimens were prepared with w/c ratio 0.26 and 0.30 by weight. At 24 hours after mixing, the specimens were demolded and then cured in water until required age. The specimens were analyzed at 7, 28, 56, 91, 182 days after mixing. In addition, the results were compared to another set of specimens, which was prepared for other purposes. These specimens were prepared with w/c ratio 0.46 by weight and cured with seal condition. 2.2 Examination methods At required age, the specimens were crashed into 2.5 to 5.0 mm and soaked in acetone to stop the hydration reaction. Next the specimens were further dried at 105°C. Then they were ground in a disc mill. Particles smaller than 75μm were used in analysis. The total nonevapolable water was measured from the weight loss of specimen ignited from 105 to 950°C. The compositions and the hydration reaction of hydrating specimens were estimated by the XRD-Rietveld analysis (Termkhajornkit, et al., 2005a, 2005b and 2006). As for the XRD-Rietveld analysis, Cukα X-ray diffraction equipment was used. The experiments were carried out in the range of 5-70°2θ with 0.02 step scan and 1.00 s/step speed. Divergence slit, scattering slit and receiving slit were 1/2°, 1/2° and 0.3 mm respectively. The SIROQUANT version 2.5 was used as the software for the Rietveld analysis.

3. RESULTS

3.1 Hydration of cement The amount of each compound in the unhydrated cement during the hydration reaction can be estimated by the XRD-Rietveld analysis. Figures 1 and 2 show the amount of each hydrated compound for the specimens before and after 3 days, respectively. The rapid hydration of alite can be seen before 3 days. After 3 days, the hydration reaction still gradually increased. In contrast, the hydration of belite hardly occurred at early age, however, its hydration increased slowly after 3 days.

0102030405060708090

100

6 12 18 24 48 60

Age ( hours)

Am

ount

of h

ydra

ted

parti

cles

(%

by w

eigh

t)

AliteBeliteFerriteAluminate

0102030405060708090

100

7 28 56 91 182

Age (days)

Am

ount

of h

ydra

ted

parti

cles

(%

by w

eigh

t)

Aluminate Ferrite Belite Alite

Figure 1. The amount of each compound in

unhydrated cement during hydration reaction before 3 days.

Figure 2. The amount of each compound in unhydrated cement during hydration reaction after 3

days.

C3S C2S C3A C4AF SiO2 Al2O3 Fe2O3 CaO MgO Na2O K2O TiO2 MnO (%) (m2/kg) (kg/m3) (%) (%) (%) (%) (%) (%) (%) (%) (%) (%) (%) (%) (%)0.77 347 3150 63.09 12.99 11.78 9.23 20.84 5.95 2.62 63.63 1.79 0.18 0.33 0.34 0.10

Blaine surface area

Phase compositions ( by mass)Ignition loss

Chemical compositions ( by mass)Density

14 CMT - 101


CMT -

As for the hydration of aluminate phase, its hydration was almost complete before 6 hours after mixing so that the change of reaction was hardly detected. The hydration of ferrite phase was over expectation. This is because a mass absorption coefficient of Fe for the X-rays of CuKα is considerably high (Termkhajornkit, et al., 2006). The amount of Ferrite would not be used in the analysis of this study. The total hydrated cement as a function of age and w/c ratio is shown in Figure 3. It confirms the effect of w/c ration on the hydration reaction; cement hydrated faster when the w/c ratio increased from 0.26 to 0.46. 3.2 Hydration products Not only unhydrated cement, hydration products such as Ca(OH)2 and C-S-H gel can be estimated likewise. The amount of the total unhydrated cement, the Ca(OH)2 and the C-S-H gel of the specimens before and after 3 days are elucidated in Figures 4 and 5, respectively. The rapid change was observed before 24 hours after mixing. The amount of the unhydrated cement decreased while those of the Ca(OH)2 and the C-S-H gel increased. The slops of all graphs suddenly changed after 24 hours. In Figure 5, the unhydrated cement gradually decreased while the C-S-H gel increased, however, the Ca(OH)2 was almost steady. This means the production of the Ca(OH)2 stopped before that of the C-S-H gel. The reaction of alite and belite may produce the Ca(OH)2 and the C-S-H gel at early age, but produced only the C-S-H gel at later age. Figure 6 shows the effect of w/c ratio on the amount of Ca(OH)2. The amount of Ca(OH)2 increased as the increase of w/c ratio. Figure 7 demonstrates the C-S-H gel as a function of w/c ratio. The amount of C-S-H gel of specimens prepared with w/c ratio 0.26 was lower than others. The difference between those prepared with w/c ratio 0.30 and 0.46 was small. The amount of Ca(OH)2 and C-S-H gel as functions of the hydrated calcium silicate phases are expressed in Figure 8. There is a certain relation between the C-S-H gel and the hydrated calcium silicate phases, regardless w/c ratio. As for the Ca(OH)2, however, the relation between the Ca(OH)2 and the hydrated calcium silicate phases was effected by the w/c ratio. When the w/c ratio changed from 0.26 to 0.30, the Ca(OH)2 slightly increased. However, when the w/c ratio changed from 0.30 to 0.46, the Ca(OH)2 distinctly increased.

60

65

70

75

80

85

90

95

0 50 100 150 200

Age (days)

Hyd

rate

d ce

men

t (%

by

wei

ght)

W /C 0.26

W/C 0.30

W/C 0.50

Figure 3. The total hydrated cement as a function of age and w/c ratio.

Figure 4. The amount of total unhydrated cement, Ca(OH)2 and C-S-H gel before 3 day.

0102030405060708090

0 20 40 60 80

Age ( Hours)

Am

ount

(% b

y w

eigh

t)

Ca(OH)2

Unhydrated cement

C-S-H gel

14 CMT - 102


CMT -

0102030405060708090

0 50 100 150 200

Age ( days)

Am

ount

(% b

y w

eigh

t)C-S-H gel

Unhydrated cement

Ca(OH)2

0

5

10

15

20

25

30

0 50 100 150 200

Age (days)

Ca(

OH

) 2 (%

)

W /C 0.26

W/C 0.50

W/C 0.30

Figure 5. The amount of total unhydrated cement,

Ca(OH)2 and C-S-H gel after 3 day.

Figure 6. The effect of the w/c ratio on the amount of Ca(OH)2.

60

65

70

75

80

85

0 50 100 150 200Age (days)

C-S

-H g

el (%

)

W /C 0.26

W/C 0.30

W/C 0.50

0102030405060708090

100

40 50 60 70 80Amount of hydrated calcium silicate phases(%)

Am

ount

(%)

W /B0.26 W/B 0.30 W/B0.50

C-S-H gelCa(OH)2

Figure 7 The C-S-H gel as a function of the w/c

ratio.

Figure 8. The amount of Ca(OH)2 and C-S-H gel as function of hydrated calcium silicate phases.

It should be noted that ettringite and monosulfonate can also be estimated from XRD-Rietveld analysis since both of them are crystals. However, they will not be discussed in this paper. 3.3 Compositions of the C-S-H gel The CaO/SiO2 molar ratio of the C-S-H gel can be estimated by following. The calcium oxide supplied by C3S and C2S was consumed to generate the Ca(OH)2 and the C-S-H gel. Since the amount of Ca(OH)2 was known from XRD-Rietveld analysis, by subtracting the calcium oxide in Ca(OH)2 from the calcium oxide provided by dissolution from C3S and C2S, the amount of calcium oxide in C-S-H gel can be estimated. On the other hand, the silica oxide contributed by C3S and C2S was supplied to produce hydrated gel only. Since the amount of C3S and C2S in hydrated samples are known, CaO/SiO2 can be estimated. The nonevapolable water in the C-S-H gel can be estimated by following. The total ignition loss of paste is equal to the sum of chemical bonding water in Ca(OH)2, ettringite, monosulfate hydrate and C-S-H gel. Since the Ca(OH)2, the ettringite and the monosulfate hydrate are crystals, the amount of chemical bonding water in these crystalline compounds can be stoichiometrically calculated. After subtract the chemical bonding water in the crystalline phases from the total bonding water, the amount of nonevapolable water in C-S-H gel can be estimated. The relation between the CaO/SiO2 molar ratio at early age and time is shown in Figure 9. The CaO/SiO2 molar ratio increased until 18 hours and then became constant. Odler and Dörr, 1979,

14 CMT - 103


CMT -

reported that the C-S-H phase formed in the induction period differs from that formed later on: it has a lower CaO/SiO2 molar ratio. Normally, the induction period of the cement reaction stops around 3 to 5 hours after mixing. Therefore, the change of CaO/SiO2 molar ratio was expected to stop around 3 to 5 hours. However, as shown in Figure 9, the CaO/SiO2 molar ratio still continually increased until 18 hours. This extension can be explained by the effect of superplasticizer. In this study, the superplasticizer was used to control a rheology. It seemed that addition of superplasticizer extended the induction period. In general the end of induction period can be noticed by the increase of temperature in specimens. Figure 10 shows the temperature rises of specimens. Compared with the specimens prepared without superplasticizer, those prepared with superplasticizer showed a delay of temperature rise. The temperature started to increase after 18 hours, which meant that the induction period finished at 18 hours. In Figure 9, the CaO/SiO2 molar ratio increased until age 18 hours. After that, it became almost constant. Therefore, the change of the CaO/SiO2 molar ratio at early age found in this study is in fact similar to those found by Odler and Dörr , 1979. Figure 11 shows the relation between the H2O/SiO2 molar ratio at early age and time. In contrast with the CaO/SiO2, the H2O/SiO2 molar ratio before 18 hours decreased as age increased and then became almost constant. Similar results were again reported by Odler and Dörr, 1979. It seemed that there is the relation between CaO/SiO2 molar ratio, H2O/SiO2 molar ratio. Figures 12 and 13 show the CaO/SiO2 and the H2O/SiO2 molar ratio at long period as functions of time. The results showed that the CaO/SiO2 molar ratio decreased when the w/c ratio increased. The CaO/SiO2 molar ratio of specimens prepared by the w/c ratio 0.30 was slightly lower that those prepared by the w/c ratio 0.26. However, when the w/c ratio changed from 0.30 to 0.46, the CaO/SiO2 molar ratio markedly decreased. The effect of the w/c ratio on the CaO/SiO2 molar ratio agrees well with those reported by Locher, 1967. In contrast, the H2O/SiO2 molar ratio increased when the w/c ratio increased. In fact these results are similar to those observed in early age; it seemed that there is balance or relation between the CaO/SiO2 and the H2O/SiO2 molar ratio.

4. CONCLUSIONS

1. The results confirm the effect of the w/c ratio on the hydration reaction; cement hydrated faster when the w/c ratio increased.

0

0.5

1

1.5

2

2.5

3

0 1 2 3Age (days)

CaO

/SiO

2 m

olar

ratio

W/C 0.30

Figure 9. The relation between the CaO/SiO2 molar ratio at early age and time.

Figure 10. The temperature rises of specimens.

15

20

25

30

35

40

0 1 2 3Age (days)

Tem

pera

ture

(o C

) with superplasticizer

without superplasticizer

14 CMT - 104


CMT -

0

12

3

45

6

7

89

10

0 1 2 3

Age (days)

H2O

/SiO

2 m

olar

ratio

W/C 0.30

Figure 11. The relation between H2O/SiO2 molar ratio at early age and time.

Figure 12. The CaO/SiO2 molar ratio at long period as functions of time.

Figure 13. The H2O/SiO2 molar ratio at long period as functions of time.

2. There is a certain relation between the C-S-H gel and the hydrated calcium silicate phases, regardless the w/c ratio. In contrast, the relation between the Ca(OH)2 and the hydrated calcium silicate phases was affected by the w/c ratio. 3. The CaO/SiO2 and the H2O/SiO2 molar ratio of C-S-H gel varied until end of induction period. 4. The composition of C-S-H gel at later age was affected by the w/c ratio. 5. It quit likely that there is relation between the CaO/SiO2 and the H2O/SiO2 molar ratio of C-S-H gel.

5. REFERENCES

Locher, F. W. (1967). “Zement-Kalk-Gips.” 20, 402. Odler, I. and Dörr H., (1979). “Early hydration of tricalcium silicate I. Kinetics of the hydration

process and the stoichiometry of the hydration products.” Cem. Concr. Res. [online], (9), 239-248, Available from: < http://www.sciencedirect.com>

Termkhajornkit, P., Nawa, T. and Kurumisawa, K. (2005a). “A study of fly ash-cement hydration by Rietveld analysis and selective dissolution”, JCI, Vol. 27,pp. 169-174.

Termkhajornkit, P., Nawa, T. and Kurumisawa, K. (2005b). “Quantitative study on hydration of fly ash and Portland cement.” In: N. Banthia, T. Uomoto, A. Bentur and S.P. Shah, Eds.

0

0.5

1

1.5

2

2.5

0 50 100 150 200

Age (days)

CaO

/SiO

2 m

olar

ratio

W/C 0.30

W/C 0.26

W/C 0.50

00.5

11.5

22.5

33.5

44.5

5

0 50 100 150 200

Age (days)

H2O

/SiO

2 m

olar

ratio

W/C 0.26

W/C 0.30

W/C 0.50

14 CMT - 105


CMT -

Proceedings of ConMat’05, Vancouver, 399. Termkhajornkit, P., Nawa, T. and Kurumisawa, K., (2006). “ Effect of water curing conditions on the

hydration degree and compressive strengths of fly ash-cement paste.” Cement and Concrete Composites, in press.

14 CMT - 106


CMT -

NO-FINES CONCRETE FOR LOW COST MASS HOUSING

Kardiyono1

ABSTRACT: Indonesian population is already more than 200 millions, where one of the crucial problems to be solvent is to fulfill the very high demand of housing. As the consequence is high demand of material used and require a method to speed up the construction, yet to keep the cost as low as possible. In the process, the cost of wall can be minimized by using local material and the speed can be increased by developing a non-traditional method. One of the solution for fulfill this requirement is by using no-fines concrete technology for constructing the wall. Although the technology has already been popular in overseas a quite sometimes ago, no or rare house was yet to be built in Indonesia using this technology. According to Moss (1979). the first no-fines concrete houses were built in the Netherlands after World War I with crushed clinker as the aggregate. Later about 50 houses were built in Scotland using the same material. In 1937 the Scottish Special Housing Association was set up to provide work for unemployment coal miners. As the labor was inexperienced in construction, traditional methods of building were discharged in favor of the no-fines system, but the construction manager decided to use crushed whinstone-plenty-fully available there. About 900 houses were built by this method in Scotland before World War II and these are still in good condition. It is noted that no-fines concrete system needed minimum of labor and yet without sacrificing design standard to build 2 to even 10 stories of homes or apartment. Based on this background, some researches of no-fines concrete have been carried out in the Department of Civil Engineering, Gadjah Mada University. The researches were emphasized on the application of the plenty-unused local material such pumice in Bawuran and cinder aggregate in the area around Merapi Volcano in Yogyakarta for the application in low cost mass housing.

The tests results showed that the most suitable no-fines concrete made of the above material can be achieved if uniform gradation with the maximum aggregate of 20 mm was applied. The water cement ratio was 0.4 and the cement content was around 150 kg to 300 kg per cubic meter of no-fines concrete. Using this mix proportion a compressive strength of up to 18 MPa can be achieved and the specific gravity is around 1.8. To see its applicability on site, a prototype of house funded by Sleman Government has also been build. It can be concluded that no-fines concrete is superior to the conventional brick masonry. No-fines concrete is cheaper and easy to be constructed that can speed up the construction.

KEYWORDS: No-fines concrete, pumice, cinder aggregate, low cost mass housing.

1. INTRODUCTION

Indonesian population is already more than 200 millions, where one of the crucial problems to be solved is to fulfill the very high demand of housing. As the consequent there are high demand of material to be used and requires a method to speed up the construction, yet to keep the cost as low as possible. In the process, the cost of wall can be minimized by using local material and the speed can be increased by developing a non-traditional method.

On the other side, there are many natural aggregate which not yet been exploited, for example rock pumice in Pleret, Bantul district, south of Yogyakarta, and stone of volcanic cinder in Merapi Volcano inside, north-side of Yogyakarta. The aggregate unused because its strength is low, disqualification to make the normal concrete.

1 Lecturer, Civil Engineering Department, Engineering Faculty, Gadjah Mada University, Yogyakarta, Indonesia.

15 CMT - 107


CMT -

One of the solutions to fulfill this requirement is by using no-fines concrete technology from local aggregate that not be used yet, for constructing the wall. Although this technology has already been popular in overseas a quite sometimes ago, no or rare house is yet to be built in Indonesia.

2. REVIEW OF LITERATURE STUDIES

Macintosh et al., (1956) explained that no-fines concrete might be defined as a concrete from which the fine aggregate is almost, if not entirely, omitted. The aggregate is of a single size, usually 20 mm in Britain. In this paper it is explained that the type of concrete can be used for both external and interior partition walls, and it is normal for all window and doorframes, beams, and other fixtures that are to form part of the wall to be fixed between the forms and the no-fines concrete to be poured around them. The concrete is usually made with water-cement ratios 0.35 to 0.50 and the aggregate-cement ratios 6 to 10. The compressive strength among 7 – 14 MPa, and the weight about 1500 kg per cubic meter.

According to Moss (1979). the first no-fines concrete houses were built in the Netherlands after World War I with crushed clinker as the aggregate. Later about 50 houses were built in Scotland using the same material. In 1937 the Scottish Special Housing Association was set up to provide work for unemployment coal miners. As the labor was inexperienced in construction, traditional methods of building were discharged in favor of the no-fines system, but the construction manager decided to use crushed whinstone-plenty-fully available there. About 900 houses were built by this method in Scotland before World War II and these are still in good condition. It is noted that no-fines concrete system needed minimum of labor and yet without sacrificing design standard to build 2 to even 10 stories of homes or apartment.

Based on this background, some researches of no-fines concrete have been carried out in the Department of Civil Engineering, Gadjah Mada University. The researches were emphasized on the application of the plenty-unused local material such pumice in Bawuran village, Bantul District, Yogyakarta Special Province, and cinder aggregate in the area around Merapi Volcano in Yogyakarta for the application in low cost mass housing.

3. REVIEW OF PREVIOUS RESEARCH RESULTS

In this section, some research, which has done in laboratory, will be submitted successively.

3.1. Water/cement ratio In the year 1992 a research of about no-fines concrete with the aggregate from ceramic tile fraction has been done ( Kardiyono, 1992). The tile fraction which the size between 5 until 40 mm and specific gravity 1.8 and also its water absorption 17 percent has been made for no-fines concrete with the aggregate-cement ratio by volume were varies from 6 to 10, and water-cement ratio by weight were varies from 0.3 until 0.46. From this research is obtained that the specifics gravity of the no-fines concrete are varies 1.4 until 1.7, the porous is varies between 20-25 percent, and the compressive strength were varies between 4 MPa and 10 MPa, with the maximum compressive strength at the water-cement ratio around 0.38-0.42 (see Figure 1). This matter is similar with Raju (1983) guide, that water-cement ratio range from 0.37 and 0.45.

3.2. Pumice aggregate In the year 2000 a research to study of no-fines concrete that is made from pumice aggregate obtained from Bawuran Village, southern Yogyakarta, was done (Sulistiyowati, 2000). The aggregate’s specific gravity is 1.20 and the water absorption is 35 percent. With this aggregate that the size 5-20 mm and the aggregate-cement ratio by volume varies from 2 until 10 and water-cement ratio 0.4 was

15 CMT - 108


CMT -

obtained that the no-fines concrete weight is about 1600 – 1700 kg per cubic meter, the pore volume is varies among 1.6 - 3.0 percent, and the compressive strength is varies between 7-20 MPa (see Figure 2). Its elasticity modulus is varies about 3,200-10,000 MPa.(see Figure 3). Cement requirement is varies 173 -593 kg per cubic meter concrete (see Figure 4) depend on the aggregate-cement ratio . The pore volume is small and the cement requirement is great because its aggregate is brittle.

0,34 0,36 0,38 0,4 0,42 0,44 0,46 0,48

Water/cement ratio

0

2

4

6

8

10

12

Note:A/C=aggregate/cement ratio

Com

pres

sive

str

engt

h ( M

Pa)

2 4 6 8 1Aggregate/cement ratio by volume

30

25

20

15

10

5

0

Cinder aggregate(Subkhannur,2002)

Pumice aggregate(Sulistyowati,2000)

Co

mp

ress

ive

stre

ng

th (

MP

a)

Figure 1. Relationship of Water/Cement Ratio and Figure 2. Relationship of Aggregate/Cement Compressive Strength (Kardiyono,1992) Ratio and Compressive Strength

2 4 6 8 10

Aggregate/cement ratio by volume

0

5

10

15

20

25

30

35

Ela

stic

ity

mod

ulus

(10

00 M

Pa)

2 4 6 8 10

A ggregate/cement rat io by volume

0

100

200

300

400

500

600

Cinder aggregate(Subkhannur,2002)

Pumice aggregate(Sulistyowati,2000)

Figure 3. Relationship of Aggregate/Cement Figure 4: Relationship of Aggregate/Cement Ratio and Elasticity Modulus and Cement Requirement

3.3. Cinder aggregate In the year 2002 a research to study of no-fines concrete that is made from cinder aggregate from Merapi volcano, northern Yogyakarta city, was done (Subkhannur, 2002). Result of this aggregate inspection obtained that its specific gravity is 2.06 with the water absorption 10 percent, and hardness is 29 percent (with Rudeloff tube tester) and abrasion value is 70 percent (with Los Angeles machine tester). This aggregate is uniform size of 10 - 20 mm. The volume aggregate-cement ratio is varies from 2 until 10 and water-cement ratio is 0.4. The weight of no-fines concrete obtained is varies from 1790 – 2260 kg per cubic meter, porosity volume is varies among 1.06 – 25.67 percent, and the compressive strength is varies between 5 - 32 MPa (see Figure 2). elasticity modulus is varies about 7,500 – 25,000 MPa (see Figure 3). Cement Portland requirement is varies 136 - 605 kg per cubic meter of concrete (see Figure 4). The variation is depended on the value of the aggregate-cement ratio.

15 CMT - 109


CMT -

3.4. Moment strength and ductility of reinforced no-fines concrete beam A research to obtain moment strength and ductility of reinforced concrete beam from no-fines concrete with the aggregate of expanded clay (shale) have been done by Tjokrodimuljo (1995 ). No-fines-concrete is made with the aggregate-cement ratio 4 and water-cement ratio 0.40. Weight of the concrete is 1600 kg per cubic meter, and the compressive strength is 15 MPa. The concrete beam specimens are width 150 mm, height 200 mm, and length 2000 mm (span length 1800 mm and tested with two points load apart 600 mm). Five beams are with upper and lower reinforcements of equal size, which are 2D8, 2D10, 2D12, 3D12, and 3D16. The test result is indicated that the concrete beam have the moment strength which is equal to ordinary concrete beam (by compressive strength 15 MPa.). that are maximum elastic moment successively 4.9 kN-m, 7.4 kN-m, 12.5 kN-m, 18 kN-m, 26 kN-m, and maximum plastics moment are 5.1 kN-m, 7.7 kN-m, 14 kN-m, 20 kN-m, 30 kN-m (see Figure 5). The value of beam ductility is great enough. It can be seen that the greater area of steel reinforcement the greater stiffness of the concrete beam, this matter because of low concrete elasticity modulus, so that the influence of steel reinforcement is very dominant.

3.5. Shear strength of reinforced no-fines concrete beam A research to obtain shear strength of reinforced concrete beams from no-fines concrete with the aggregate of expanded clay (shale) has been done by Basewed (1997). At this research the experiment concrete beams are width 150 mm, height 200 mm, and length 2000 mm ( span length 1800 mm and tested with two points load apart 600 mm) with the longitudinal steel reinforcement 4D16, as tensile and compressive reinforcement. The specimens are ten concrete beams. Five first beam with the shear steel reinforcement of the diameter 6 mm with the distance 50 mm, 75 mm, 100 mm, 125 mm, and 150 mm. As for five second beam with the steel reinforcement is the diameter 8 mm with the distance varies, equal to the five first beams. The yield strength of the steel are 320 MPa and 420 MPa for 6 mm and 8 mm steel diameter successively. The compressive strength of no-fines concrete is 15 MPa. Result of this research is indicated that the shear force can be supported by the no-fines concrete beams are among 30-48 kN (that are 55-80 percent of theoretically analysis for normal concrete). Observation during experiment indicated that the damage type is brittle (see Figure 6). Thereby, no-fines concrete beam has to be prevented from shear crack or needed by the way to increase the shear strength.

0 10 20 30 40 50 Displacement at mid span of beams (mm)

30

25

20

15

10

5

0

tensile steel reinforcement: 3D16compr. steel reinforcement: 3D16

3D12

2D12

2D10

2D8

50

40

30

20

10

0

She

ar fo

rce

(kN

)

Displacement at mid span (mm)

Note :1 : specimen 12 : specimen 23 : specimen 34 : specimen 45 : specimen 5

Figure 5. Moment-Displacement Curves at Mid Figure 6. Shear Force-Mid Span Displacement Span of The Five Experimental No-Fines Curves of Beams With 6 mm Ties Concrete Beams (Tjokrodimuljo,1995) (Basewed and Kardiyono, 1997)

3.6. Reinforced no-fines concrete column

15 CMT - 110


CMT -

A research to study the properties of reinforced no-fines concrete column has been done by Setiyawan (2002). The both end parts of column experiments are made from a normal concrete in order not to destroy effect of examination. Fifteen columns with the dimension 150 mm x 150 mm and length 2000 mm, additionally head both its back part with the wide dimension 450 mm, length 200 mm, have been made. Five given by the first column group of longitudinal reinforcement 4D8, five of the second column group 4D10, and five third column group 4D12. All columns given by the shear reinforcement with the ties diameter is 6 mm with the distance 160 mm. Examination done with the axial load with the eccentricity distance varies for each specimen tested, that are 75 mm, 150 mm, 225 mm, 300 mm, and 375 mm. The testing of no-fines concrete indicated that the concrete compressive strength is 9 MPa., the modulus of elasticity is about 13.300 MPa, and the strain at maximum load is about 0.0009. The the yield of reinforcing steel are 293 MPa, 403 MPa, and 361 MPa for the steel of diameter 8 mm, 10 mm, and 12 mm successively. From this research, its can be seen that load-displacement relationship diagram is equal to theoretic analysis result (see Figure 7). Column fractures are occurred by concrete compressive failure.

0 5 10 15 20 25

0

5

10

1

5

Axi

al lo

ad (k

N)

Mid Column Displacement (mm)

Experiment

Theoretical

Figure 7. Axial Load – Mid Column Displacement Curves For Column With 4D8 Reinforcement and 300 mm Eccentricity (Setiyawan,2002)

3.7. Reinforced no-fines concrete wall Research to obtain the behavior of no-fines concrete wall was done by Baasir (2005). In this research, the walls tested are thickness 100 mm, length 3000 mm, and high 3000 mm (see Figure 8). The bottom side of wall set by a horizontal beam of size 150 mm x 300 mm (with the longitudinal reinforcing steel of upper side to the 2D10 and of lower side of 3D14). In upper side of the wall is given a horizontal beam 100 mm x 100 mm (with the reinforcing steel 4D8). and both wall besides given by the column of size 100 x 100 mm (with the reinforcing steel 4D8). The specimens are two no-fines concrete walls. The first wall is without reinforcing steel, and the second wall is with horizontal reinforcing steel. The steel bars are diameter 6 mm with the vertical distance 200 mm. The weight of the no-fines concrete wall is 1500 kg per cubic meter, compressive strength is 4.45 MPa, tensile strength is 0.248 MPa, and initial elastic modulus 1600 MPa. Tensile test of reinforcing steel indicated that yield stress is 281 MPa, maximum tensile stress is 398 MPa, and the modulus of elasticity is 194983 MPa.

15 CMT - 111


CMT -

loading frame

hydraulics jack

load cell

linear gauge

NO-FINES CONCRETE WALL

concrete beam (100X100)

concrete column (100X100)

3000 mm

3000 mm

rigid floor

concrete beam (150X300)

Figure 8. Arrangement of No-Fines Concrete Wall Experiment (Baasir, 2005)

These walls are tested with static horizontal load at side to the unidirectional of wall area. Results of these wall tests are can be seen in Figure 9 and Table 1. From these data it can be seen that first crack on 2.972 kN for wall without reinforcing steel and 15.716 kN for wall with reinforcing steel.

0 1

0 2

0 3

0 4

0 5

0 6

0 7

0To

p ho

rizo

ntal

load

(kN

)

0 3 6 9 12 15 18 21

Top w all displacement (mm)

No-fines concrete wall with reinforcing steel

No-fines concrete wall without reinforcing steel

Figure 9. Top Horizontal Load – Top Wall Displacement Curves of No-Fines Concrete Wall

(Baasir, 2005)

Table 1. Experiment Result of No-fines Concrete Wall (Baasir, 2005).

No-fines Concrete Wall without Reinforcing Steel

No-fines Concrete Wall With Reinforcing Steel

Horizontal Load (kN)

Crack width (mm) Horizontal Load (kN) Crack width (mm)

First Crack 2.972 0.230 15.716 0.510

Yield Condition 22.268 2.172 51.909 1.662

Maximum Load 28.851 9.045 59.978 3.740

Fracture Load 25.175 17.379 47.983 6.299

15 CMT - 112


CMT -

4. FIELD EXPERIMENTS IN YOGYAKARTA AREA After these researches have finished in laboratory, then field experiments are carry out. Antecedent attempt in this field executed near the aggregate quarry, were in Bawuran Village, and in Purwobinangun Village. Both villages are in Yogyakarta Special Province (see Figure 10). Bawuran Village is in the location of pumice quarry, southern of Yogyakarta city (see Figure 11). Around this countryside there are some mounts of pumice, that are about 25 million meter cubic (Widiasmoro et al., 1993). The other location is in Purwobinangun Village, northern of Yogyakarta city. Purwobinangun Village is in the slanting side of Merapi Volcano, close to Boyong River where the volcanic cinder is taking over (see Figure 12). There is volcanic cinder aggregate about 750 thousand cubic meter (Setiawan, 2003).

Centre Jawa Province

Yogyakarta City

Merapi Volcano

Yogyakarta Special ProvinceOcean

Borobudur Temple

Prambanan Temple

Purwobinangun Village

Bawuran Village

10 km

Sleman

Bantul

Figure 10. Map of Yogyakarta Special Province

Figure 11. Pumice Hill in Bawuran Village. Figure 12. Boyong River Near Purwobinangun Village in

The Slanting Side of Merapi Volcano

The first attempt was to train local community to make no-fines concrete block with length 400 mm, width 200 mm, and thickness 100 mm (see Figure 13, Figure 14, and Figure 15). The second attempt was to train the local community to make a wall from no-fines concrete block (see Figure 16). The third attempt was to train the local community for constructing no-fines concrete column and no-fines concrete wall for making low cost mass housing (see Figure 17 and Figure 18). Later the activity was to train the local community for practicing this method for constructing a low cost mass house from no-fines concrete (see Figure 19).

15 CMT - 113


CMT -

Figure 13. Training Community in Bawuran Figure 14. Pouring No-fines Concrete in to

Village Concrete Block Mould

Figure 15. Concrete Block of (400 mm x Figure 16. Wall That is Made From No-fines Concrete

200 mm x 100 mm) Size Block

Figure 17. No-fines concrete column and wall Figure 18. Column and wall of low cost mass

during construction housing have finished

15 CMT - 114


CMT -

Figure 19. Prototype of Mass House that was Made From No-fines Concrete in Sleman

5. CONCLUSIONS AND COMMENTARY

By these antecedent attempts had been done in the field can be elaborated local community in the location of aggregate intake can accept. The process of constructing wall from no-fines concrete with pour in place is easier, quicker, and cheaper than traditional method that is it constructed from red brick. Making wall from no-fines concrete can replace the ordinary society habit that make the wall from red brick. The replacement will stop the damage process of rice field because of the red brick is made of rice field land, and in the other side, using of materials that not be used before is produced. Next attempt is socialize of newly method is being developed to low cost house contractors in Special Province of Yogyakarta and Indonesian country.

6. ACKNOWLEDGEMENTS

The author acknowledges many supporters from Academic Staff, Students, and Laboratory Technicians of Gadjah Mada University. These laboratory researches and field attempts were carry out with Gadjah Mada University Students in Research Program and in Community Service Activity Program, and supported by Local Government Regency. Partly author fund aid was accepted from Portland Cement Factory of “Semen Nusantara”. Grateful thanks are to: Lecturers, Students, and Technicians of Engineering Material Laboratory, in Civil Engineering Department of Engineering Faculty of Gadjah Mada University, The Government of Bantul District, The Government of Sleman District, The Community of Bawuran Village and Purwobinangun Village, Portland Cement Factory of “Semen Nusantara”, and whosoever assisted these attempts to develop of no-fines concrete in Indonesia.

15 CMT - 115


CMT -

7. REFERENCES

Baasir,H.(2005). “Characteristic of Post Elastics of No-fines Concrete Wall with Horizontal Reinforcing Steel under Static Loading.” Thesis, Masters Program, Civil Engineering Department, Engineering Faculty of Gadjah Mada University, Yogyakarta.

Basewed, F., and Kardiyono (1997). “The Influence of Shear Reinforcement to The Strength of No-fines Concrete Beam with Expanded Shale Aggregate”, Forum Teknik, Volume 21, No.2. Juli 1997, Technological Journal, Engineering Faculty of Gadjah Mada University, Yogyakarta.

Kardiyono (1992). “No-fines Concrete made from Aggregate of Ceramic Tile Rubble.” Research Report, Engineering Faculty, Gadjah Mada University, Yogyakarta.

Mcintosh,R.H, Bolton,J.D., Muir,C.H.D. (1956). “No-fines Concrete as A Structural Material.” Proceedings of the Institution of Civil Engineers, Paper No. 6136, London.

Moss, J.K.(1979). “No-fines Building Gives Energy-Conserving Homes, A System that Saves Both Energy and Cement.” International Construction Sutton, Surrey, England, Publication#C790123, Copyright©1979,The Aberdeen Group.

Raju.,N.K.(1983). “Design of Concrete Mixes.” Second Edition, College Book Store Publishers & Distributors, Delhi.

Setiawan,Y.R.(2003). “Geology of Purwobinangun Area of Pakem sub-District of Sleman District of Yogyakarta Special Province and Calculation of Reserve of Resource Cinder Aggregate.” Research Report, Research Program and Community Service Activity Program, Gadjah Mada University, Yogyakarta.

Setiyawan,P.(2002). “Behavior of No-fines Concrete Column under Eccentric Axial Load.” Thesis, Masters Programme, Civil Engineering Department, Engineering Faculty of Gadjah Mada University, Yogyakarta.

Subkhannur,A.(2002) “Cinder Aggregate from Merapi Volcano as Aggregate in Making of No-fines Concrete.” Final Report, Civil Engineering Department of Gadjah Mada University, Yogyakarta.

Sulistiyowati,E.E.(2000). “Use of Pumice of Size Measure 5mm-20mm as Aggregate in Making No-fines Concrete.” Final Report, Civil Engineering Department of Engineering Faculty of Gadjah Mada University, Yogyakarta.

Tjokrodimuljo,K. (1995). “Strength and Ductility of Reinforced Concrete Beam from No-fines Concrete with Expanded Shale Aggregate.” Media Teknik, No.1. Year of XVII of April Edition 1995 ISSN No:0216-3012, Four Monthly Magazine of Engineering Faculty of Gadjah Mada University, Yogyakarta.

Widiasmoro, Tjokrodimuljo,K., and Fatimah,S.(1993). “Study On Petrology, Potency, and Application of Pumice Conglomerate in Piyungan, Yogyakarta, for Basic Raw Material for Light Bricks and Light Tile”, Challenging The Frontier Geology in Indonesia, The Twenty Second Annual Convention Indonesian Association of Geologists (IAGI). December 6-9, 1993, Bandung

15 CMT - 116


CMT -

”JALAN BATU BETON” A STRONG AND LOW COST ROAD

USING A SUBSTANDARD CONCRETE MATERIAL

Abdul Basyit

ABSTRACT : Substandard concrete material is easily available in most villages in Indonesia while standard quality material is becoming rare and expensive. This paper is focusing in utilizing this type of material more efficiently in appropriate application to produce a strong and high quality road using as minimum as possible cement to support a friendly and sustainability drive promotion.

Literature study to back up this paper is carried out to see any possibility to widely use this kind of material in Indonesia and apply it in road constructions , to produce a strong , flood proof and bad drainage conditions as well as low in price using non import paving material. Limited field tests had been carried out to validate assumptions which is taken in this paper. KEYWORDS : low cost, strong, low cement content, waste material, monolithic

1. PREFACE

Substandard concrete material is easily available in most village or aggregate quaries in Indonesia while standard quality material is becoming rare and expensive. For most parts of Indonesia, concrete is still regarded as expensive surfacing material even though its advantage and performance has been widely accepted. A breakthrough should be obtained especially for rural area. Starting from that perspective , I try encourage myself to combine both theory obtained from literature with a limited field tests with a great hope that this paper will help to enriched the material technology in Indonesia road construction.

Maintenace is major cost component in road funding,. Building concrete road will be the solution, since concrete road had been accepted as having good characteristics in its withstanding for bad drainage and flood conditions. Even though Indonesia is oil exporting country, but still importing asphalt for its road maintenace system.Asphalt is still mayor road material in Indonesia which absorb government budget for road sector especially for maintain the road network in good condition. This paper is started with a cover of one of references , which write down an efforts to get a inexpensive road in Cambodia where concrete road is one of it solutions. The aims of this efforts is to lift poverty in some “neglected area”. This LCS paper is result of cooperation between the ILO Upstream Project Cambodia and the Low Cost Road Surfacing Initiative: “Low-cost, Labor-based Roads for Poor Communities”.

If all governments realizing such noble efforts, and cement base road is widely utilized, Insyaallah with a better road in villages and small town, rural communities poverty could be reduced substantially. If road maintenance cost could be drastically reduced, more strong new road could be built.

16 CMT - 117


CMT -

In many developing and emerging economies and regions, many public road network is still unpaved. Many of these road are not surfaced. However, where provided, the usual constructed running surface for these roads is selected natural gravel or cementitious materials. These naturally occurring materials are usually excavated from pits or quarries and hauled by trucks or tractors and trailers to be laid on the previously shaped formation or road surface, watered to achieve a suitable moisture content and compacted . To obtained a strong and durable road surfacing this LCS paper propose and tests several alternatives where one of it is concrete road , utilizing bamboo reinforcement as substitute for ordinary steel reinforcement. Starting from this LCS paper, my paper will focus on developing a concrete road which costs lower than what is written in .this LCS paper If steel could be replaced with bamboo reinforcement so the remaining target is lowering cement content to reduced the price of concrete road.

2. MAIN DIFFERENCE BETWEEN ASPHALT ROAD AND CONCRETE ROAD.

In asphalt road , the strength of the structure are built on layers of sub base under asphalt layers. Structurally this kind of road is called flexible, since it’s flexural ability to retains its original position under imposing load.

On concrete road, because of its rigidity the strength of the system is forms by the concrete layer itself. The concrete acts as a bridge over sub grade and the load is distributed to a wider area than asphalt road. Concrete road could be build direct on soil or on sand, gravel or low-grade concrete sub-base.

Maintenance of asphalt road should be carried out regularly to prevent water penetrating into underlying sub-grade. Penetrating water thru cracks will degrade the supporting capability of sub-grade layers.

The following picture is example of neglected maintenance which lead to road user big losses and nightmare where loss of time, money and cost of spare-part are very high. Asphalt is a byproduct of crude oil, an organic material where it’s quality will degrade upon time, asphalt road is getting brittle after some time and could not withstand flood or bad drainage. On other hand cement is processed from minerals and concrete is

gaining more strength over life time and more susceptible to flood. Choosing concrete road for village and rural community is a solution for long-lasting road system which is not only using locally produced material but also strong and able to withstand rainy country conditions such in Indonesia.

16 CMT - 118


CMT -

3. LITERATURE STUDY.

The following chapter is taken from “Thickness design for Concrete highway and street pavement “ Published by Canadian Portland Cement Association

Design of Concrete Pavement with Lean Concrete Lower Course.

Following is the thickness design procedure for composite concrete pavements incorporating a lower layer of lean concrete, either as a sub base constructed separately or as a lower layer in monolithic construction Lean concrete is stronger than conventional sub base material and is considered to be nonerodable. Recognition of its superior structural properties can be taken by a reduction in thickness design requirements. Analysis of composite concrete pavements is a special case where the conventional two layer theory (single slab on foundation) is not strictly applicable. The design procedure indicates a thickness for a two-layer concrete pavement equivalent to a given thickness of normal concrete. The equivalence is based on providing thickness for a two-layer concrete

pavement that will have the same margin of safety* for fatigue and erosion as a single-layer normal concrete pavement. In the design charts, Figs. B1 and B2, the required layer thicknesses depend on the flexural strengths of the two concrete materials. Since the quality of lean concrete is often specified on the basis of compressive strength it has to be converted to an estimated flexural strength (modulus of rupture) for use in preliminary design calculations.

Lean Concrete Sub base

The largest paving use of lean concrete has been as a Sub base under a Conventional concrete pavement. This is non monolithic construction where the Surface course of normal concrete is placed on a hardened lean concrete sub base. Usually, the lean concrete sub base is built at least 600 mm wider than the pavement on each side to support tile tracks of the slip form pavers. This extra width is structurally beneficial for wheel loads applied at pavement edge.

The normal practice has been to select a surface thickness about twice the sub base thickness; for example, 220 mm of concrete on a 100 or 120 mm sub base. Fig. B1 shows the surface and sub base thickness requirements set to be equivalent to a given thickness of normal concrete without a lean concrete sub base. A sample problem is given to illustrate the design procedure. From laboratory tests, concrete mix designs have been selected that give modules of rupture of 4.5 and

16 CMT - 119


CMT -

2.0 Mpa,* respectively, for the surface concrete and the lean concrete sub base. Assume that a 260-mm thickness requirement has been determined for a pavement without lean concrete sub base by the procedures set forth in Chapter 3 or 4 of this book. As shown by the dashed example in 10 in Fig. B1, designs equivalent to the 260-mm pavement are: (1) 185- mm concrete on a 130 - mm lean concrete sub base, and (2) 200-mm concrete on a I 00-mm lean concrete sub base.

Monolithic Pavement

In some areas, a relatively thin concrete surface course is constructed monolithically with a lean concrete lower Layer. Local or recycled aggregates can be used for the lean concrete, resulting in cost savings and conservation of high-quality aggregates. Unlike the lean concrete sub bases discussed in the previous section, the lower layer of lean concrete is placed at the same width as the surface course, and the joints are sawed deep enough to induced full depth cracking through both layers at the joint locations. Fig. B2 is the design chart for monolithic pavements. To illustrated its use, assume that the design strength of the two concretes are 4.5 and 2.0 Mpa, and the design procedures of chapter 3 and 4 indicates the thickness requirement of 250 mm for full depth normal concrete. As shown by dashed example line in fig. B2, monolithic design equivalent to the 250-mm pavement are (1) 100-mm concrete surface on 220 mm lean concrete, or (2) 80-mm surface on 245 lean concrete. The criteria are that (1) stress ratios in either of the two concrete layers not exceed those of reference pavement; and (2) erosion values at the sub base-sub grade interface not exceed those of the reference pavement. Rationale for the criterion is given in reference 56 plus two additional considerations 1)erosion criteria are included in addition to the fatigue approach given in reference; and (2) for non monolithic construction, some structural benefit is added because the sub base is constructed wider than pavements. ** Flexural strength of lean concrete to be used as sub base is usually selected to be between 1.0 to 1.7 Mpa (Compressive Strength, 5.2 to 8.3 Mpa) These relative low strengths are used to minimize reflective cracking from unjointed sub base through the concrete surface. From the above Publication, we could draw 2 important points which seldom used in Indonesia.

1. The thickness of lean concrete sub base could as thick as 300 mm while the surface layer could be as thin as 80 mm.

2. Local or recycled aggregates can be used for the lean concrete, resulting in cost savings and conservation of high-quality aggregates

3. The lean concrete sub base is constructed monolithically with the surface layer.

16 CMT - 120


CMT -

STANDARD CONCRETE AND CONCRETE WITH SUBSTANDARD MATERIAL.

When concrete is mixed with a composition of cement, water, fine and coarse aggregate, a sieve analysis is provided to obtained an economical and workable mixture is assumed as normal or standard concrete in this paper. A usual concrete road use at least K350 concrete to get good and long lasting road. This mixture will require cement amount which in some area will contributing the high cost of concrete road. To reduced the cement amount, sieve analysis is a must to obtained the smallest empty area between aggregate which should be filled with cement and water. Good graded aggregate should be produced by sieving and mixing aggregates. In villages and small town where batching plant is unavailable, it is difficult to educate private users to use the good aggregate and they prefer to use an all in aggregate to make concrete road even for concrete structure.

• On road with relatively high traffic , Concrete with nonstandard material is ordinary concrete (K350) with intrusions of big stones to reduce the amount of concrete itself thus reducing cement content drastically in every cubic meter of concrete road.

• On rural area and housing with low number of traffic, concrete with non standard material is concrete which is made from “ all-in aggregate” which is obtained directly from quarry and being added with bigger stone to reduce the amount of concrete..

• On the very extreme side, a test of road using concrete with non standard material use building waste for its aggregate.

Concrete with substandard material explained in this paper could be used for base and sub base in asphalt road and concrete substitution for ordinary concrete road. From the nature of concrete with substandard material, it could be classified as econocrete. In econocrete the material could be from recycle road material, low quality concrete aggregate or recycle concrete aggregate. This type of material could be use as sub base of asphalt road or concrete road.

16 CMT - 121


CMT -

Since this type of road material consist concrete and stone , writer called it “ Stone Concrete Road” which in Indonesia language it is translated into “ Jalan batu beton” This type of road is closer to “Telford road” than to concrete road. On very simple Telford system the road consist of big stones and the space between it is filled with smaller stone to lock the movement of big stones.. In Jalan batu beton, instead of using small stone we use concrete to fill the gap between the big stone. For practical and to ease the tests, concrete quality is made similar for both layer so both layer was poured in one go. The only difference is that the lower layers contains stone to reduce the volume.

5. FIELD TEST.

Base on the above literature study, a field test was carried out in public road with a high volume of traffic but with low axle load and ICCI area. The public road is use to simulate high traffic city road with low load and ICCI area to simulate village road. To simulate village concrete work, the test in ICCI area, the concrete is hand mixed. For public road we used ready-mix using K350 concrete quality.

The test use 50 % of stone by volume. This means that the volume of concrete is also reduced up to 50 % .The test also shows that for public road with high traffic, lowering the concrete quality lead to excessive abrasion of its surface. All tests shows that in 4 years no significant distressed occurred. Similar work is done in housing area in Bekasi West Java. The road is also made with high intrusion of big stone.

6. CONCRETE ROAD USING WASTE MATERIAL.

When we deal with concrete road for rural area and measures to lift the poverty, we always facing the budget problem. There are still many road even in Jakarta using gravel as surfacing, in housing area , rural, villages and slump areas. Usually this type of road bear a very low load. Mostly 1 ton axle cars and once in a week a truck collecting garbage. In this following test writer try to use waste material from housing renovation and construction work.Usually such material is only used for filling material. Since this type of waste consist of sand, concrete ruble and brick. After screening , an all in sand -aggregate and boulder size material are obtained. The all-in aggregate are corrected with natural sand to get better performance, while the big size material is broken into pieces smaller than side form size. This big size material is use to reduce the concrete volume.

16 CMT - 122


CMT -

The result is a concrete road with performance similar to ordinary concrete road .

This test opens a good possibility of using waste material for concrete road to support the green concrete drive, promoting the use of waste with minimum cement content in concrete. Concrete road has very long life. The first concrete road in America is still in use after more than 100 years of service.These tests were carried out 4 years for concrete road with stone intrusions and less than 6 months for concrete road with cement base waste material. This span of time is far beyond the age of first concrete road in America. However this test was done with a great hope that other research and test could be continued to obtained a low cost concrete road dedicated to urban and poor communities.

7. WHY THIS IS IMPORTANT FOR POVERTY REDUCTION.

If concrete road could be built in larger scale, transportation could easily reached poor community and enable it to improve trading. Concrete road could be made with simple screed and is possible to use ordinary mason with a little training. For villages and kampongs, this work could be made thru “gotong royong” scheme, where people work together unpaid to improve their infrastructures system. With more concrete road built, less fund for road maintenance is needed and this means that more new road could be built.

16 CMT - 123


CMT -

Lifting poverty means more job is opened and village trading is getting heated. Reducing the cost of concrete road will increase the possibility of constructing new concrete in wider area. So it is our duty to support such drive. From internet publication, India has very intensive village concrete road system. It is stated that in Sept 2000 , 50 % of road connecting villages all over India has been concretizes. This number comprising total 1 million kilometers of road connecting villages.

8. CONCLUSION.

• It is possible to make low cost concrete road by using Jalan batu beton. • Although it is low in price, the strength still be maintained. • Reduction of cement content in concrete road will support sustainability drive. • This type of road also support using waste cement base building material if use for roads with

very low traffic. • More tests, more research are still needed for exploring further the use and reliability of

this type of road. If Indonesia could take the benefit from other country experiences in using concrete road combine with efforts to reduce the cost. A brighter hope will emerge from villages and poor community of possibility in having good , strong and low cost road system. Insyaallah.

9. REFERENCES

Department of Army. US Army (1988). “Substandard Material For Pavement Construction.” Canadian Portland Cement Association. “Thickness Design For Concrete Highway And Street

Pavement.” Perie, B. “Low –Volume Concrete Road.” Cement and Concrete institute. ACI (1997). “ACI Manual Of Concrete Inspection.” ACI Publication. Portland Cement Association. “Soil Cement Construction Handbook.” “Bamboo Reinforced Concrete Pavement ,Road Construction In Cambodia.” LSC Working Paper

No.7. “Paving The Way For Rural Development & Poverty Reduction.” LCS Working Paper No. 12.

16 CMT - 124


CMT -

LIGHTWEIGHT STYROFOAM CONCRETE FOR LIGHTER AND MORE DUCTILE WALL

Iman Satyarno1

ABSTRACT : A Solid clay-brick unit masonry or brick masonry has been commonly used in buildings as partitions or architectural accessories. There are three major problems arise in the application of brick masonry wall. Firstly, a serious environmental impact has currently occurred in some places in Indonesia due to the over exploitation of rice field soil for brick production. A long term impacts should be anticipated that might cause a reduction of rice production. Secondly, the unit weight of brick masonry wall, which is around 1700 kg/m3, can be considered high. Its self weight might introduce forces in terms of gravity forces and seismic forces for buildings in the high seismic zones. Thirdly, brick masonry wall is very brittle unless it is reinforced. Therefore, it is necessary to find alternative materials for masonry wall to be used in buildings. Researches carried out in the Department of Civil Engineering Gadjah Mada University had shown that lightweight Styrofoam concrete might be the solution to overcome the above problems. This is because the unit weight of lightweight Styrofoam concrete can be made as low as 330 kg/m3 or around 0.2 unit weight of brick masonry. The full scale tests of 3m by 3m wall as prototype of brick masonry walls and lightweight Styrofoam concrete walls have also been carried out. The test results showed that the lateral strength of both walls are almost the same but the lightweight Styrofoam concrete wall was lighter and showed a more ductile behaviour than the one of brick masonry wall.

KEYWORDS: Brick, brittle, ductile, hysteretic loops, masonry, seismic force, self weight.

1. INTRODUCTION Brick masonry wall is very popular as wall material in Indonesia. The brick was commonly made of soil taken from rice field. Currently serious environmental impact has occurred in some places in Indonesia such as Java due to the over exploitation of rice field soil for brick production. A long term impacts should be anticipated that may cause a reduction of rice production as shown in Figure 1 [Satyarno (2004)].

1 Senior Lecturer, Department of Civil Engineering, Gadjah Mada University, Indonesia.

Figure 1. Environmental Impact on Rice Fields as the Soil was Used for Brick Industry.

17 CMT - 125


CMT -

The unit weight of brick masonry wall, which is around 1700 kg/m3 or around 250 kg/ m2 can be considered high. Its self weight might introduce forces in terms of gravity forces and seismic forces for buildings in the high seismic zones. In some places, the price of land has already increased very high so that the requirement of high rise buildings cannot be avoided. In this case, self weight effect becomes dominant to the developed forces in the structural elements [Satyarno (2005)]. Besides its self weight that is considered high, brick masonry is also categorized as brittle material. In the recent Yogyakarta earthquake on 27 May 2006, it can be seen some damaged buildings. It is believed that the damaged was participated by the high self weight as can be seen in Figure 2. Had the walls were built using lighter material but had the same strength, the damaged could be reduced.

Therefore, it is necessary to find an alternative material that has a compressive strength close to the common brick masonry but has a lighter self weight and has more ductile behaviour.

2. STYROFOAM CONCRETE Styrofoam is a trademark for a light plastic material which is known as a form of foam polystyrene packaging which is widely used for packaging electronic items. Polystyrene itself is produced from styrene (C6H5CH9CH2), which has phenyl groups (six-member carbon ring) attached in random locations along the carbon backbone of the molecule. The random attachment of benzene prevents the molecules from becoming highly aligned. As a result, polystyrene is an amorphous, transparent, and somewhat brittle plastic [Crawford (1998)].

If the form of the Styrofoam is granular, it can be used as aggregate in the concrete. However, it is noted here that the density of Styrofoam is very low that is only around 15 to 20 kg/m3. Due to this low density, the Styrofoam in the concrete can be considered as air or foams. The advantage of using Styrofoam rather than using air entranced in the aerated concrete is the Styrofoam has a tensile strength. Therefore beside it will reduce the concrete density, it can also work as fibre that can improve the concrete mechanical behaviour. In the application, concrete density can be adjusted by controlling the amount of Styrofoam mixed in the concrete. The more the Styrofoam used in the concrete, the lower density of concrete will be achieved. Nonetheless, the lower achieved compressive strength of concrete should be anticipated and be taken into account.

Lightweight Styrofoam concrete can be made from the mix of water, cement (white or grey), sand, and Styrofoam grain as can be seen in Figure 3. As mentioned above the unit weight and the mechanical properties of lightweight Styrofoam concrete can be adjusted based on the amount of Styrofoam grain used in the mix. Styrofoam grain unit weight is only around 15 kg/m3, which is much lower than the other common building materials unit weight. For example, the approximate unit weight of brick is 1200 kg/m3, mortar is 2000 kg/m3 and concrete is 2400 kg/m3.

The physical and mechanical properties of lightweight Styrofoam concrete depend on the following parameters [Satyarno (2004)]: 1) amount of cement per cubic meter,

Figure 2. Wall Self Weight on Building may Participate the Cause of Damage due to Earthquake.

17 CMT - 126


CMT -

2) volumetric ratio of sand and Styrofoam in the mix. The physical and mechanical properties of lightweight Styrofoam concrete with white cement can be explained as follows.

2.1 Unit Weight of Styrofoam Concrete Unit weight or weight per cubic meter of Styrofoam concrete can be seen in Table 1 and Figure 4. As a comparison, common unit weights of brick masonry of 1700 kg/m3 is also plotted in Figure 4. It can be seen that the unit weight of Styrofoam concrete can be made to be lower than that of brick masonry unit weight if the Styrofoam content is more than 20%. For more Styrofoam content, the unit weight will be much lower that can reach only 330 kg/m3.

2.2 Compressive strength of Styrofoam Concrete Compressive strength of Styrofoam concrete can be seen in Table 2 and Figure 5, where usual compressive strengths of common bricks in Yogyakarta markets are also plotted in the figure. It can be seen that the compressive strength of Styrofoam concrete is close to the compressive strength of brick for the Styrofoam content of 80% or a little bit less than that. If the Styrofoam content is much lower than 80%, the compressive strength of Styrofoam concrete is much higher than the compressive strength of brick.

Table 1. Unit Weight of Styrofoam Concrete (kg/m3)Styrofoam Cement Content

Content 250 kg/m3 300 kg/m3 350 kg/m3 400 kg/m3

100% 330 459 497 58980% 800 844 845 91360% 1094 1199 1179 125540% 1457 1476 1455 153620% 1635 1824 1841 18230% 2040 2089 2109 2167

Figure 3. Materials for Styrofoam Concrete.

Water Cement

Styrofoam grain Sand

17 CMT - 127


CMT -

It is noted here that the price of Styrofoam grain is quite expensive. To minimize the price, Styrofoam waste that is commonly used such as for electronics or fruits packaging can be applied [Musana (2006)]. In the application Musana cut the Styrofoam waste to become dices with dimension of 10 mm as shown in Figure 6. The dices were then used as a replacement of Styrofoam grain shown in Figure 3. Test results of unit weight and compressive strength of Styrofoam concrete using this Styrofoam waste are shown in Figures 7 and 8. It can be seen that while the unit weight of the Styrofoam concrete using Styrofoam grain is close to the one of Styrofoam concrete using Styrofoam waste, the compressive strength is quite different. The compressive strength of Styrofoam concrete using Styrofoam waste significantly drops for the volumetric Styrofoam waste content of 20%. This compressive strength is quite constant for more volumetric content of Styrofoam waste and drops again for volumetric content of 100%.

Figure 4. Unit Weight of Styrofoam Concrete.

0

500

1000

1500

2000

2500

0 20 40 60 80 100Volumetric Styrofoam content (%)

Uni

t wei

ght (

kg/m

3 )

Amount of cement = 250 kg/m3Amount of cement = 300 kg/m3Amount of cement = 350 kg/m3Amount of cement = 400 kg/m3Common brick masonry unit weight

Figure 5. Compressive Strength of Styrofoam Concrete.

0

5

10

15

20

25

30

0 20 40 60 80 100

Volumetric Styrofoam content (%)

Com

pres

sive

stre

ngth

(MPa

) Amount of cement = 250 kg/m3Amount of cement = 300 kg/m3Amount of cement = 350 kg/m3Amount of cement = 400 kg/m3Bricks in Yogyakarta markets

Table 2. Compressive Strength of Styrofoam Concrete (MPa)Styrofoam Cement Content

Content 250 kg/m3 300 kg/m3 350 kg/m3 400 kg/m3

100% 0.35 0.74 0.91 1.2380% 2.09 2.38 2.61 4.4660% 4.36 5.79 6.78 8.0240% 5.87 5.53 9.71 13.0120% 10.70 9.66 13.32 18.480% 12.09 15.49 21.63 28.00

17 CMT - 128


CMT -

3. LIGTWEIGHT STYROFOAM CONCRETE PRODUCTS The lightweight Styrofoam concrete can be formed to be as follows: 1) concrete block [Satyarno (2004)], 2) concrete panel [Darmawan (2004)], 3) direct pour [Satiawan (2005)]. The concrete block production can be explained as follows, see Figure 9.

Figure 6. Styrofoam Waste Dices with Dimension of 10 mm for Styrofoam Concrete.

Figure 7. Unit Weight of Styrofoam Concrete Using Styrofoam Waste.

0

500

1000

1500

2000

2500

0 20 40 60 80 100Volumetric Styrofoam content (%)

Uni

t wei

ght (

kg/m

3 )

Amount of cement = 250 kg/m3Amount of cement = 300 kg/m3Amount of cement = 350 kg/m3Common brick masonry unit weight

Figure 8. Compressive Strength of Styrofoam Concrete Using Styrofoam Waste.

0

5

10

15

20

25

0 20 40 60 80 100

Volumetric Styrofoam content (%)

Com

pres

sive

stre

ngth

(MPa

) Amount of cement = 250 kg/m3Amount of cement = 300 kg/m3Amount of cement = 350 kg/m3Bricks in Yogyakarta markets

17 CMT - 129


CMT -

1) Prepare the concrete block mold or a parallel concrete block mould, where the mould can be made of wood or steel.

2) Fill the mould with Styrofoam concrete and try to compact it. 3) The block can be lifted after one day and can be used for construction after seven days. The concrete panel production can be explained as follows, see Figure 10. 1) Prepare the concrete panel mold, where the mould can be made of wood or steel. 2) Fill the mould with Styrofoam concrete and try to compact it. 3) The panel can be lifted after one day and can be used for construction after seven days.

The direct pour production can be explained as follows, see Figure 11. 1) Prepare the reinforcement according to the design. 2) Put the door or window frames on their position if any. 3) Prepare the form work, where it is better to limit the height is around 1 m. 4) If the form work is ready, pour the lightweight Styrofoam concrete in to it and try to carry out

compacting using stick of steel, wood or bamboo. 5) After one day (it would be better after three days), the form work can be dismantled. 6) Arrange the form work for another higher position. 7) Repeat the above procedure until the required height is met, and the lintel reinforced concrete

beam must be put on top of the wall.

Figure 9. Production of Lightweight Styrofoam Concrete Block.

Figure 10. Production of Lightweight Styrofoam Concrete Panel.

Pouring the Styrofoam concrete Opening the moulding

Lifting the panel Constructing the panel

17 CMT - 130


CMT -

4. FULL SCALE WALL TESTS To see the wall behaviour of brick masonry and lightweight Styrofoam concrete under lateral load that simulates earthquake force, a series of full scale tests have been carried out in the Department of Civil Engineering Gadjah Mada university [Agustin (2005), Raharjo (2005), Satiawan (2005), Setyawati (2005)]. It is noted here that while the brick masonry wall was constructed using common layer by layer method, the lightweight Styrofoam concrete wall was constructed using direct pour method as shown in Figure 11. Using this direct method, the speed of construction is faster than that of layer by layer method. Moreover, the wall also becomes more monolithic. The setup of the specimen in general is shown in Figure 12.

Figure 11. Construction process of Direct Pour Lightweight Styrofoam Concrete [Satiawan (2005)].

Preparing the formwork. Pouring the Styrofoam concrete

Dismantling the formwork. Preparing the next formwork.

Figure 12. General Setup of Full Scale 3 m by 3m Wall Tests [Agustin (2005), Satiawan (2005)].

17 CMT - 131


CMT -

The descriptions of brick masonry wall and lightweight Styrofoam concrete specimens are given in Table 3, it is noted that brick masonry wall unit weight and compressive strength are much higher than that ones of lightweight Styrofoam concrete. The hysteretic behaviours of both walls under static and cyclic lateral load are shown in Figures 13 and 14. It can be seen that the lateral load strength of both walls are almost the same, but the lightweight Styrofoam concrete shows a little bit more ductile behaviour than that of brick masonry wall.

5. CONCLUSIONS From the above discussions, the following conclusions can be made.

1) Lightweight Styrofoam concrete can be used as an alternative material for wall to prevent the exploitation of soil in the rice field for brick production.

2) Lightweight Styrofoam concrete unit weight and compressive strength can be adjusted by controlling the Styrofoam content in the mix.

Table 3. Description of Brick Masonry and Lightweight Styrofoam Concrete WallsNo Description Brick Masonry Wall Lightweight Styrofoam

Concrete wall1 Material Bricks and mortar bed joint with Content per cubic meter:

the volumetric proportion - cement 350 kg1 cement : 6 sand - volumetric content of

Styrofoam 95% 2 Thickness 110 mm 100 mm3 Method of construction Layer by layer Direct pour4 Unit weight Brick = 1260 kg/m3 680 kg/m3

Mortar bed joint = 1950 kg/m3 No bed joint5 Compressive strength Mortar bed joint = 4.7 MPa 2.0 MPa

Brick = 2.6 MPa No bed joint

Figure 13. Hysteretic Loops of Unreinforced Brick Masonry Wall [Raharjo (2005), Satiawan (2005)]

-60

-45

-30

-15

0

15

30

45

60

-45 -30 -15 0 15 30 45Lateral Deformation (mm)

Late

ral L

oad

(kN

)

-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5Storey Drift (%)

Cyclic LoandingStatic Loading

17 CMT - 132


CMT -

3) Lightweight Styrofoam concrete wall can be constructed using direct pour method, which is faster than that of layer by layer method in the construction of brick masonry wall construction.

4) Full scale test results of brick masonry and lightweight Styrofoam concrete wall under static and cyclic load tests showed that lightweight Styrofoam concrete wall is superior than brick masonry in terms of unit weight and ductility.

6. RECOMMENDATIONS The researches of Styrofoam concrete that have been carried out in the Department of Civil Engineering Gadjah Mada University were limited to the study of physical and mechanical property only. The following further studies need to be carried out: 1) durability, 2) fire resitance, 3) impact resistace, 4) economic factor.

7. REFERENCES Agustin, R.S (2005). “Post Elastic Characteristic of Styrofoam Concrete Wall with Horizontal

Reinforcement under Static Horizontal Force.” (in Indonesian), Master Thesis, Department of Civil Engineering, Gadjah Mada University, Yogyakarta.

Crawford, R.J. (1998). Plastic Enggineering, Third Edition. Darmawan, F. (2004). “Lightweight Styrofoam Concrete for Precast Wall Panel.” (in Indonesian),

Master Thesis, Department of Civil Engineering, Gadjah Mada University, Yogyakarta. Musana (2006). “The Application of Styrofoam Waste for Lightweight Styrofoam Concrete with the

Cement Content of 250, 300, 350 kg/m3.” (in Indonesian), Master Thesis, Department of Civil Engineering, Gadjah Mada University, Yogyakarta.

Raharjo, E.P. (2005) “Post Elastic Characteristic of Brick Masonry Wall with Horizontal Reinforcement under Cyclic Horizontal Force.” (in Indonesian), Master Thesis, Department of Civil Engineering, Gadjah Mada University, Yogyakarta.

Satiawan, B. (2005). “Post Elastic Characteristic of Styrofoam Panel Wall with Horizontal Reinforcement under Cyclic Horizontal Force.” (in Indonesian), Master Thesis, Department of Civil Engineering, Gadjah Mada University, Yogyakarta.

Figure 14. Hysteretic loops of unreinforced Styrofoam concrete wall [Agustin (2005), Satyawan (2005)].

-60

-45

-30

-15

0

15

30

45

60

-45 -30 -15 0 15 30 45Lateral Deformation (mm)

Late

ral L

oad

(KN

)

-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5Storey Drift (%)

Cyclic LoadingStatic Loading

17 CMT - 133


CMT -

Satyarno, I (2004). “The Application of Cement Content for Lightweight Styrofoam Concrete.” (in Indonesian), National Seminar of Innovation in Building Material Technology, Joint Cooperation between Department of Civil Engineering Gadjah Mada University and PT Indocement Tunggal Prakarsa Tbk.

Satyarno, I (2005). “Lightweight Styrofoam Concrete Panel for Wall.” (in Indonesian), Proceeding National Seminar of Research Development in Material and Process, Center of Engineering Study, Gadjah Mada University, Yogyakarta.

Setyawati (2005). “Post Elastic Characteristic of Brick Masonry Wall with Horizontal Reinforcement under Static Horizontal Force.” (in Indonesian), Master Thesis, Department of Civil Engineering, Gadjah Mada University, Yogyakarta.

17 CMT - 134


CMT -

SYNTHETIC FIBER-MATRIX INTERFACE BOND PROPERTIES IN CEMENTITIOUS COMPOUND

Hexiang Dong1 and Tadashi Nishimura2

ABSTRACT : In this paper, the synthetic fiber pullout test by using the briquette molds according to JCI-SF 8 was conducted for evaluating the interface bond property of fiber and cementitious-base matrix. Five kinds of commercial available short length synthetic fibers, which have different shapes and were made of different raw materials such as polyolefin and polyvinyl alcohol, were used. The test results show that the fiber-matrix interface bond property is affected by geometry factor of fiber, such as the aspect ratio, the specific surface area and the property of surface layer of fiber. The fiber-matrix interface bond property of polyolefin fibers is based on the pullout of fiber and was greatly influenced by the fiber surface properties, I.e., roughness, embossment and fibril, so as to improve the frictional resistance of fiber pullout from matrix. Polyvinyl alcohol fibers are based on the fracture of fiber itself because it is stronger chemical bond between the matrixes than that of polyolefin finer. KEYWORDS: Polyolefin, polyvinyl, pullout test, short synthetic fiber, interface bond property, Chemical bond, frictional resistance

1. INTRODUCTION

Cementitious Matrix is a brittle material with the low ratio of tension and compression strength, it is common knowledge that Material is reinforced using the fiber with high-tension strength that is effective.

By the fiber reinforce the Cementitious Matrix (FRC), the energy absorptions of matrix, the extension modification capability, the load capability post cracking, and the resistance over shock/fatigue load can be raised. And toughness of matrix can be increased remarkably. Moreover, in the crack produced on the FRC structure object, the fiber reinforce can be distributed crack to smaller; the crack progress (development) can also be controlled. The fiber reinforcement Cementitious matrix is the composite material of the reinforce fiber and Matrix. It is thought that which is being united and functioning by the firm bond at the mutual interface, therefore, interfaces bond strength has decisive influence on the performance of FRC. The interface bond morphology of a fiber, especially the synthetic fiber with cementitious matrix is various, and the relation between the bond mechanism and reinforcement property has many still unknown portions. Moreover, many researchers have investigated about the characteristic of matrix and fiber, and there was little research on both interface bond morphologies. Thus, in this study, the interfacial bond properties (bond strength, pullout load – displacement curve, and so on) between synthetic fibers and cementitious matrices were investigated using Plurality fiber pullout test ． The synthetic fibers have different composition material properties, such Polyethylene/Polypropylene (PE&PP) fiber, Polyvinyl Alcohol (PVA) fiber and Polypropylene (PP) 1 Chief research engineer, Engineering Department of Grace Chemicals K.K 2 Technical Director, Grace Chemicals K K.

18 CMT - 135


CMT -

]n

-CH-CH2-CH3[ ]n

[-CH2-CH2-]n

PE PP PVA

Fig –1 The molecular formula of various fibers

OH

-CH2-CH- -[ ]n

-CH-CH2-CH3[ ]n

[-CH2-CH2-]n

PE PP PVA


OH

-CH2-CH- -[

-CH-CH2-CH3[ ]n

-CH-CH2-CH3[ ]n

[-CH2-CH2-]n

PE PP PVA


OH

-CH2-CH- -[ OH

-CH2-CH- -[

Picture-1 F-1～5 Synthetic FibersPicture-1 F-1～5 Synthetic Fibers

fiber; and different geometry size form, such circular section, rectangle section. The strength of the matrix, material properties, geometry size form, and the surface morphologies of reinforcement fibers were to affect the reinforcement effect greatly, which is well-known fact. On the other hand, the reinforcement effect of the fiber is in the bridging action which straddles the crack both ends of a matrix, fiber embedded length at a matrix and angle are various that is influence the reinforcement effect directly. Therefore, in this research, the embedding length of a fiber was set up with three levels, the embedding angle was set up with four levels, and the influence which they have on the bond characteristic was investigated.

2. EXPERIMENTAL METHODS

2.1 Synthetic Fibers Synthetic fibers used for this research being shown in Table –1,and, the molecule structure of them as shown in Fig. –1. 1) F-1 and F-3 fiber are the blend monofilament fiber of PE (Polyethylene) and PP (Polypropylene), the ratio of width and thickness is high as the shape of thin sheet, therefore, aspect ratio and Specific surface area of fibers are big value. One side, the width of them is the same; F-3 fiber is thicker and longer than F-1 fiber only. In addition, the surface of this kind of fiber with changes a lot through mixing process in cementations matrix (I.e., concrete). 2) F-2 is the monofilament fiber of PVA (Polyvinyl Alcohol), there is a hydrophilic group called hydroxyl group in a molecular formula as shown in Fig.-1, and with strong chemical bond to surrounding matrix. 3). F-4 and F-5 fibers are the monofilament fibers of PP (Polypropylene), and has given macroscopic emboss processing to the surface in the molding process of the fiber. In addition, F-5 is wave shape from by bends linear shape fiber of F-4 to make. These are molecular formula of PE, PP, and a PVA fiber is shown in Fig.-1, and appearance being shown in a picture -1.

2.2 The-geometry form element of a synthetic fibers The items showing the geometry-element of the fiber was aspect ratio and specific surface area. The aspect ratio (it expresses with As below) is defined by the following formula-1

dLAs = （1）

Here，L is the length (mm) of fiber and d is the cross-sectional diameter (mm) of fiber. However, when the section of a fiber is not circular, the conversion diameter when converting circularly equivalent area is used.Moreover, the specific surface ratio of fiber is the

Figure 1. The molecular formula of various fibers

Photo 1. F-1~5 Synthetic fibers

18 CMT - 136


CMT -

0

2

4

6

8

10

F-1 F-2 F-3 F-4 F-5fiber types

Asp

ect R

atio

0

5

10

15

20

25

Spec

ific

Surf

ace

area

(mm2 /m

m3 )

Aspect ratio Specific surface area

Fig-2 Fibers Geometry properties

0

2

4

6

8

10

F-1 F-2 F-3 F-4 F-5fiber types

Asp

ect R

atio

0

5

10

15

20

25

Spec

ific

Surf

ace

area

(mm2 /m

m3 )

Aspect ratio Specific surface area

Fig-2 Fibers Geometry properties

Table-2 Mortar mixture of Bond test

* The standard sand and cement use which standard sand portolandcement applies to GB(China Standard)178-77.

CompressiveStrength(Mpa)

30 5040 40

SandType*

FiberType

Standard Standard F-1～5

W/C（％）

C：S

1：1.7

CementType*

Table-2 Mortar mixture of Bond test

* The standard sand and cement use which standard sand portolandcement applies to GB(China Standard)178-77.

CompressiveStrength(Mpa)

30 5040 40

SandType*

FiberType

Standard Standard F-1～5

W/C（％）

C：S

1：1.7

CementType*

Picture-2 Test Piece and direct tension testPicture-2 Test Piece and direct tension test

Table 2. Mortar mixture of bond test

Photo 2. Test piece and direct tension test

ratio of the surface area and volume of fiber, and it is expressed with SV is defined by the following formula-2

V

ΣSSv =

（2）

Here，∑S is the surface area (mm2) ，V is the volume of fiber(mm3)．The aspect ratio and specific surface area of the various synthetic fibers used for this research, is as being shown in Table -1 and Fig.-2. 2.3 The-bond test of fibers The bond test of Synthetic fibers is preformed accordingto JCI-SF8「Method of test for bond of fibers」1）

1） about casting test piece The mixture design of mortar being shown in

Table –2, the change in the amount of Admixture addition adjustedthe flow of mortar to 240-250mm. in addition，Because F-1 and F-3 fibers，F-1 and F-3 fiber’s surface property (It becomes coarse) was changed in after mixing in cementitious matrix, After mixing in concrete beforehand and performing Mixing, the fiber probed and dried was used for this test. 2） Pullout test method

Test piece form of bond test as shown in a picture -2.Thelength of test piece is 240mm and the rectangle minimumsection size of the central part is 24×26mm. The central partof the length direction was installed the partition board made from plastic of 1mm thickness (there are some small holes for

Tensile Strngth

Young Modulus Density Length Specific

Surface area Circumferen

tial length Cross-sect

ion area Items

Fiber Material (MPa) (GPa) (g/cm3) (mm)

Aspect Ratio

(mm2/mm3) (mm) (mm2)

F‐1 PE＆PP 540 9.5 0.92 40 90 19.7 3.02 0.154

F‐2 PVA 880 29.4 1.30 40 60 6.93 2.36 0.342

F‐3 PE＆PP 540 9.5 0.92 48 82 12.0 3.18 0.266

F‐4 PP 440 9.8 0.91 48 60 6.04 3.00 0.500

F-5 PP 440 9.8 0.91 48 60 6.04 3.00 0.500

Table-1 Properties of Synthetic Fibers

Type

Tensile Strngth

Young Modulus Density Length Specific

Surface area Circumferen

tial length Cross-sect

ion area Items

Fiber Material (MPa) (GPa) (g/cm3) (mm)

Aspect Ratio

(mm2/mm3) (mm) (mm2)

F‐1 PE＆PP 540 9.5 0.92 40 90 19.7 3.02 0.154

F‐2 PVA 880 29.4 1.30 40 60 6.93 2.36 0.342

F‐3 PE＆PP 540 9.5 0.92 48 82 12.0 3.18 0.266

F‐4 PP 440 9.8 0.91 48 60 6.04 3.00 0.500

F-5 PP 440 9.8 0.91 48 60 6.04 3.00 0.500

Table-1 Properties of Synthetic Fibers

Type

Figure 2. Fibers geometry properties

Table 1. Properties of synthetic fibers

18 CMT - 137


CMT -

Picture-3 Test for bond of fibers SetupPicture-3 Test for bond of fibers SetupPhoto 3. Test for bond of fibers setup

carrying out fiber installation)beforehand, and 5 fibers were divided near the rectangle peak (But,6 fibers are used in the test of the embedding angle of fibers, two fibers are made into 1 set, the two fibers was bent to so that it may become the angle of a plan,), remaining one into the rectangular center, and they are install through a board. After carrying out installation of the fiber beforehand (At embedding length and angle as planned ), mortarwas divided into two layers, and was placed. We tested by having pullout directly and decided to bond properties between fibers and cementitious matrix. Usingthe electronic omnipotent test machine CMT5105（Picture-3）, the amount of pullout load and displacement of fiber(displacement of a crosshead quantity) was simultaneouslyreading by eight frequency/second, and the loading speed is controlled so that the increase speed of the displacement of fiber becomes to 0.4-0.5mm/second. The test piece of bond test and synthetic fibers situation of after the completion of direct pullout test was shown in picture -2. 3） the evaluation method of the fibers bond characteristics As the bond characteristic, the test results that are the pullout load-displacement curve until the amount of displacement reaches predetermined value,( 2.5mm for JCI-SF8) and the bond strength of fibers, and so on are obtained. The pullout stress （σf） and bond stress（τ） of fibers are calculated by the following formulas -3-4 respectively2）.

df Sn

Pσ×= （3）

SSnP×

=τ （4）

Here，P is pullout load（N），n is the numbers of a fiber，Sd is the cross-section area（mm2

），Ss is the bond area （Embedding length ×circumferential length, mm2

）. Like the flexural toughness of fiber reinforcement concrete bend test, the area under the pullout – displacement curve the fiber became the absorption energy capacity in fiber pullout process, from cementitious matrix, and we defined this area as "bond toughness" like "flexural toughness”, and decided to express with Te (N×mm). Therefore, bond toughness relation with the pullout load directly; and, since as the fibers is thick (cross-section area is high), the pullout load become larger, and the toughness also becomes higher. That is, the bond toughness is directly dependent on the volume factor of the fibers.

18 CMT - 138


CMT -

0

5

10

15

20

25

30

35

40

0 5 10 15Displacement(mm)

Pullo

ut

Load

(N)

Debondingwhole Fiber Slippage

Fig-3 Embedding Length – max Pullout Load

bonding regime

0

5

10

15

20

25

30

35

40


Pullo

ut

Load

(N)

Debondingwhole Fiber Slippage


bonding regime

Fig-4 Displacement - Pullout Load Curve(F-1,F-2 Fibers)

01020304050607080


Pullo

ut L

oad(

N)

F-1F-2

^

30 MPa,30℃

Fig-4 Displacement - Pullout Load Curve(F-1,F-2 Fibers)

01020304050607080


Pullo

ut L

oad(

N)

F-1F-2

^

30 MPa,30℃

Figure 3. Embedding length – max. pullout load

Figure 4. Displacement-pullout load curve (F1, F2 fibers)

One side, if the dimension of the fiber differs, he specific surface area is also differing. The specific surface area is larger, bond area between fibers and the matrix becomes large, and the total energy absorption capacity is also considered to become higher. Therefore, Here, when expressing the bond characteristic of fibers better, we newly introduced the indices of relative insensitive to fibers dimension factor, suppose that index called the bond toughness coefficient Tb (MPa). Suppose Tb is the product of bond toughness (Te（N×mm）／Ss) in the unit bond area, and specific surface area (Sv（mm2

／mm3）), it calculates by the following formulas-5.

vb ST e ×=sS

Τ （5）

3．RESULTS AND DISCUSSION

3.1 Fiber pullout Process from Cementitious matrix An example of the curve of pullout load – displacement (L-D Curve) of polypropylene type synthetic fibers were shown in Fig. -3. In general, the pullout process of this kind fiber can be classified in the following three stages.3）

1 ） The bonding stage until bond loses （ bonding regime）：in this stage, with the increase in pullout load,

the elastic extension of free end of the fiber s equal to displacement, and the fibers bond is held. 2）The bond was broken（debonding）：bond is lost and the fibers begins to be slippery in matrix． 3）Sliding starts（fiber slippage）stage：the fiber continues undergo pullout load through interface friction, sliding in

matrix.. The pullout load – displacement Curves of various Synthetic fibers obtained by the direct pullout test used mortar test piece was shown in Fig. 4-7. According to this results, 1) In bonding regime, since the cross-section area of F-4 and F-5PP fiber was large (it is shown in table-1), the maximum pullout load was the largest, and the maximum pullout load of F-2 PVA Fiber, which has strong chemical bond with matrix, was also bigger than which of F-1 and F-3 PE & PP fibers.

18 CMT - 139


CMT -

Fig-5 Displacement – Pullout Load Curve(F3～5　Fibers）

0

20

40

60

80

100

120


Pullo

ut L

oad(

N)

F-3F-4F-5

30 MPa,30℃

Fig-5 Displacement – Pullout Load Curve(F3～5　Fibers）

0

20

40

60

80

100

120


Pullo

ut L

oad(

N)

F-3F-4F-5

30 MPa,30℃

Fig-6 Displacement – Pullout Curves(F-4 Fiber,different Embedding Length)

0

2040

60

80

100120

140

160

0 5 10 15 20Displacement(mm)

Pullo

ut L

oad(

N)

10mm15mm20mm

40 MPa,0℃

Fig-6 Displacement – Pullout Curves(F-4 Fiber,different Embedding Length)

0

2040

60

80

100120

140

160

0 5 10 15 20Displacement(mm)

Pullo

ut L

oad(

N)

10mm15mm20mm

40 MPa,0℃

0

50

100

150


Pullo

ut L

oad(

N)

0℃30℃45℃60℃

Fig-7 Displacement – Pullout Curves(F-4 Fiber,different Embedding Angle)

30 MPa

0

50

100

150


Pullo

ut L

oad(

N)

0℃30℃45℃60℃

Fig-7 Displacement – Pullout Curves(F-4 Fiber,different Embedding Angle)

30 MPa

Figure 5. Displacement-pullout load curve (F3 ~ F5 fibers)

Figure 6. Displacement-pullout load curve (F4 fiber, different embedding length)

Figure 7. Displacement-pullout load curve (F4 fiber, different embedding angle)

2) Debonding, the pullout load falls of F-2 PVA fiber was greatly, it’s occurs was very suddenly and vertical. This is considered to be based on the loss of chemical bond between fiber and matrix4）, and the fracture of fibers. Moreover, for PF-4 and F-5 PP fibers, the load down after reaching the maximum is large, because F-4 and F-5 PP fibers havemacroscopic surface emboss, it’s to conquer the frictional resistance of surrounding matrix,

and to sliding that need bigger pullout load .

Conversely, once fibers sliding, the down of load is also was large. Others, F-1 and F-3 PE&PP fibers have small cross-section area, microscopic surface emboss(after mixing).the maximum pullout loads was not high, so the fall of pullout load was also small. 3) Fiber slippage, Since F-1 and F-3 PE＆PP Fiber, which has the microscopic surface emboss undergo pullout load, pullout load to conquer frictional resistance was necessity not much high, and the down after sliding are small also, the L-D curve was smooth. For F-4 and F-5 PP fibers have macroscopic surface emboss, in order to conquer frictional resistance, it is required large pullout load, and the load down after sliding also becomes large,. the pullout load change is greatly, so that the L-D curve is gone up and down wavelike. One side, Since F-2 PVA fiber was fractured one after another, the Pullout load was downed rapidly gradually. Therefore, the actions of which various synthetic fibers in the pullout process from cementitious matrix was very different 3.2 The bond strength of difference fibers The test result of the bond strength of various synthetic fibers in different embedding length, and different embedding angle was shown in Fig -8-9. According to this results,

18 CMT - 140


CMT -

Fig-8 Bond Strength in different Embedding Length

0

1

2

3

4

5

F-1 F-2 F-3 F-4 F-5Fiber types

Bond

Stre

ngth

（ Mpa

）

10mm15mm20mm

40 MPa

Fig-8 Bond Strength in different Embedding Length

0

1

2

3

4

5


Bond

Stre

ngth

（ Mpa

）

10mm15mm20mm

40 MPa

Fig-9 Bond Strength in different fiber type and embedding angle

0

1

2

3

4


Bond

Stre

ngth

（ MPa

）

0℃30℃45℃60℃

30 MPa,

Fig-9 Bond Strength in different fiber type and embedding angle

0

1

2

3

4


Bond

Stre

ngth

（ MPa

）

0℃30℃45℃60℃

30 MPa,


0

50

100

150

200

F-1 F-2 F-3 F-4 F-5Fiber type

Max

Pul

lout

Loa

d （N

）

10mm15mm20mm

40 MPa


0

50

100

150

200


Max

Pul

lout

Loa

d （N

）

10mm15mm20mm

40 MPa

Figure 8. Bond strength in different embedding length

Figure 9. Bond strength in different fiber type and embedding angle

Figure 10. Embedding length – max. pullout load

1) It is tendency was shown that the embedding length of a fiber waslarger, the bond strength becomes lower. This is as having shown formula-4, by the direct pullout test of fibers, it is surmised that he minus influence of the increase in the bond area given to the bond strength is larger than the plus influence of increase of the pullout load accompanying increase of embedding length.

2) The difference of F-1 and F-3 PE&PP fibers in the bond strength is also simply depended on the size of the cross-section area of fibers, and since F-5 PP fiber had bent like a wave shape in Length direction, the bond strength showed the tendency which becomes some high than F-4 straight PP fiber, which is same kind fiber.

3) Various synthetic fibers showed same tendency, that while the fiberembedding angle, was 0 degree C - 45 degrees C, the bond strength continued increasing, and which becomes low bordering on 45 degrees C. It has suggested that the plus influence on the frictional resistance byincrease of the embedding angle of a fiber changed this to reduction from the increase bordering on 45 degrees C.

4) The results of the bond strength of F-2 PVA fiber, F-4 and F-5 PP fibers was higher than F-1 and F-3 PE&PP fibers was obtained. This is considered for F-4 and F-5 PP fibers to be thick, and the cross-section area to be larger than F-1and F-3 PE&PP fibers, and for F-2 PVA fiber to be for Strong chemical bond to act with matrices.

3.3 Fibers bond properties in differ embedding length It's was shown in Fig 10-12 about the test results of the maximum pullout load, bond toughness, and bond toughness coefficient of various synthetic fibers in different the embedding length. According to this results, 1) It is tendency was shown. That the more the embedding length is large of fibers; the more high of maximum pullout load (pullout load in the case of bond with the matrix is lost) of the fibers. This is considered that because the fibers bond area with the matrix also to become large, and the total frictional resistance power over pullout of a fiber also to becomes large, if the embedding length of the fiber is large.

18 CMT - 141


CMT -

Fig-11 Embedding Length – Bond Toughness

0

50

100

150200

250

300

350


Bon

d To

ughn

ess （

N×

mm

）

10mm15mm20mm

40 MPa

Fig-11 Embedding Length – Bond Toughness

0

50

100

150200

250

300

350


Bon

d To

ughn

ess （

N×

mm

）

10mm15mm20mm

40 MPa

Fig-12 Embedding Length – Bond Toughness Coefficient

0

10

20

30

40

50

60


Bon

d To

ughn

ess

Coe

ffic

ient

（

MPa

） 10mm15mm20mm

40 MPa

Fig-12 Embedding Length – Bond Toughness Coefficient

0

10

20

30

40

50

60


Bon

d To

ughn

ess

Coe

ffic

ient

（

MPa

） 10mm15mm20mm

40 MPa

Fig-13 Embedding Angle – max Pullout Load

020406080

100120140160


Max

Pul

lout

Loa

d （N

）

0℃30℃45℃60℃

30 MPa

Fig-13 Embedding Angle – max Pullout Load

020406080

100120140160


Max

Pul

lout

Loa

d （N

）

0℃30℃45℃60℃

30 MPa

Figure 11. Embedding length – bond teoughness

Figure 12. Embedding length – bond teoughness coefficient

Figure 13. Embedding angle – max. pullout load

Moreover, it is though it was natural the results that when the maximum pullout load became large, the area under the L-D curve, i.e., bond toughness, becomes large, and the energy absorption capacity becomes high was obtained. 2) On the other hand, the bond toughness coefficient of various synthetic fibers when embedding length sets to 20mm low than15mm a little. It is thought that it is bond toughness coefficient was the bond toughness in unit bond area(shown in formula -5),although both bond toughness and bond area become larger as embedding length is large, the ratio is becomes small a little. 3) The result of high bond toughness of F-4 and F-5 PP fiber have large cross-section area, and F-2 PVA fiber have high pullout load was obtained. But, for the bond toughness coefficient of various synthetic fibers, in consideration of the influence (in order to lose influence) of the volume element of fibers like the cross-section area , the bond toughness coefficient of F-1 and F-3 PE&PP fibers have high aspect ratio and specific surface ratio (more thin fiber, in other words), was higher value obtained. Namely, the various synthetic fibers, which do not change the difference in physical properties (tensile strength and young modulus) so much, by raising aspect ratio and specific surface area of the fibers, the energy absorption capacity in unit volume of fibers can be heightened. 4) The difference of F-1 and F-3 PE&PP fibers in the maximum pullout load, bond toughness, and bond toughness coefficient there are depended on the difference of aspect ratio and specific surface area of them. In addition, the difference in the maximum pullout load, bond toughness, and bond toughness coefficient of straight F-4 and waveform-like F-5PP fiber was not seen so much. 3.4 Fibers bond properties in differ embedding Angle The maximum pullout load, bond toughness, and bond toughness coefficient of the various synthetic fibers in the case of different embedding angle was shown in Fig13-15.

18 CMT - 142


CMT -

Fig-14 Embedding Angle – Bond Toughness

0

50

100

150

200

250


Bon

d To

ughn

ess （

N×

mm

）

0℃30℃45℃60℃

30 MPa

Fig-14 Embedding Angle – Bond Toughness

0

50

100

150

200

250


Bon

d To

ughn

ess （

N×

mm

）

0℃30℃45℃60℃

30 MPa

Fig-15 Embedding Angle – Bond Toughness Cioefficient

0

10

20

30

40

50

60


bong

Tou

ghne

ss C

oeff

icie

n（

Mpa

）

0℃30℃45℃60℃

30 MPa


0

10

20

30

40

50

60


bong

Tou

ghne

ss C

oeff

icie

n（

Mpa

）

0℃30℃45℃60℃


0

10

20

30

40

50

60


bong

Tou

ghne

ss C

oeff

icie

n（

Mpa

）

0℃30℃45℃60℃

30 MPa

Figure 14. Embedding angle – bond toughness

Figure 15. Embedding angle – bond toughness coefficient

According to this results, 1) This tendency was shown that the results of the maximum pullout load, bond toughness, and bond toughness coefficient of various synthetic fibers, were become higher accompanying increase of embedding angle becomes large ( from 0 degree C, to 30 degrees C, and 45 degrees C). But, when the embedding angle became 60 degrees C, the above factor showed becomesconversely lower than the case of 45 degrees C. The influence of the embedding angle on the bond properties between the fibers and matrix was shows the different tendencybordering on 45 degrees C. 2) On the other hand, in change of embedding angle of fibers, the results of the maximum pullout load and bond toughness of various synthetic fibers become large in order of F-1, F-3 PE&PP fibers, F-2 PVA fiber, and F-4,and F-5 PP fibers (F-4 and F-5 fibers do not have differ so much) was obtained. This is considered that because F-4 and F-5 PP fiber have greatly the cross-section area, and F-2 PVA fiber have the strong chemical bond between fiber and matrix, the maximum pullout load and bond toughness was highly little than F-1 and F-3 PE&PP fibers. 3) For the bond toughness coefficient results of F-1 and F-3 PE &PP fibers, which has high aspect ratio and specific surface area high (it described to 3.3.3) like), and the next is F-2 PVA fibers comparatively high, This value of F-4, and F-5 PP fiber is lower than F-1 and F-3 PE &PP fibers, and F-2 PVA fibers. Moreover, the value of bond toughness coefficient of F-4 and F-5 PP fiber did not much have differed.

4. CONCLUSION

The results obtained from this study are helpful to for better understanding of the role of synthetic fibers in improving the properties of brittle cementitious composites. Plurality fibers pullout tests were preformed to investigate the interfacial bond properties (bond strength, pullout load - displacement curve, and so on), and therefore to obtain a preliminary knowledge about the different interfacial bond properties As opposed to different synthetic fiber types (for example: different material), and it’s different properties (for example: geometry factor, and surface properties). 1） The L-D curves which show the pullout process in the cementitious matrix of various synthetic fibers was differs, the load change of PVA fiber which have chemical bond with matrix was distinct

18 CMT - 143


CMT -

and suddenly; PP fiber which has macro surface emboss showed sharp load change, the load change of PE &PP fiber which has micro surface emboss, was quiet. 2） The bond strength of PP and PVA fiber was higher than PE&PP fiber, the macro surface emboss of PP fiber and chemical bond of PVA fiber was strong to the resistance to pullout load. 3） Although max pullout load and bond toughness became large with embedding length from 10mm to 20mm increase; the bond toughness coefficient was the maximum in embedding length 15mm. One side, although max pullout load, bond toughness and bond toughness coefficient became large with embedding angle from 0� to 45� increase; when embedding angle from 45� to 60� increase, then, they are changed to decreasing. 4） however, the wave shape fiber and the straightly shape fiber was almost no interface bond performance difference shown in this study.

ACKNOWLEDGEMENT

I received counsel precious to Professor Hiroshi Seki,waseda university(Japan), and, Professor Wei Sun southeast university(China) in carrying out this research. Moreover, in this test of this research, I received help of the doctor Jianzhong Lai, and,Jinyang Jiang of Wei Sun Lab, and Grace Chemicals Engineering Department members. Gratitude here is expressed.

5. REFERENCES

Japan Concrete Institute (1984). “JCI Standard for Test Methods of Fiber Reinforced Concrete.” JCI-SF

Kobayashi K. (1981). “The characteristic and application of fiber reinforcement concrete.” Ohm-Ltd,1981.

Youjiang，W.，Victor，C.L.，and Stanley，B. (1988). “Modeling of fiber pull-out from a cement matrix.” The International Journal of Cement Composites and Lightweight Concrete.，Volume.10，

Number.3，August. Carl，R.，Victor，C.L.，Cynthia，W.，Hideki，H.，Tadashi，S.，and Atsuhisa，O．(2001). “Measuring

and Modifying Interface Properties of PVA Fibers in ECC Matrix.” Journal of materials in civil engineer，November/December , pp.399�406.

Sehaj，S.，Arun，S.，and Richard，B． (2004). “Pullout behavior of polypropylene fibers from cementitious matrix.”Cement and Concrete Research，No.34，pp1919-1925.

18 CMT - 144


CMT -

FRESH PROPERTIES OF HIGHLY-FLOWABLE STEEL FIBER-REINFORCED CONCRETE

Morio Kurita1 and Hirokazu Tanaka2

ABSTRACT : With the aim of gaining knowledge about the fresh properties of highly-flowable steel fiber-reinforced concrete, the effects of three mix design factors, namely, the absolute unit volume of coarse aggregate, steel fiber content and the dosage of the viscosity agent, on flowability, deformability and segregation resistance were investigated experimentally. The slump flow of highly-flowable steel fiber-reinforced concrete is influenced by the absolute unit volume of coarse aggregate, steel fiber content and the dosage of viscosity agent. The segregation properties of highly-flowable steel fiber-reinforced concrete should be evaluated in terms of the segregation index of steel fibers. Under constant slump flow conditions, there are combinations of the absolute unit volume of coarse aggregate and the steel fiber content that provide excellent deformability.

KEYWORDS: steel fiber, highly-flowable concrete, flowability, viscosity, deformability, segregation resistance, mortar

1. INTRODUCTION

Steel fiber reinforced concrete (SFRC) is a composite material developed to reduce the brittleness of concrete and dramatically increase its ductility. SFRC is used extensively to line tunnels and other underground structures, increase the thickness of pavements, and repair and strengthen various structures. Steel fibers used in SFRC are 20 to 60 mm long and 0.4 to 0.9 mm in diameter and have aspect ratios ranging from about 35 to 80. The most common steel fiber content is about 0.5 to 1.0 percent (by volume). Attempts are being made at using SFRC for the Extruded Concrete Lining (ECL) Method of tunnel construction, in which concrete is cast in place without using lining segments in order to shorten the period of shield tunneling, improve the economy of tunnel construction and prevent land subsidence caused by tail void formation. In these attempts, concrete with relatively high flowability is being used with the aim of improving construction efficiency. Research is also underway on the use of SFRC columns as structural members. The Japan Society of Civil Engineers has published a design guideline indicating the basics in the design of steel fiber reinforced concrete columns. Since, however, SFRC is not as easy to work with as conventional concrete, SFRC's excellent hardened properties are not currently being used effectively. To address this problem, it is essential to make SFRC easier to work with, and it is therefore necessary to develop SFRC with fresh properties that make SFRC easier to work with. Although one way to do it is to pursue the possibilities of highly flowable concrete, little research is being done on highly flowable concrete containing steel fibers. For the purpose of obtaining basic data for the mix design of highly-flowable steel fiber-reinforced concrete for primary linings built by the ECL Method, an experimental study was conducted on flowability, deformability and segregation resistance, parameterizing the absolute unit volume of coarse aggregate, steel fiber content and viscosity. In this study, the influence of various factors on the fresh properties of highly-flowable steel fiber-reinforced concrete was investigated.

1 Chief Research engineer, D. Eng., Shimizu Corporation, Japan 2 Research engineer, M. Eng., Shimizu Corporation, Japan

19 CMT - 145


CMT -

2. EXPERIMENTAL

2.1 Materials used Table 1 shows the materials used. In the tests conducted for the purposes of this study, the viscosity of concrete was used as a parameter, and viscosity was varied by changing the amount of the viscosity agent used. The viscosity agent was composed mainly of cellulose, and a polycarboxylic acid type air-entraining high-range water reducer was used. Steel fibers with hooked ends, which are 0.6 mm in diameter and 30 mm long, were used.

Table 1. Materials used 2.2 Test method Table 2 shows the test method used. Because there is currently no standard method for testing the deformability of highly-flowable steel fiber-reinforced concrete, the flowability properties of the concrete was evaluated through V75-funnel testing. Mini slump flow tests and K-funnel tests (Kawai et al. 1993), in which highly-flowable steel fiber-reinforced concrete was wet-screened with a 5 mm sieve, were also conducted to test the properties of mortar excluding coarse aggregate and steel fibers. It has been shown that there are strong correlations between mini slump flow and the yield value of mortar and between K-funnel flow time and the plastic viscosity of mortar. The slump cone used in the mini slump flow tests was of the type specified in JIS A 1173 Method of Test for Polymer-Modified Mortar, and a cone that is the half size of the slump cone specified in JIS A 1101 (referred to as a "mini slump cone") was used. The "K-funnel" consists of a section of the J14-funnel described in JSCE-F531-1994, a straight pipe 30 mm in inside diameter and 60 mm long attached to the 30-mm-or-less-in-inside-diameter end of the funnel, and a straight pipe 70 mm in inside diameter and 160 mm long attached to the upper end of the funnel. Segregation properties, which indicate the quality of concrete, can be conservatively evaluated by conducting a test under more demanding conditions than the actual concreting conditions. It was decided, therefore, to use vibration for the testing. In the segregation tests, a cylindrical steel form 15

Materials Characteristics Cement, C Ordinary Portland cement, Density 3.16g/cm3

Sand Aggregate, S Crashed sand, Density 2.74g/cm3, Water absorption 0.86％, Fineness modulus 2.80

Coarse aggregate, G Crashed stone Density 2.72g/cm3, Water absorption 0.68％, Fineness modulus 6.43

Air-entraining high-range water reducing agent, SP Polycardoxylic acid type

Viscosity agent, VA Cellulose type Steel fiber, SF Hooked end type, Φ0.6mm×L30mm

Test items Test methodSlump flow JSCE-F503Flow-out time of V75-funnel JSCE-F512Air content JSCE-F513Temperatuer Bar thermometerMini slump flow reference to the textFlow-out time of K-funnel reference to the textSegregation index of Coarse aggregate reference to the textSegregation index of Steel fiber reference to the text

Table 2. Test methods

19 CMT - 146


CMT -

cm in diameter and 30 cm high was filled with freshly mixed concrete, and the formed concrete was vibrated on a shaking table with a frequency of 50 Hz and amplitude of 1.1 mm for 30 seconds. After the vibrating, 2 liters of specimen were taken out from the upper and lower parts of the steel form. Then, the segregation indexes of the coarse aggregate and steel fibers, SI(G) and SI(SF), were calculated from the formulas shown below to compare the segregation properties of the two types of materials(Kawai et al. 1993).

SI(G) ={(G2−G1)/(G1+G2)}×100 (%) SI(SF) ={(SF2−SF1)/(SF1+SF2)}×100 (%)

where SI(G): segregation index of coarse aggregate SI(SF): segregation index of steel fibers G1: mass of coarse aggregate in the 2 liters of upper specimen G2: mass of coarse aggregate in the 2 liters of lower specimen SF1: mass of steel fibers in the 2 liters of upper specimen SF2: mass of steel fibers in the 2 liters of lower specimen

2.3 Types of test Within the ordinary slump range, the slump of concrete tends to decrease as the steel fiber content increases. It is generally said that this is because of mechanical engagement between steel fibers and coarse aggregate. In the mix design of steel fiber-reinforced concrete, therefore, the fine aggregate content was increased (i.e., the absolute unit volume of coarse aggregate was decreased) as the steel fiber content increased and the water content was increased accordingly. Properties such as flowability, deformability, self-compacting property and segregation resistance that indicate the fresh properties of highly flowable concrete are greatly influenced by the absolute unit volume of coarse aggregate and viscosity. When investigating the fresh properties of highly-flowable steel fiber-reinforced concrete, therefore, it is necessary to investigate the factors affecting the fresh properties of both steel fiber-reinforced concrete and highly flowable concrete. Since the density of steel fibers (7.85 g/cm3) is by far higher than those of other materials, it is thought likely that the density of steel fibers greatly influences segregation resistance, which is one of the fresh properties. In the tests, attention was paid to the steel fiber content, the absolute unit volume of coarse aggregate and the viscosity of concrete. It was decided to use a viscosity agent, which makes it easy to vary the viscosity of concrete with relative ease, and the viscosity was varied by adjusting the amount of the viscosity agent used. In order to investigate the influence of various factors on the fresh properties of highly-flowable steel fiber-reinforced concrete, the following tests were conducted: a) Tests to evaluate the influence of the viscosity agent and the steel fiber content (Series 1) The effects of the steel fiber content and the dosage of the viscosity agent on the fresh properties while the absolute unit volume of coarse aggregate and the dosage of the air-entraining high-range water reducer are kept constant were evaluated. The steel fiber contents of 0.5, 0.75 and 1.0 vol.% were used, and the dosage of the viscosity agent was varied from 600 to 1,800 g/m3 at 300 g/m3 intervals. The absolute unit volume of coarse aggregate and the dosage of the water-entraining high-range water reducer were kept at 200 L/m3 and C×2.0%, respectively (Kurita & Tanaka 1998). b) Tests to evaluate the influence of the absolute unit volume of coarse aggregate and the steel fiber content (Series 2) The absolute unit volume of coarse aggregate and the steel fiber content were varied while the dosages of the air-entraining high-range water reducer and the viscosity agent were kept constant. The steel

19 CMT - 147


CMT -

fiber content was set at two levels (0.75%, 1.0%), and the absolute unit volume of coarse aggregate was varied from 150 to 250 L/m3 at 25 L/m3 intervals. The dosages of the air-entraining high-range water reducer and the viscosity agent were kept at C×2.0% and 1,200 g/m3, respectively (Kurita & Tanaka 1999). c) Tests to evaluate the influence of various factors under constant slump flow conditions (Series 3) The fresh properties under the conditions that make slump flow more or less constant in cases where the steel fiber content and the absolute unit volume of coarse aggregate are varied were investigated. Under different combinations of various factors, the dosages of the air-entraining high-range water reducer and the viscosity agent were varied from C×1.4 to 2.0% and 900 to 2,000 g/m3, and the mix proportions from which the target slump flow level of about 60 cm can be achieve were chosen.Table 3 shows the combinations of Series 1 to Series 3 tests and the mix proportions used. The target air content level was set at 2% because underground structures were under consideration. In the Series 1 to Series 3 tests, the air content and concrete temperature were 1.1 to 2.5% and 21 to 22°C, respectively.

Table 3. Test combinations and mix proportions used


3.1 The influence of the dosage of the viscosity agent and the steel fiber content on fresh properties (Series 1) a) Fresh concrete Figure 1 to Figure 4 show the test results for fresh concrete. In the tests, segregation was observed visually, regardless of the steel fiber content, in the cases where the dosage of the viscosity agent was 600 and 900 g/m3. The visible segregation occurred after the concrete mix was dumped into the mixing pan, the mix was turned over and the mix was left to stand for a while. This indicates that steel fibers and coarse aggregate settled so that segregation resulted. As the dosage of the viscosity agent increased, slump flow decreased linearly. It is thought that as the dosage of the viscosity agent increased, the viscosity and yield value of the concrete increased. As the steel fiber content increased, slump flow generally decreased. The likely reason is that as the steel fiber content increased, mechanical engagement with the coarse aggregate increased so as to cause flowability to decrease. This is similar

Unit weight（kg/m3） Chemical admixture Types of tests

W/C (%)

Vg (L/m3)

Steel fiber content (vol%) W C SP (Cx%） VA (g/m3)

0.5 0.75 Series 1 200 1.0

600-1800

175-250 0.75 Series 2

150-225 1.0

2.0

1200

150-300 0.5 150-275 0.75 Series 3

40

150-250 1.0

200 500

- -

50

55

60

65

70

75

80

300 600 900 1200 1500 1800 2100

Vf:1.0 vol%Vf:0.75 vol%Vf:0.5 vol%

Slum

p flo

w(c

m)

Viscosity agent(g/m3)

Vg:200L/m3

Figure 1 Relation between viscosity agent and slump flow

19 CMT - 148


CMT -

Figure 5. Relation between mini slump flow and slump flow

Figure 2. Relation between steel fiber content and slump flow

to the tendency shown by conventional steel fiber-reinforced concrete. Slump flow changed by about 7 cm as the dosage of the viscosity agent increased by 500 g/m3, and a similar tendency was observed regardless of the different steel fiber content. At the same dosage of the viscosity agent, slump flow decreased linearly as the steel fiber content increased. Slump flow changed by about 5 cm as the steel fiber content changed by 0.5%. The tendency of this relationship remained more or less the same regardless of the dosage of the viscosity agent. Figure 3 shows the relationship between the dosage of the viscosity agent and V-funnel flow time. At the steel fiber content of 1 vol.%, when the dosage of the viscosity agent was 1,200 g/m3 or more, as the dosage of the viscosity agent increased, the viscosity of the concrete increased so that V-funnel flow time increased. When the dosage of the viscosity agent was 900 g/m3 or less at which segregation was observed, deformability decreased and V-funnel flow time became longer because of the settlement of the coarse aggregate and steel fibers and mechanical engagement between them. When the dosage of the viscosity agent was 600 g/m3, the tendency of segregation became pronounced and the funnel was blocked. In the cases where the steel fiber content was 0.75 vol.% and 0.5 vol.%, V-funnel flow time increased linearly as the dosage of the viscosity agent increased. In the case where the dosage of the viscosity agent was 900 g/m3 or less at which segregation was observed, the funnel was not blocked unlike in the case where the steel fiber content was 1 vol.%. From these results, it can be inferred that within the range in which segregation does not occur, the highest level of deformability can be obtained by the mix proportions that minimize V-funnel flow time, and this relationship can be used as a yardstick for determining the dosage of the viscosity agent. In the tests, in the cases where the dosage of the viscosity agent was 900 and 600 g/m3, segregation was observed visually. In the V-funnel tests, however, there were cases in which funnel blocking did not occur. This result indicates that there are cases in which funnel blocking did not occur even when the concrete whose segregation had been visually observed was subjected to V-funnel testing. This suggests that it is necessary to express segregation properties quantitatively. b) Fresh mortar Figure 4 shows the relationship between the properties of mortar obtained by wet-screening produced highly-flowable steel fiber-reinforced concrete mixes with a 5

50

55

60

65

70

75

80

0.25 0.5 0.75 1 1.25

VA:600g/m3

VA:900g/m3

VA:1200g/m3

VA:1500g/m3

VA:1800g/m3

Steel fiber content(vol%)

Vg:200L/m3

Slum

p flo

w(c

m)

0

10

20

30

40

50

60

70

300 600 900 1200 1500 1800 2100

Vf:1.0 vol%Vf:0.75 vol%Vf:0.5 vol%

V75

-fun

nel f

low

tim

e(se

c) (blockage)


Vg:200L/m3

280

300

320

340

360

380

400

420

0

5

10

15

20

25

30

35

300 600 900 1200 1500 1800 2100Viscosity agent(g/m3)

Kﾛｰﾄ流下時間

Mini slunp flowy=452*e^(-0.0002x) r=0.970

K-funnel flow-out time(sec)y=5.44*e^(0.0009x) r=0.970

Min

i slu

mp

flow

(mm

) K-funnel flow

time(sec)

50

55

60

65

70

75

80

280 300 320 340 360 380 400 420

Vf:1.0vol%Vf:0.75vol%Vf:0.5vol%

Mini slump flow(mm)

Vg:200L/m3

Slum

p flo

w(c

m)

Figure 3. Relation between viscosity agent and V-funnel flow

Figure 4. Relation between viscosity agent and mini slump flow

19 CMT - 149


CMT -

mm sieve to remove the coarse aggregate and steel fibers, and the dosage of the viscosity agent. As shown, as the dosage of the viscosity agent increased, mini slump flow decreased linearly and K-funnel flow time increased exponentially. This indicates that the viscosity agent influences both the viscosity and flowability of the mortar, and this result agrees with previous study results (Hujiwara et al. 1992). Figure 5 shows the relationship between mini slump flow and slump flow. As mini slump flow increased, slump flow increased linearly. This relationship is influenced by the steel fiber content, and slump flow decreased as the steel fiber content increased. It has been shown11) that when the absolute unit volume of coarse aggregate is constant, the slump flow and the mini slump flow of highly flowable concrete are highly correlated (Hujiwara et al. 1992). Thus, it has been confirmed that there is a similar relationship for highly-flowable steel fiber-reinforced concrete, too. c) Segregation properties Figure 6 shows the relationship between the dosage of the viscosity agent and the segregation index for different steel fiber contents. As shown, as the dosage of the viscosity agent increased, the steel fiber content and the segregation index decreased exponentially, indicating that as the viscosity of the concrete increased, segregation resistance increased. It is also shown that the segregation index is not significantly influenced by the steel fiber content. Comparison of the segregation indexes of the coarse aggregate and steel fibers reveals that the segregation index of the steel fibers is greater than that of the coarse aggregate. Since the degree of change in the segregation index of the steel fibers due to changes in the dosage of the viscosity agent is higher than that of the coarse aggregate, it is thought that the segregation index of steel fibers is easier to use than that of coarse aggregate. It is appropriate, therefore, to evaluate the segregation properties of highly-flowable steel fiber-reinforced concrete in terms of the segregation index of steel fibers. There were two cases (i.e., the case in which the dosage of the viscosity agent is 900 g/m3 and the case in which it is 600 g/m3) in which segregation was observed visually. The segregation indexes in these cases were equivalent to segregation indexes of steel fibers greater than 30%. The segregation indexes of coarse aggregate did not show significant differences, so it is difficult to determine threshold values of the segregation index. Figure 7 and Figure 8 show the relationship between the properties of mortar and the segregation index. As K-funnel flow time, which indicates the plastic viscosity of the mortar, increased, the segregation index decreased exponentially. As mini slump flow, which indicates the yield value of the mortar, increased, the segregation index increased exponentially. This indicates that segregation properties are highly correlated with the rheological properties of mortar. Thus, a K-funnel

0

10

20

30

40

50

300 600 900 1200 1500 1800 2100


SI(%

)

SI(SF)y = 105 * e^(-0.0014x) r= 0.957

SI(G)y = 49.6 * e^(-0.0021x) r= 0.833


0

10

20

30

40

50

280 300 320 340 360 380 400 420


SI(S

F)(%

)

Mini slump flow(mm)

Vg:200L/m3

SI(SF)y = 0.018* e^(-0.0194x) r= 0.955

0

10

20

30

40

50

5 10 15 20 25 30 35


SI(S

F)(%

)

K-funnel flow time(sec)

SI(SF)y = 81.5* e^(-0.0835x) r= 0.945

Vg:200L/m3

Figure 7. Relation between K-funnel flow time and segregation index

Figure 8. Relation between mini slump flow and segregation index

Figure 6. Relation between viscosity agent and segregation index

19 CMT - 150


CMT -

flow time of about 13 seconds or more and a mini slump flow of about 380 mm or less can be used as yardsticks of mortar properties that do not cause segregation of steel fibers.

3.2 The influence of the absolute unit volume of coarse aggregate and the steel fiber content on fresh properties (Series 2) a) Fresh concrete Figure 9 shows the relationship between the absolute unit volume of coarse aggregate and slump flow. Slump flow tended to increase linearly as the absolute unit volume of coarse aggregate increased. This tendency was observed regardless of the steel fiber content. The likely reason is that as the absolute unit volume of coarse aggregate increased, the fine aggregate content decreased in relative terms so that the amount of water confined by fine aggregate decreased and slump flow increased. As the absolute unit volume of coarse aggregate changed by 50 L/m3, slump flow changed by about 5 cm. Figure 10 shows the absolute unit volume of coarse aggregate and the V-funnel test results. As the absolute unit volume of coarse aggregate increased, funnel flow time showed a downwardly convex curve. According to the V-funnel test results, when the steel fiber content was 1.0 vol.%, V-funnel flow time decreased until the absolute unit volume of coarse aggregate reached 200 L/m3, and the funnel was blocked at 225 L/m3. When the steel fiber content was 0.75 vol.%, funnel blocking did not occur even when the absolute unit volume of coarse aggregate reached 225 L/m3, and the funnel flow time curve was downwardly convex. Funnel blocking occurred, however, at the absolute unit volume of coarse aggregate of 250 L/m3. These results indicate that there are combinations of the absolute unit volume of coarse aggregate and the steel fiber content that provide excellent deformability. b) Fresh mortar Figure 11 shows the properties of wet-screened mortar. As the absolute unit volume of coarse aggregate changed, mini slump flow and K-funnel flow time changed linearly. As the absolute unit volume of coarse aggregate increased, mini slump flow increased and K-funnel flow time decreased. These results are consistent with the observed fact that slump flow increased as the absolute unit volume of coarse aggregate increased. c) Segregation properties Figure 12 and Figure 13 show the results of the concrete segregation test in which the absolute unit volume of coarse aggregate and the steel fiber content were varied. As shown, the steel fiber content

280

300

320

340

360

380

400

420

0

5

10

15

20

25

30

35

125 150 175 200 225 250 275

Min

i slu

mp

flow

(mm

) K-funnel flow

-out time(sec)

Vg(L/m3)

Mini slunp flowy=169+0.923x r=0.983

K-funnnel flow-out timey=51.3-0.171x r=0.991

○，□：Vf：0.75vol％●，■：Vf：1.0vol％

0

10

20

30

40

50

125 150 175 200 225 250 275

Vf=0.75vol%Vf=1.0vol%

V-f

unnn

el fl

ow ti

me(

s)

SP:Cx2.0%

VA:1200g/m3

Vg(L/m3)

(blockage)

Figure 9. Relation between Vg and slump flow

Figure 10. Relation between Vg and V-funnel flow time

Figure 11. Relation between Vg and mini slump flow

50

55

60

65

70

75

125 150 175 200 225 250 275

Vf=0.75vol%Vf=1.0vol%

Slum

p flo

w(c

m)

Vg(L/m3)

SP:Cx2.0%

VA:1200g/m3

19 CMT - 151


CMT -

and the segregation index of coarse aggregate are strongly correlated with K-funnel flow time and mini slump flow, which are indicators of mortar properties. As K-funnel flow time increased, the segregation index decreased; and as mini slump flow increased, the segregation index increased. These relationships were more or less the same regardless of the steel fiber content. Thus, it has been confirmed that the test results are similar to the Series 1 test results.

3.3 The influence of different factors under constant slump flow conditions (Series 3) The effects of changes in different factors on fresh properties under constant slump flow conditions were investigated experimentally. The target level of slump flow was set at about 60 cm, and the dosages of the air-entraining water reducer and the viscosity agent were adjusted. In the test, slump flow ranged from 57 to 63 cm, which were mostly within the target range. Figure 14 shows the segregation test results. All mixes subjected to the test were visually inspected to make sure that segregation had not occurred. The segregation indexes of steel fibers were not higher than about 20%. The relationship between K-funnel flow time and the segregation index showed a strong correlation. Studies have shown(Kurita et al.,1996) that under constant plastic viscosity conditions, the segregation resistance of highly flowable concrete is governed by the yield value, and that under constant yield value conditions, segregation resistance is governed by plastic viscosity. Under more or less constant yield value conditions as in the tests conducted for the purposes of this study, therefore, the segregation index is strongly correlated with K-funnel flow time. It may be concluded, therefore, that segregation resistance can be evaluated in terms of the viscosity of mortar. Figure 15 shows the relationship between the absolute unit volume of coarse aggregate and V-funnel flow time. In the case where the steel fiber content was 0.5 vol.%, V-funnel flow time did not change significantly until the absolute unit volume of coarse aggregate reached 275 L/m3. After that, V-funnel flow time increased sharply. In the case where the absolute unit volume of coarse aggregate was 300 L/m3, the V-funnel was not blocked, either. In the case where the steel fiber content was 0.75 vol.%,

0

10

20

30

40

50

0 5 10 15 20 25 30 35 40

SI(%

)

K-funnel flow-out time(sec)

○，□:Vf:0.75vol%●，■:Vf:1.0vol% Vg:150-250 L/m3

SI(G):y = 13.7 * e^(-0.07x) r= 0.639

SI(SF):y = 53.7 * e^(-0.060x)r= 0.955

0

10

20

30

40

50

280 300 320 340 360 380 400 420

SI(%

)

Mini slump flow(mm)

SI(G):y = 0.073* e^(0.011x)

r= 0.600

SI(SF):y = 0.37 * e^(0.011x)

r= 0.970

○，□:Vf:0.75vol%●，■:Vf:1.0vol% Vg:150-250 L/m 3


Figure 13. Relation between mini slump flow and segregation index

0

10

20

30

40

50

0 5 10 15 20 25 30 35

Slump flow: 57-63cm

Vf: 0.5-1.0vol%, Vg:150-300L/m3

y = 49.6 * e^(-0.062x) r= 0.777

SI(%

)

K-funnel flow time(sec)

0

20

40

60

80

100

125 150 175 200 225 250 275 300 325

Vf=0.5vol%Vf=0.75vol%Vf=1.0vol%

Vg(L/m3)

V-f

unne

l flo

w ti

me(

sec) (blockage)

(blockage)



19 CMT - 152


CMT -

F-funnel flow time increased gradually until the absolute unit volume of coarse aggregate reached 250 L/m3, and funnel blocking occurred at 275 L/m3. In the case where the steel fiber content was 1.0 vol.%, V-funnel flow time increased with the absolute unit volume of coarse aggregate, increased sharply after 200 L/m3 was reached, and funnel blocking occurred at 250 L/m3. After V-funnel flow time exceeded about 20 seconds, deformability tended to decline sharply. It is thought that this may be used as an indicator of the deformability of highly-flowable steel fiber-reinforced concrete. These results indicate that under more or less constant slump flow conditions, the deformability of highly-flowable steel fiber-reinforced concrete is greatly affected by the combination of the absolute unit volume of coarse aggregate and the steel fiber content, and that in cases where the steel fiber content is fixed, there is an upper limit to the absolute unit volume of coarse aggregate that provides excellent deformability. Under the test conditions covered in this study, the following combinations of the steel fiber content (Vf) and the absolute unit volume of coarse aggregate can be used as rough guidelines for the mix design of highly-flowable steel fiber-reinforced concrete: • For a steel fiber content of 0.5 vol.%, use an absolute unit volume of coarse aggregate of about 275 L/m3 or less. • For a steel fiber content of 0.75 vol.%, use an absolute unit volume of coarse aggregate of about 250 L/m3 or less. • For a steel fiber content of 1.0 vol.%, use an absolute unit volume of coarse aggregate of about 200 L/m3 or less.

4. CONCLUSION

With the aim of gaining knowledge about the fresh properties of highly-flowable steel fiber-reinforced concrete, the effects of three mix design factors, namely, the absolute unit volume of coarse aggregate, steel fiber content and the dosage of the viscosity agent, on flowability, deformability and segregation resistance were investigated experimentally. The findings of this study can be summarized as follows: (1) The slump flow of highly-flowable steel fiber-reinforced concrete is influenced by the absolute unit volume of coarse aggregate, steel fiber content and the dosage of viscosity agent. As the absolute unit volume of coarse aggregate changes by 50 L/m3, slump flow changes by about 5 cm. As the steel fiber content changes by 0.5 vol.%, slump flow changes by about 5 cm. (2) The segregation properties of highly-flowable steel fiber-reinforced concrete should be evaluated in terms of the segregation index of steel fibers. The segregation properties are highly correlated with the properties of wet-screened mortar. Segregation does not occur when K-funnel flow time is about 13 seconds or more or mini slump flow is about 380 mm or less. (3) Under constant slump flow conditions, segregation resistance can be evaluated in terms of the viscosity (K-funnel flow time) of mortar. (4) The deformability of highly-flowable steel fiber-reinforced concrete can be evaluated through V-funnel testing. Under constant slump flow conditions, there are combinations of the absolute unit volume of coarse aggregate and the steel fiber content that provide excellent deformability.

5. REFERENCES

Fujiwara, H., Shimoyama, Y., Tomita, R. and Kubota, Y.(1992). “Fundamental study on filling ability of highly flowable concrete.” (in Japanese) Proceedings of the Japan Concrete Institute, Vol. 14, No. 1, pp. 27–32.

Kawai, T., Hashida, H., Kuroda, Y. and Inoue, H.(1993). “Fundamental study on rheological properties of highly flowable concrete (part 2, characteristics of mortar). (in Japanese), Summaries of Technical Papers of Annual Meeting, AIJ, pp. 1129–1130.

Kawai, T., Kuroda, Y. and Takekawa, Y.(1993). “Experimental study on properties of highly flowable concrete containing low-heat cement.” (in Japanese), Proceedings of the Japan Society of Civil Engineers, No. 462/VI-18, pp. 111–120.

19 CMT - 153


CMT -

Kurita, M. and Tanaka, H.(1998). “Fresh properties of highly flowable concrete mixed with steel fibers.” (in Japanese), Proceedings of the 53rd Annual Conference of Japan Society of Civil Engineers, V-269, pp. 538–539.

Kurita, M. and Tanaka, H.(1999). “The influence of mix design factors on fresh properties of highly flowable concrete mixed with steel fibers.” (in Japanese), Proceedings of the 54th Annual Conference of Japan Society of Civil Engineers, V-453, pp. 906–907.

19 CMT - 154


CMT -

POLYCARBOXYLATE BASED ADMIXTURES: THEIR CHARACTERISTICS AND ROLE IN MANAGING CONCRETE

PROBLEMS

Baha Abdelrazig1

ABSTRACT:Concrete has been the major construction material in the past, present and is going to be with us for sometime to come. One of the main reasons for this is its continual adaptation for the ever changing demands. Admixtures have played a major role in upgrading the concrete quality & improving its versatility over the past several decades. Hot climates pose a special challenge and concretes placed in such environment may suffer loss in fresh properties and consequently strength & durability. Problems are frequently encountered in producing good quality concrete in hot climates. Inadequate workability in placing results in cold joints and consequently porous and permeable concrete., The use of cement alone results in thermal gradients which may lead to cracking. Waste materials are more frequently used and recycling of building materials may become a hot topic in the near future.. Changes in availability and price of aggregates and cement could have a significant effect. It is therefore more important that admixtures be insensitive to quality changes in materials, achieve fluid, stable, cohesive and durable concrete at low water additions. Recently, more prestigious projects with increased architectural concepts in ambition and size are being built. Longer pumping distances and stringent water tightness qualities are often needed This has placed additional serious challenges on normal super-plasticized concrete. Work to mitigate structural failures in concrete is decades old endeavourer. This work has culminated in a series of new raw materials. In turn these new materials have necessitated new chemistry in concrete admixtures. Polycarboxylates (PCEs), are a new generation of 3-dimensional comb-type polymers . They have gained wide acceptance as powerful dispersants. The polymer chemistry can be used to customize admixtures to meet the needs of specific jobs. They have been frequently used to address many of the above site problems. In this presentation, tackling some concrete problems with the use of PCE based admixtures shall be discussed. Case studies from different parts of the World ,but particularly Asia, shall be presented.

1. CONCRETE IN HOT CLIMATES

Concrete placed in a hot environment may suffer a loss of strength and durability for several reasons. First, the higher the early curing temperature of concrete, the more rapid the rate of early-age strength gain, but the lower will be the long-term strength of the concrete. Second in dry environment accelerated water loss due to evaporation leads not only to a decreased degree of hydration in the surface , but can also lead to shrinkage and accompanying shrinkage stresses in the freshly place concrete. During the first few hours after batching the concrete has little or no tensile strength, making it particularly vulnerable to plastic shrinkage cracking. The drying which initiate this cracking begins the moment that the rate of evaporation of water from the surface of the concrete exceeds the rate at which bleed water is supplied to the surface. The earlier the drying begins, the lower the cracking resistance of concrete. Whether or not cracking occur further depends on the complicated and time-dependant relationships between water loss and shrinkage, shrinkage strain and shrinkage stress, and the continuous race between development of both shrinkage stress and tensile strength. Cracks caused by rapid and early drying can be very deep, range in width between 0.1 mm and 3 mm and as long as one metre or more. Moreover, drying terminates the continued hydration of the cement, increasing permeability at the surface and decreasing abrasion resistance. Efficient curing is therefore important for maintaining uniformity of concrete surfaces as well as ensuring proper hydration and

1 Sika Regional Technical Support Centre Asia Pacific, Lot 689, Nilai Industrial Estate, 71800 Nilai, NSDK, Malaysia

20 CMT - 155


CMT -

strength development. Curing is even more important in good quality concretes made with cement replacement materials and superplasticized concrete when the bleeding rate is near zero. If the concrete has zero connected porosity and insufficient water available for the processes of dissolution and reaction to occur, then the degradation reactions would be much reduced and perhaps even stopped. Transport processes are the root of degradation and service life and water is the medium in which transport or movement can accur. To achieve durable concrete and structures, it is also necessary in the first place to select materials and mix proportions which, in hardened concrete, are inherently both physically and chemically resistant to their operating environment. However, this alone is not enough. The concrete must be designed such that it can be placed and compacted with minimum defects. It must be recognized that changes in the availability, chemistry and price of aggregates and cement have occurred over time and could have a significant effect. The use of cement alone results in thermal gradients which may lead to cracking. Concretes containing mineral additions can be beneficial in all those respects. High temperatures result in reduced workability and accelerated setting, with the increased risk of poor compaction or “cold” joints. Once the concrete is in place drying can occur, leading to plastic cracking as discussed above. The high initial mix temperature accelerates the rate of hydration, resulting in high early temperature rise in cement rich or thick sections. This increases the risk of early thermal cracking and , while accelerating the early strength gain, may impair the long term development of strength and other properties, including those influencing durability. This in turn can cause premature deterioration of either the concrete itself or the reinforcement. To reduce the incidence of deterioration, concreting materials should impart the following properties to concrete:

• Increased resistance to environmental effects and/or high tolerance to them. • Improved and extended workability or reduced water demand. • Reduced rate of heat generation during hydration. • In hot climates, tolerance of and , if possible, benefit from elevated temperatures.

It is also essential that other properties are not adversely affected to the extent that another deterioration mechanism becomes predominant. Chemical admixtures have played a key role in the past decades in improving the quality and upgrading the performance of concrete. Today they are required more than any other time to be insensitive to quality changes in materials, achieve fluid, stable, cohesive, long pumping distances and durable concrete at low water additions. The emergence of the third generation polycarboxylate Ether (PCE) based admixtures have made it possible to meet such requirements and address the many problems. While PCE based admixtures may sometimes come at a cost, the question ,in the short term, should not be “ can we afford to use them?’ but “Can we afford not to?” This can only be answered by recognizing the benefit in whole life cost which can be achieved, the many difficulties that can be overcome and problems that can be solved.

2. POLYCARBOXYLATE ETHER (PCE) BASED ADMIXTURES

Work to mitigate structural failures in concrete is a decades-old endeavoure. This work has culminated in a series of new raw materials. In turn, these new materials have necessitated new chemistry in concrete admixtures. Polycarboxylates (PCE) , in particular, have gained wide acceptance as powerful dispersants in admixtures. The polymer chemistry can be used to customize admixtures to meet the needs of specific construction jobs. They are the third generation of water reducers and are manufactured by the combination of a number of building blocks which are synthesized by polymerization. The new generation comb-type polycarboxylate polymers are ideal for:

Powerful plasticizing adding to the strength Special formulations to keep concrete cohesive and homogeneous. Controlled workability.

20 CMT - 156


CMT -

A greater water reducing effect in concrete can be obtained by enhanced dispersibility and stable dispersion of the cement particles. The retention of dispersion of inorganic microparticles is due to electrical and steric repulsion of the adsorbed surfactant. Their chemical and physical properties are controlled by:

• Backbone • Side chain length & density • Electric Charges, and • Functional groups of the polymer.

It is, therefore, clear that the polymer chemistry can be used to customize admixtures by regulating the above parameters or other variables to meet the needs of specific construction jobs. PCEs can be made by design to affect water reduction, workability retention and can also be formulated to keep the concrete cohesive and homogeneous. Stable dispersion due to electrical repulsion can be explained by the well known DLVO theory. The greater this energy barrier the more stable the dispersion, which is found to correlate well with the value of the zeta potential. The repulsion due to the steric effect can be explained by entropy effect theory. The water reducing effect of cement composites such as concrete is obtained by increasing the dispersion of the cement particles. The water reducing effect in cement composites such as concrete is obtained by increasing the dispersion of the cement particles. Water reducing agents are roughly divided into two types, those which enlarge the zeta- potential and increase the repulsion and those which increase the force of repulsion by sterically expanding the adsorption layer. Melamine & naphthalene sulphonate formaldehyde condensates based water reducers are adsorbed in the shape of a rod in several layers, in which the cement particles are dispersed due to the strong electrical repulsion of the negative ions of the sulfonate group. The size of this repulsion can be estimated by measuring the zeta potential of the surface of the cement particle. For polycarboxylate based water reducing agents, cement particles are dispersed and water reducing effect is obtained by the electrical repulsion of the negative ions of the carboxylic group and steric repulsion of the main and side chains (6). Therefore PCE based water reducing agents can give water reductions equal to that of the conventional admixtures at much lower dosages or much higher water reduction at equal dosages. Diagrammatically this can be illustrated as shown overleaf: Conventional High Range Water Reducers disperse the cement particles through electrical repulsion give Water reductions of up to 20%.

20 CMT - 157


CMT -

The new generation three dimensional comb-type polycarboxylate polymers disperse the cement particles through a double action of electrical repulsion and steric hinderance resulting in higher water reductions up to 40% and at the same time much longer slump retentions. The Sika technology of PCE based admixtures is termed ViscoCrete technology. Since its introduction in the late nineties, many projects around the world have been executed using it, giving solutions to many difficult problems . In the next few pages some challenges are highlighted and the solutions provided by the use of PCE based admixtures, through the maximization of the design freedom, enhanced concrete placement leading to increased productivity and enhanced overall working environment, are discussed.

3. SELECTED CASE STUDIES

Challenge 1: Vertical & Long distance pumpability 1.1 Cut and Cover Tunnel in SMART 1.1a: SMART stands for Stormwater Management And Road Tunnel. Two objectives:

1) To channel stormwater from the Klang & Ampang rivers to the Taman Desa retention pond.

2) Part of the tunnel will also be used as a motorway when not fully flooded.

Project details: Project Cost RM 2.5bn 9.7km water-cum-motorway tunnel 4km motorway - 49,000 cars per day 15 mins to clear traffic : fully used

to channel floodwater. Completion scheduled for 2006

1.1.b : Requirements • Concrete Grade 40 • High workability of the concrete • Pump mix design • Long Slump life • Initial slump < 200 mm • Minimum slump at 2 hrs 125 mm • Reduction in cement (plus) • Single admixture (plus) • High placing temperatures

Figure 1. SMART Tunnel route. Inset Showing the two deck, 3 chambers

tunnel.

20 CMT - 158


CMT -

1.1.c Concrete mix design: • Mascrete 390 kg/m3 • Coarse Agg. 974 kg/m3 • Sand 834 kg/m3 • Water 156 l/m3 • Admixture: (PCE) @ 2l/m3

1.1.d Typical Results: • W/C ratio 0.4 • Slump Initial 180 mm • At 2 hrs 120 mm • Good pumpability at low pump pressure • Compressive strength

(target + Margin achieved) • Cement reduction achieved • Single admixture • At Competitive price

1.2) Vertical Pumping 1.2a Project: 450 m building in Hong Kong 1.2b Requirements:

• SCC, C100 concrete • Low and stable concrete viscosity • To be pumped up to 480 m up • Constant slump life for 180 min even

after pumping

1.2c Mix design: • OPC 358 kg • PFA 192 kg • M/silica 42 kg • 10 mm 1000 kg • R sand 680 kg • Water 130 kg (w/b= 0.21) • Admixture (PCE) @ 6l/m3

Figure 3. First trial pour of 400 m3

Figure 2. Cut & cover Tunnel

20 CMT - 159


CMT -

Figure 7. Concreting of upper deck

1.2d Results:

Figure 4. Initial Slump Flow Figure 5. Slump Flow after pumping 450m up and 170 min. No bleeding or segregation. 1.3 Long distance pumpability 1.3a Project: SMART Road Deck- Malaysia 1.3b Requirements:

• Internal concreting for decks & walls pumped from tunnel shaft Pumping Distance up to 1500m !

• Concrete Grade 40 • Initial slump 220 mm • Slump at 3 hrs not less than

160 mm • 24 hr strength 14 N/mm2.

1.3c Mix design:

• Mascrete 380 kg/m3 • LSP 40 kg/m3 • C. Agg 940 kg/m3 • Sand 801 kg/m3 • Water 172 kg/m3 • Admixture (PCE): @ 2.3 l/m3.

1.3d Long Distance Pumping Trial

• Batch slump – 200mm – started pumping after 40mins

• 200m length – Slump 200mm - Conc Temp 30C

• 600m length – Slump 200mm • 1000m length – Slump 160mm • 1400m length – Slump 155mm

Figure 6. Two deck tunnel for stormwater & traffic

20 CMT - 160


CMT -

Figure 8. Concrete after 2000m pumping

• 2000m length – Slump 140mm - Conc Temp 34C • Initial set 5h • Final set 6.5h • 24h strength 14 MPa

Challenge 2: Water tight concrete to replace membranes 2.1 Watertight Tunnel 2.1a Project : Watertight KLCC Tunnel- Malaysia This tunnel connects KLCC towers with KLCC convention centre. 2.1b: Requirements

• Replace external Waterproofing Membrane with high- performance, Watertight Concrete. - Permeability & Absorption specified.

• Eliminate potential workmanship problems with concrete placement and waterproofing installations.

• 28d Compressive Strength: - Min 50 N/mm2

• Hotel proximity to Site: Reduce or Eliminate Noise - Self-Compacting Concrete - Flow: Min 600mm @ 2hours 2.1b: Mix Design

• Low heat Mascrete 450kg/m3 • Sand 700kg/m3 • 20mm aggregate 595kg/m3 • 10mm aggregate 455kg/m3 • w/c ratio 0.35 • Admixture (PCE based) 1.5%bwoc • Integral waterproofing admixture 1.5%bwoc • Viscosity Modifying Agent 0.6% bwoc

2.1c: Results SCC used to construct all floors, walls and soffits of tunnel. The mix design included an integral waterproofing admixture and a viscosity modifying agent to give the mix more cohesion and avoid any bleeding and segregation. Hardened concrete achieved both permeability ( RPC < 1000 coul.) & absorption ( < 2%) to BS 1881 specification allowing the elimination of waterproofing membranes in tunnel construction.

20 CMT - 161


CMT -

Challenge 3: Non Stop pouring of 28000m3 of uniform quality concrete: 3.1a: Project: Mori Tower- Shanghai China 3.1b:Project details

• Project period: end 2004 - end 2007 • High-rise building: 492m high, • 101 floors above ground with 3 basements plus a foundation with depth of 4.5m – 4.7m • Total concrete volume: 300,000 m3 including SC& CFT from grades C30 to C60. • Admixture: PCE based 1200 tons in total

3.1c Requirements Table 1. Three Foundation Concrete Pours with congested reinforcement

Involved

Concrete Pouring

Date Started

Hours of Pouring

RMC plants Trucks

1 4000m3 26/12/04 15 4 80

2 4500m3 08/01/05 11 5 100

3 28’000m3 28/01/05 40 7 350

Figure 9. A good base with cast in drainage channels

Figure 10. Excellent compaction around water bars

Figure 11. Smooth walls treated with curing agent for optimum curing. Figure 12. Finished tunnel as it stands today

20 CMT - 162


CMT -

Figure 15. Congested reinforcement

3.1d Mix Design & Results: OPC 270 kgs PFA 90 kgs Slag 70 kgs Sand 715 kgs Aggregate (5-25mm) 1029kgs Water 170 kgs Admixture (PCE) 0.8% by wt. (OPC + Slag) Designated initial slump 200mm 2-hr slump 150+30mm

Figure 13. Trucks in que for turn Figure 14. More than 6 pumps working simultaneously

Figure 16. Normal Slump concrete flowing with a bit of help to spread the concrete.

20 CMT - 163


CMT -

4. CONCLUSIONS

• The use of PCE based admixtures has allowed the optimization of concrete properties and has lead to incremental advances in design and placement of concrete.. Long distance pumpability, water tightness, long slump retentions at low water additions, making of slender and more complex shapes have all been made simple. PCE based admixtures have maximized the design freedom leading to the production of SCC, CFT & ultra high strength concrete.

• SCC can eliminate or drastically reduce the need for vibration, saves time making it possible to reduce labour costs while improving the overall work environment and gives higher casting guarantees of uniformity and quality even of complex shapes.

• Faster placement, reduced construction time and less finishing time can improve productivity & profitability. Increased flowability and consolidation can improve appearance and enhance the durability of finished elements.

5. REFERENCES

Menzel, C.A (1954). “Causes and prevention of crack development in plastic concrete.” Proceedings of the Portland cement association, Annual meeting, pp130-136.

Hover, K.C. (1992). “ Evaporation of surface moisture: a problem in concrete technology and human physiology.” Concrete in hot climates, Proc. Of the third Int. Rilem conference, Torquay, England, Sept. 12-25.

Soroko, I. (1992). “Concrete in Hot Environment.” Chapman and Hall, London. ACI (1989). “Hot Weather Concreting.” ACI 305R-89, Detroit. Abdelrazig, B.(2004). “A contribution towards the study on: developing applications of SCC to

enhance quality, cost effectiveness, buildability and to reduce noise in public housing construction by City University of Hong Kong.“ Sika Internal publication.

20 CMT - 164


CMT -

EFFECT OF A NEUTRALISED BAUXITE REFINERY RESIDUE AS FINE AGGREGATE ON PROPERTIES OF CEMENT MORTAR

S A Barbhuiya1, P A M Basheer2, D McPolin3, B Sengupta4, G I B Rankin5, M W Clark6

ABSTRACT: Bauxite refining residues (BRR) from Bayer alumina refining are being generated at a rate of about 70 million tons per annum. This caustic material represents a major disposal problem in the alumina industry; hence many attempts have been made over the years to find a use for BRR. However, despite some 40-50 years of research, no significant quantity is actually being utilised anywhere in the world. A neutralised BRR is different from caustic BRR, the alkalinity of which is converted from soluble to insoluble forms and the causticity is substantially reduced. Therefore, there is the possibility of using neutralised BRR in concretes, as a means of its disposal, or to provide some chemical benefits to concrete, which are being investigated in a research project by the authors. In this paper, early results of an experimental investigation to evaluate properties of cement mortar for which natural sand was replaced with a neutralised BRR at 0%, 5%, 10% and 20% by mass of cement are reported. The properties presented in this paper are the consistency of fresh cement mortar and the compressive strength of the hardened cement mortar at 1, 3, 7, 28, 56 and 90 days. These results indicated that the water demand of fresh cement mortar increases with an increase in the neutralised BRR content. At a fixed water-cement ratio (W/C), the compressive strength of cement mortar containing different quantities of neutralised BRR was found to be comparable with that of the control mix containing natural sand. At a fixed flow (a measure of the consistency), the compressive strength of cement mortar containing neutralised BRR was found to be comparable with that of the control up to a replacement level of 10%, but the strength decreased when the replacement level was 20%.

KEYWORDS: Compressive strength, consistency, mortar, neutralised bauxite refinery residues.

1. INTRODUCTION

Bauxite refining residues (BRR) are a by-product of the bauxite refining to produce alumina using the Bayer process and globally this generates about 70 million tons of BRR per annum. BRR is a complex material where the chemical and mineralogical composition may vary, depending upon the source of the bauxite and the plant operations (Komnitsas et al., 2004). It is highly caustic owing to sodium hydroxide (caustic soda) used in the Bayer process to solubilise the alumina from the bauxite and, hence, this represents a major disposal problem in the alumina industry. BRR usually has a pH greater than 13.5 and exits the process stream as slurry with 15-30% solids (Zambo and Solymar, 1973; Ostap, 1984; Brunori et al., 2005). The production of 1 ton of alumina generally results in the generation of 1-1.5 tons of BRR. The disposal of such a large quantity of this alkaline by-product is expensive (up to 1-2% of the alumina price) because it requires a lot of land. At present, it is disposed of mainly by dumping on land in constructed dams and dykes, or in few cases, in natural valleys; a small percentage (nearly 10%) is disposed of in the sea. However, the dumping of BRR with high metal values along with pollutant alkali is not particularly far sighted and creates several environmental problems (Das et al., 1988).

1 Research Scholar, Centre for Built Environment Research, Queen's University Belfast, Northern Ireland, UK. 2 Professor & Director, Centre for Built Environment Research, Queen's University Belfast, Northern Ireland. UK. 3 Post-doctoral Research Fellow, Centre for Built Environment Research, Queen's University Belfast, Northern Ireland, UK. 4 Lecturer, Environmental Engineering Research Centre, Queen's University Belfast, Northern Ireland, UK. 5 Manager, Construction Division, Northern Ireland Technology Centre, Queen’s University Belfast, Northern Ireland, UK. 6 Senior Research Fellow, School of Environmental Science, Southern Cross University, Australia.

21 CMT - 165


CMT -

Over the years, many attempts have been made to find a use for BRR and initial attempts were based mainly on vegetating the abandoned ponds (Das et al., 1988). However, this is not an ideal solution because a pollutant covered or stored is not the solution, but temporarily pushing the problem out of sight. Moreover, since BRR contains few nutrients, humus or organic matter, these must be supplied from external sources to sustain plant growth. Research work has been done on the use of BRR as a partial substitute for clay in the production of bricks and other ceramic products as well (Wagh and Douse, 1991; Newton, 1997; Valcin and Sevinc, 2000; Sglavo et al., 2000). Bricks containing BRR using a sodium silicate binder have been made in Jamaica (Newton, 1997). Addition of reactive silica (often sodium silicate) to the caustic sodium aluminate in the BRR allows the formation sodium alumina-silcates (water glass), which effectively glues the particles together (Sagoe-Crestsil and Brown, 2005). This technique does not require firing of the bricks and, although not as strong as fired bricks, they are substantially cheaper to produce (Newton, 1997). Despite some 40-50 years of research, no significant quantity of BRR has actually been utilised anywhere in the world. However, BRR can be neutralised using the patented Basecon™ technology, which allows the conversion of soluble alkalinity in BRR into low soluble minerals (essentially Ca and Mg hydroxides, carbonates and hydrocarbonates; McConchie et al., 2002). The effect of this is that the pH falls from 13 down to about 9 (Genc-Fuhrman et al., 2003). The neutralised BRR material provides acid buffering capacity, excellent trace metal and phosphorous binding and, hence, it is being used in environmental remediation (Genc-Fuhrman et al., 2004). Further treatment of the material, such as acid treatment, heat treatment, addition of ferric sulphates or aluminium sulphates, has been studied by Genc-Fuhrman et al. (2004a) in order to enhance specific geochemical properties of the final product (McConchie et al. 2001). Due to the low causticity of the neutralised BRR and the potential for disposing this in cementitious materials, or for improving certain properties of hydrated cement, a comprehensive research work is in progress at Queen’s University Belfast, UK to explore its utilisation in manufacturing cement mortar and concrete. An initial study by the authors has indicated that the neutralised BRR has no pozzolanic properties and, hence, it is unsuitable as a cement replacement material. Therefore, the use of the neutralised BRR in cement mortar and concrete as a sand replacement material is being investigated. This paper reports the early results of an investigation in which the neutralised BRR was used as a partial sand replacement material in cement mortar.

2. EXPERIMENTAL PROGRAMME

2.1 Materials Class 42.5N Portland cement (OPC), complying with BS EN 197-1:2000, and medium graded natural sand obtained from local sources in Northern Ireland, complying with BS EN 12620:2002, were used to manufacture the cement mortar mixes. Where the neutralised BRR was used to replace the natural sand, the material from the Euralumina Company at Portoscuso, Sardinia, Italy was used. The physical properties of the OPC, the neutralised BRR and the natural sand are shown in Table 1. It can be seen that both the natural sand and the neutralised BRR have almost the same specific gravity (determined as per ASTM C 188-1995). The Blaine fineness (ASTM C 204a-1996) of the neutralised BRR ranged from 190 to 210m2/kg. A comparison of the particle size distributions of the OPC, the neutralised BRR and the natural sand is presented in Fig. 1. It can be observed that the neutralised BRR particles have a size distribution much finer than that of the natural sand, although grain sizes up to 100µm are observed (Fig. 1). The chemical composition of the OPC and the neutralised BRR, obtained using X-ray fluorescence (XRF) spectrometer, is shown in Table 2. This indicates that the neutralised BRR contains six major constituents, namely Fe2O3, Al2O3, SiO2, TiO2, Na2O and CaO. It also contains small quantities of numerous minor/trace elements (as oxides) such as V, Cr, Ce, Sc, Th, Nb, Pb, Ga, Sr, Ni, La, U, Cu, As, Ba, Co, Cd and Zn. Scanning electron micrographs of the neutralised BRR and the natural sand particles at two different magnifications are shown in Figs. 2 and

21 CMT - 166


CMT -

3 respectively. It can be seen that the neutralised BRR is a conglomerate of very small particles and appears to be relatively more porous in nature than the natural sand.

Table 1. Physical properties of materials used

Physical properties Specific gravity Blaine fineness (m2/kg) OPC 3.18 352-359

Neutralised BRR 2.75 190-210 Natural sand 2.72 -

0

20

40

60

80

100

0.1 1 10 100 1000 10000

Particle size (μm)

Cum

ulat

ive

pass

ing

(%)

OPCNeutralised BRRNatural sand

Figure1. Particle size distributions of OPC, neutralised BRR and natural sand

Figure 2. SEM of neutralised BRR particles at (a) 400x magnification (b) 1000x magnification and

(c) 3500x magnification

Table 2. Chemical composition of major elements in OPC and the neutralised BRR

Oxides (%) SiO2 Al2O3 Fe2O3 MgO CaO Na2O K2O TiO2 SO3 P2O5 LOI

OPC 21.41 5.11 2.61 1.78 61.50 0.33 0.61 -- 3.03 0.16 2.58 Neutralised

BRR 16.59 23.26 30.12 0.70 3.17 7.40 0.13 6.66 0.09 0.17 0.81

cba

21 CMT - 167


CMT -

Figure 3. SEM of natural sand particles at (a) 150x magnification (b) 300x magnification and (c) 500x magnification

2.2 Mix proportions The experimental work was carried out in two series- Series A and Series B. For both the series, the natural sand was replaced with the neutralised BRR at 0%, 5%, 10% and 20% by mass of cement and the aggregate-cement ratio was fixed at 2.75. Series A was designed at a fixed water-cement ratio (W/C) of 0.62 and Series B at a fixed flow of 400(±10) mm (to control the consistency) by adjusting the water content. The details of the mix proportions for Series A and Series B are summarised in Tables 3 and 4 respectively.

Table 3. Mix proportions for Series A

Mix ID OPC (g) Natural sand (g) Neutralised BRR (g) Water (g)

A-S100B0 1000 2750 0 620 A-S95B5 1000 2700 50 620 A-S90B10 1000 2650 100 620 A-S80B20 1000 2550 200 620

Table 4. Mix proportions for Series B

Mix ID OPC (g) Natural sand (g) Neutralised BRR (g) Water (g)

B-S100B0 1000 2750 0 590 B-S95B5 1000 2700 50 600

B-S90B10 1000 2650 100 620 B-S80B20 1000 2550 200 650

2.3 Manufacture and curing of samples Twelve 100 mm size cubes were cast for each mix to determine the compressive strength at 1, 3, 7, 28, 56 and 90 days. Cement mortar was manufactured in accordance with BS EN 1015-2:1999 using a 60 kg capacity pan mixer. All specimens were cast in two layers and compacted on a vibrating table until air bubbles appearing on the surface stopped. The specimens in their mould were covered with a plastic sheet and kept in the casting room at 20(±1) 0C for 24 hours. These were then demoulded and transferred to a water bath (maintained at 20(±1) 0C) for curing them in water. After 2 days of water curing the specimens were wrapped in polythene sheet and kept in an environmental chamber at

a b c

21 CMT - 168


CMT -

20(±1) 0C and 40(±1)% RH, where they remained until required for testing. One-day old cubes were tested just after demoulding, whereas those at the age of 3 days were tested immediately after removing from the water bath. The 2 days of water curing was adopted to simulate the curing regime in most concrete construction sites.

2.4 Test procedure The consistency of fresh cement mortar was measured by the flow table test, in accordance with BS EN 12350-5:2000. The compressive strength test was conducted as per BS EN 12390-3:2002 on three 100 mm cubes at the specified ages, using a universal testing machine and a consistent rate of loading of 0.2MPa/s.


3.1 Properties of fresh cement mortar The properties of fresh cement mortar are shown in Figs. 4 and 5 for Series A and Series B, respectively. Figure 4 shows that, at a fixed W/C of 0.62, the flow decreased with an increase in the neutralised BRR content, which suggests that there was an increase in water demand with the increase in the neutralised BRR content. For Series B, (Fig. 5), for which the water content was adjusted to have a fixed flow of 400(±10) mm, there was a uniform increase in the water demand to maintain the flow with the increase in the neutralised BRR content. Both data sets show that when the neutralised BRR was used as a sand replacement material in cement mortar, the water demand would increase. The increased water demand with an increase in the neutralised BRR content may be due to an increased particle surface area provided by the fine grained fraction of the neutralised BRR (Fig. 1), which requires more water to lubricate the particles. The increased water demand could also be attributed to the relatively more porous nature of the neutralised BRR when compared to the natural sand (Figs.2 and 3).

310

410450470

050

100150200250300350400450500

A-S100B0 (0%) A-S95B5 (5%) A-S90B10 (10%) A-S80B20 (20%)

Flow

(mm

)

Figure 4. Flow at fixed W/C of 0.62

21 CMT - 169


CMT -

3.2 Compressive strength Figures 6 and 7 present the compressive strength results of cement mortar at the age of 1, 3, 7, 28, 56 and 90 days at fixed W/C and fixed flow respectively, which show that the strength increased with age in both test series. At fixed W/C (Fig. 6), the compressive strength of cement mortar containing different quantities of neutralised BRR was comparable to that of the control mortar. At fixed flow (Fig. 7), the compressive strength of the cement mortar containing neutralised BRR was also found to be comparable with that of the control mortar for replacement levels up to 10%. However, the strength decreased markedly for cement mortar containing the 20% replacement level up to the age of 56 days. The strength improved beyond this age and was nearly comparable with that of the control at the age of 90 days. The decrease in compressive strength with the increase in the neutralised BRR content can be attributed to the increased free water content of these mixes while maintaining a constant flow value of 400(±10)mm. By combining the results of both the series it can be concluded that a 10% replacement of natural sand by the neutralised BRR in cement mortars would produce a compressive strength of 40MPa at an age of 28 days. Although the data presented suggests that a physically robust cement mortar can be produced by replacing natural sand with the neutralised BRR, it must be noted that when cement is first mixed with water, calcium-sulpho-aluminate-hydrate (ettringite) is formed as a result of chemical processes between calcium, sulphate, aluminate and hydroxyl ions. As the neutralised BRR contains a substantial quantity of alumina (Table 2), there is a distinct possibility that, depending on the alumina/sulphate ionic ratio of the solution, ettringite may form, which may become unstable and decompose to form mono-sulphate-hydrate. The presence of a mono-sulphate-hydrate makes the mortar vulnerable to sulphate attack (Mehta and Monteiro, 1993). Therefore, it is essential to examine the sulphate expansion of cement mortar containing neutralised BRR. In addition, the porous nature of the neutralised BRR aggregates may increase the permeation properties of cement mortar, which means that the effect of neutralised BRR on permeation properties should also be investigated before it is used in the manufacture of cement mortar.

0.65

0.62

0.6

0.59

0.560.570.580.590.6

0.610.620.630.640.650.66

B-S100B0 (0%) B-S95B5 (5%) B-S90B10 (10%) B-S80B20 (20%)

Wat

er-c

emen

t rat

io

Figure 5. Water demand at fixed flow of 400(±10)mm

21 CMT - 170


CMT -

05

101520253035404550

0 20 40 60 80 100

Age (days)

Com

pres

sive

stren

gth

(MPa

)

A-S100B0 (0%) A-S95B5 (5%)

A-S90B10 (10%) A-S80B20 (20%)

0

10

20

30

40

50

60

1 3 7 28 56 90

Age (days)

Com

pres

sive

stre

ngth

(MPa

)

A-S100B0 (0%) A-S95B5 (5%) A-S90B10 (10%) A-S80B20 (20%)

Figure 6. Compressive strength at fixed W/C of 0.62 (Series A)

21 CMT - 171


CMT -

4. CONCLUSIONS

On the basis of the results obtained for the materials investigated in this work the following conclusions have been drawn:

(i) The water demand of fresh cement mortar increases with an increase in neutralised BRR content.

(ii) At fixed W/C, the compressive strength of cement mortar containing neutralised BRR up to 20%

of the cement content to replace the natural sand was found to be comparable with that of the control mortar containing only natural sand.

05

101520253035404550

0 20 40 60 80 100

Age (days)

Com

pres

sive

stren

gth

(MPa

)

B-S100B0 (0%) B-S95B5 (5%)

B-S90B10 (10%) B-S80B20 (20%)

0

10

20

30

40

50

60

1 3 7 28 56 90

Age (days)

Com

pres

sive

stre

ngth

(MPa

)

B-S100B0 (0%) B-S95B5 (5%) B-S90B10 (10%) B-S80B20 (20%)

Figure 7. Compressive strength at fixed flow of 400(±10)mm (Series B)

21 CMT - 172


CMT -

(iii) At fixed flow, the compressive strength of cement mortar containing neutralised BRR was found to be comparable to the control mortar for replacement levels up to 10%. However, the strength was substantially lower for cement mortar containing 20% neutralised BRR replacement level up to an age of 56 days, but the strength improved beyond this age and was nearly comparable with the control at an age of 90 days.

(iv) It is possible to produce cement mortar having a 28 day compressive strength of around 40 MPa

using 10% of the natural sand replaced with the neutralised BRR. Although this work suggests that the cement mortar containing neutralised BRR is physically robust, further work is required to prove that durability properties such as permeability and sulphate expansions are not affected.

ACKNOWLEDGEMENTS

The authors would like to express their gratitude to Europe Virotec International Ltd for the financial support for carrying out the research. The facilities provided by the School of Civil Engineering at Queen’s University Belfast, UK for carrying out this research are also gratefully acknowledged. Commercial products that may be developed from this research are to be marketed by Greenhouse Funds Ltd, under the trade name ViroConcrete™.

REFERENCES

ASTM C 188 (1995). “Standard Test Method for Density of Hydraulic Cement.” Annual Book of ASTM Standards.

ASTM C 204a (1996). “Standard Test Method for Fineness of Hydraulic Cement by Air Permeability Apparatus.” Annual Book of ASTM Standards.

BS EN 197-1 (2000). “Cement: Composition, Specifications and Conformity Criteria for Common Cements”. British Standards Institution, London.

BS EN 1015-2 (1999). “Methods of Test for Mortar for Masonry: Bulk Sampling of Mortars and Preparation of Test Mortars”. British Standards Institution, London.

BS EN 12350-5 (2000). “Testing of Fresh Concrete- Flow Table Test.” British Standards Institution, London.

BS EN 12390-3 (2002). “Testing of Hardened Concrete- Compressive Strength of Test Specimens.” British Standards Institution, London.

BS EN 12620 (2002). “Aggregates for Concrete.” British Standards Institution, London. Brunori, C., Cremisini, C., Massanisso, P., Pinto, V. and Torricelli, L. (2005). “Reuse of Treated Red

Mud Bauxite Waste: Studies on Environmental Compatibility.” Journal of Hazardous of Materials, B117, pp. 55-63.

Das, S.N., Thakur, R.S. and Ray, H.S. (1988). “Red Mud Pollution Problems: Some Observations.” Environmental and Waste Management, pp. 11-16.

Genc-Fuhrman, H., Tjell, J.C., McConchie, D. and Schuiling, O. (2003). “Adsorption of Arsenic from Water Using Neutralised Red Mud.” Journal of Colloid and Interface Science, 264, pp. 327-334.

Genc-Fuhrman, H., Tjell, J.C. and McConchie, D. (2004). “Increasing the Arsenate Adsorption Capacity of Neutralised Red Mud (BauxsolTM).” Journal of Colloid and Interface Science, 2714, pp. 313-320.

Genc-Fuhrman, H., Tjell, J.C. and McConchie, D. (2004a). “Adsorption of Arsenic from Water Using Activated Neutralised Red Mud.” Environmental Science and Technology, 38, pp. 2428-2434.

Komnitsas, K., Bartzas, G. and Paspaliaris, I. (2004). “Efficiency of Limestone and Red Mud Barriers: Laboratory Column Studies.” Mineral Engineering, 17, pp. 183-194.

McConchie, D. Clark, M., and Davies-McConchie F. (2001). “Processes and compositions for water treatment” International Patent No. PCT/AU01/01383 filed 26 October 2001, entitled Assigned by Deed of Agreement to Nauveau Technology Investments Ltd. on 26th October 2001.

McConchie, D. Clark, M., Davies-McConchie F. and Ryffel T. (2002). “Processes for the treatment of a waste material having a high pH and/or alkalinity.” International Patent No. PCT/AU03/00865

21 CMT - 173


CMT -

filed on 2nd July 2002, Assigned by Deed of Agreement to Nauveau Technology Investments Ltd. on 2nd July 2002.

Mehta, P.K. and Monteiro, P.J.M. (1993). “Concrete: Structure, Properties and Materials.” Second Edition, Prentice Hall Inc, Englewood Cliffs, New Jersey, ISBN: 0-13-175621-4, pp. 26.

Newton, P. (1997). “Making Bricks with Red Mud in Jamaica.” International Development Research Centre (IDRC): Reports: V. 21, No. 2.

Ostap, S. (1984). “Effect of Bauxite Mineralogy on its Processing Characteristics.” Proceedings of the Bauxite Symposium, Los Angeles, CA, pp. 651-671.

Sagoe-Crestsil, K., and Brown, T. (2005). “Bayer process waste stream as potential feedstock material for geopolymer binder systems.” 7th Alumina Quality Workshop, Perth, pp. 214-217.

Sglavo, V.M., Campostrini, R., Maurina, S., Carturan, G., Monagheddu, M., Budroni, G. and Cocco, G. (2000). “Bauxite Red Mud in Ceramic Industry, Part 2: Production of Clay-Based Ceramics.” Journal of European Ceramic Society, 20, pp. 245-252.

Valcin, N. and Sevinc, V. (2000). “Utilisation of Bauxite Waste in Ceramic Glazes.” Ceramics International, 26, pp. 485-793.

Wagh, A.S. and Douse, V.E. (1991). “Silicate Bonded Unsintered Ceramics of Bayer Process Waste.” Journal of Material Research, V. 6, No. 5, pp. 1094-1102.

Zambo, J. and Solymar, K. (1973). “Prospects of Phase Transformation in the Bayer Process.” Proceedings of the 3rd International Congress on Study of Bauxites, Alumina and Aluminum, Nice, pp. 491-502.

21 CMT - 174


CMT -

FEASIBILITY STUDY ON APPLICATION OF LOW-ACTIVATION CONCRETE TO BIOLOGICAL SHIELDING WALL IN LIGHT WATER REACTOR

Yusuke FUJIKURA1, Hirokazu NISHIDA1, Norichika KATAYOSE1, Ken-ichi KIMURA1,

Masaharu KINNO1, Mikio UEMATSU2, Katsumi HAYASHI3, Mikihiro NAKATA4 Takao TANOSAKI5 and Akira HASEGAWA6

ABSTRACT : In order to reduce the residual radionuclides in a concrete shield around a reactor, we have identified several low-activation raw materials by performing a screening test of neutron irradiation. For a biological shielding wall, we proposed two types of low-activation concrete, namely, “1/10-low-activation concrete” and “1/30-low-activation concrete”. In this paper, we present the result of the feasibility study of the low-activation concrete and the thermal stress analyses. The 1/10-low-activation concrete which is composed of the low-heated Portland cement and high-purity limestone aggregates can be utilized for a structure member of a biological shielding wall. The 1/30-low-activation concrete which is composed of the white cement and the high-purity limestone aggregates is also applicable to that. 1/30-low-activation concrete, however, requires further improvement to apply to the massive concrete. The result of this study indicates a good prospect of the practical application of the low-activation concrete to a biological shielding wall.

KEYWORDS: Low-activation concrete, Nuclear power plant, Residual radionuclide, Decommissioning, Massive concrete, Low heat Portland cement, White cement, Limestone aggregate

1. INTRODUCTION Concrete is very practical and inexpensive material for radiation shielding. While, after a long period of operation, the shielding concrete around a nuclear reactor change to low level radioactive waste because of the remaining long lived radionuclides. The disposal cost of this activated concrete is considered to be a hundred times expensive compared to that of non-activated concrete. This calls for the reduction of radioactive concrete to “clearance level” from a viewpoint of cost savings and reutilization of natural resources. Here, “clearance level” denotes the radioactive classification permissible for disposing of material as non-radioactive waste. Some efforts on low-activation concrete in a small proton accelerator of a radiopharmaceuticals factory have been reported (Kinno, 2004). However, no systematic study on the development of low-activation concrete regarding the biological shielding wall in a light water reactor have been studied up to this time. The present work aims to preliminarily study the application of low-activation concrete for a biological shielding wall in a light water reactor.

2. 1/10- AND 1/30-LOW-ACTIVATION CONCRETE 2.1 Low-activation Materials and Mix Proportions Screening tests using the thermal reactor JRR-4 of the Japan Atomic Energy Agency were performed

1 Fujita Corporation Technology Development Division 2 Toshiba Corporation 3 Hitachi, Ltd. 4 Mitsubishi Heavy Industries, Ltd. 5 Taiheiyo Cement Co. 6 Tohoku University

22 CMT - 175


CMT -

to identify low-activation raw materials for composing concrete that can be used for nuclear facilities. About 300 specimens of aggregate and cement were chosen. The dominant long-lived residual radionuclides induced in concrete are, in order of importance, 152Eu (representative single value of the clearance level of IAEA-RS-G1.7 (IAEA, 2004), C152Eu = 0.1 Bq/g, half-life (T1/2) = 13.54 yr), 60Co ( C60Co = 0.1 Bq/g, T1/2 = 5.271 yr), and 154Eu ( C154Eu = 0.1 Bq/g, T1/2 = 8.593 yr), produced by 151Eu (n, γ), 60Co (n, γ), and 153Eu (n, γ) reactions, respectively. These radionuclides are known to occupy 99 – 100 % of the total residual radioactivity induced in ordinary concrete at the time of decommissioning (Kinno, 2002). The dominant target element of concrete shield is, therefore, Eu and Co. Table 1 presents the dominant target element of reference materials obtained by such screening tests. Here, we composed two types of concrete as shown in Table 2, namely, “1/10-low-activation concrete” and “1/30-low-activation concrete”. The reduction rate of the activation for the 1/10-low-activation concrete is designed to be 1/10 compared to the andesite concrete which is considered to be “ordinary concrete”. That for the 1/30-low-activation concrete is designed to be 1/30. The densities and the water absorption rates are presented in Table 3. The compressive strength, slump, and air content are designed as 33 N/mm2, 15±2.5 cm, and 4±1 %, respectively.

Table 1. Dominant target element of reference material (Hasegawa, 2006)

No. Material Place Co (ppb)

Eu (ppb)

1 Andesite (Geostandard sample, JA-1) Japan 12,300 1,200 2 Limestone aggregate A Fukushima, Japan 20 20 3 Limestone aggregate B Okayama, Japan 70 52 4 Limestone aggregate C Fukushima, Japan 8.4 9.3 5 Limestone aggregate D Aomori, Japan 41 7.0 6 Limestone aggregate E Saitama, Japan 44 15.7 7 Ordinary Portland cement A Japan 9,000 690 8 Low heat Portland cement A Japan 10,000 290 9 White cement A Japan 1,590 280

Table 2. Mix proportions (Fujikura, 2006)

Composition (kg/m3)

Type of low-activation concrete W/C(%)

s/a (%) Water Cement Fine

aggregate Coarse

aggregate 1/10-low-activation concrete

(Limestone aggregate + Low heat Portland cement)

50 49 175 350 899 913

1/30-low-activation concrete (Limestone aggregate + White Cement) 50 46 158 316 855 1,018

Table 3. Density and water absorption rate

Material Density (g/cm3)

Water absorption rate (%)

Low heat Portland cement 3.13 － Cement White cement 3.05 －

Fine aggregate 2.67 0.99

Coarse aggregate Limestone (Aomori, Japan)

2.70 0.21

22 CMT - 176


CMT -

2.2 Test and inspection The items of the test and the criteria used in this study are listed in Table 4. Here, “JIS”, “JASS”, and “JCI” denote “Japanese Industrial Standard”, “Japanese Architectural Standard Specification”, and “Japanese Concrete Institute”, respectively. The compressive and the splitting tensile strength of the executed concrete were measured using the cylindrical test specimen of 100 mm in diameter and 200 mm in height. The test specimens were cured in water at 20 ºC or in seal at 20 ºC, for the periods of 7 days, 28 days, and 91 days. The setting time for each execution was automatically measured according to JIS A 1147. The autogenous and drying shrinkage were measured according to JCI-SAS2 and JIS A 1129-1. The bleeding water related to the physical properties of the hardened concrete was also measured according to JIS A 1123. The adiabatic temperature rise related to the crack generation in massive concrete was measured according to JCI SQA 3 by using the metallic container of trapezoid conic, 410 mm in upper diameter, 300 mm in lower diameter, and 420mm in height after the execution work for each concrete.

Table 4. Test and inspection

Item Criteria Remarks Slump JIS A 1101 -

Air content JIS A 1128 - Concrete temperature - Used thermometer Compressive strength JIS A 1108 Standard curing, sealed curing

Splitting tensile strength JIS A 1113 Sealed curing Young’s modulus JIS A 1149 Compressive meter

Density JIS A 1116 Used air meter Dry density JASS 5N T-601 100 mm diameter × 200 mm lengthSetting time JIS A 1147 Automatic machine

Autogenous shrinkage JCI-SAS2 - Length change

Drying shrinkage JIS A 1129-1 - Bleeding water JIS A 1123 -

Adiabatic temperature rise JCI SQA3 Air circulation type

2.3 Physical and mechanical properties The physical properties of the fresh concrete and the mechanical properties of the hardened concrete are expected to be excellent as indicated in Table 5 and Table 6. The 1/10-low-activation concrete, however, requires less bleeding rate in the case of applying to a nuclear facility. The data of the drying and autogenous shrinkages for the 1/10- and 1/30-low-activation concrete are also considered to be excellent as presented in Figure 1. These types of low-activation concrete are, therefore, expected to apply as a structural member of nuclear facilities.

Table 5. Physical properties of fresh concrete

Material Slump (cm)

Air content (%)

Concrete temperature (ºC) Density (kg/m3)

1/10-low-activation concrete 12.5 4.6 20.7 2,340 1/30-low-activation concrete 13.5 3.1 21.1 2,384

22 CMT - 177


CMT -

Table 6. Mechanical properties

Item Condition 1/10-low-activation concrete

1/30-low-activation concrete

7 days 17.9 35.0 28 days 40.2 47.8 91 days

Standard curing (20ºC)

52.2 50.4 7 days 23.6 34.4

28 days 45.5 42.7

Compressive strength (N/mm2)

91 days

Sealed curing (20ºC)

52.8 44.1 7 days 23,740 27,820

28 days 27,540 31,470 Young’s modulus (N/mm2)

91 days

Standard curing (20ºC)

34,820 37,960 7 days 2.27 3.18 28days 3.92 3.56 Splitting tensile strength

(N/mm2) 91 days

Sealed curing (20ºC)

4.22 4.02 Bleeding rate (cm3/cm2) - 0.502 0.091

Percentage of bleeding capacity (%) - 113.5 22.58 Initial

(hr-min) (20ºC) 7-08 4-31 Setting time Final

(hr-min) (20ºC) 9-46 6-11

1/10-low-activation concrete 1/30-low-activation concrete

-600

-400

-200

0

200

0 1 2 3 4 5 6Age (months)

Stra

in (×

10-6

)

Drying shrinkageAutogenous shrinkage

-600

-400

-200

0

200

0 1 2 3 4 5 6Age (months)

Stra

in (×

10-6

)

Drying shrinkageAutogenous shrinkage

Figure 1. Drying and autogenous shrinkage

2.4 Adiabatic temperature rise The adiabatic temperature rises of the 1/10- and 1/30-low-activation concrete are shown in Figure 2, together with the adiabatic temperature rise of the standard concrete proposed by Japan Society of Civil Engineers (JSCE, 2002). The adiabatic temperature rise, Q (t), is derived from estimated the ultimate temperature rise and the coefficient concerning temperature rise speed by the experiment of the adiabatic temperature rise. Q(t)＝Qmax [1-exp(-α×tβ)]

where: Q(t): Adiabatic temperature rise (ºC)

22 CMT - 178


CMT -

Qmax: Ultimate temperature rise (ºC) t: Age (days) α, β : Coefficient concerning temperature rise speed

The values of the ultimate temperature rise for the 1/10- and 1/30-low-activation concrete are small compared to that of the standard concrete.

0

10

20

30

40

50

60

0 2 4 6 8 10 12 14Age (days)

Adi

abat

ic te

mpe

ratu

re r

ise

(ºC

)

Q(t)=Qmax[1-exp(-α×tβ)]　1/10 : Qmax=37.5, α=0.67, β=0.6　1/30 : Qmax=41.8, α=1.55, β=1.0　OC : Qmax=46.0, α=1.104, β=1.0



Ordinaly concrete(C:300kg/m3)

Figure 2. Adiabatic temperature rise

2.5 Residual radioactivity The residual radio activities in the concrete, in ∑Di/Ci unit, are shown in Table 7. The value of ∑Di/Ci was calculated under the condition of 40 yr operation and 6 yr cooling, and at the inner part of a biological shield in a 1,100 MW BWR. Where, Di and Ci indicates the concentration of radionuclide i and the IAEA clearance level of radionuclide i, respectively. The table shows that the ∑Di/Ci rate for the 1/10 low-activation concrete to the ordinary concrete is about 1/10, and that for the1/30 low-activation concrete to the ordinary concrete is about 1/30.

Table 7. Residual radioactivity in ∑Di/Ci unit

Material 60Co 152Eu 154Eu ∑Di/Ci 1/10-low-activation concrete 0.43 1.53 0.082 2.09 1/30-low-activation concrete 0.075 0.50 0.027 0.613

Ordinary concrete (Andesite aggregate + Ordinary Portland cement) 6.52 11.5 0.618 18.8

3. THERMAL STRESS ANALYSES The thermal stress analyses were carried out under the assumption of the application for the biological shielding wall of the BWR. The thermal stress analyses were executed by a three-dimension finite element method, with the cylindrical model of 45 degrees cited by Architectural Institute of Japan (AIJ, 2001) by Q(t) from the results of the adiabatic temperature rise shown in subsection 2.4. The analytical model of the concrete wall and the boundary condition of the heat transfer are shown in Figure 3. The initial temperature of the concrete, the outside atmospheric temperature and the ground initial temperature were assumed to be 20 ºC. The placing of concrete was assumed to be divided into three portions, i.e. the base concrete, the 1st lift of the biological shielding wall, and the 2nd lift of portions in the order of the execution works.

22 CMT - 179


CMT -

The placing of the 1st lift concrete of the wall was executed after the base concrete execution for 28 days later. The placing of the 2nd lift concrete was also executed another 28 days later after the execution of 1st lift concrete. The executed concrete was assumed to be faced to the form for first seven days and be faced to the atmosphere after 7 days curing. The analysis was executed for 147 days, from the time of the placing of base concrete to the 91 days after the placing of end of the 2nd lift concrete. Figure 4 presents the assumed interval of the placing of concrete, the periods of analysis and curing. Thermo data used in thermal stress analysis is shown in Table 8. The crack resistances of the massive concrete for the node A and B, which located at the center of the section in Figure 3, were evaluated.

a) Analytical model (AIJ, 2001) b) Boundary condition of heat transfer

Figure 3. Analytical model of biological shielding wall and boundary condition

0 28days 56days 91days 147days

Start of analysis End of analysis

Placing of concrete Base

In form (7days) Remove the form(after 7 days) Placing of concrete

1st lift In form (7days) Remove the form(after 7 days)

Placing of concrete 2nd lift

In form (7days) Remove the form(after 7 days)

Figure 4. Interval of placing, analytical period and curing condition

Table 8. Thermo data used in thermal stress analysis

Heat transfer coefficient (W/m2･ºC)

Material Specific heat (kJ/kg･ºC)

Thermal conductivity (W/m･ºC)

Density (kg/m3) Form Concrete surface

Concrete 1.15 2.7 2,350 8.0 12.0 Ground 0.800 1.7 2,600 －－

* Coefficient of thermal expansion α=0.00001

1200mm

2500mm

5300mm

11000 mm

27600 mm

Base concrete

2nd lift

1st lift

Biological shielding wall Form or concrete surface

Adiabatic boundary A

B

Concrete surface

2800mm

22000 mm

22 CMT - 180


CMT -

Table 9. Maximum temperature and minimum crack index (1200 mm in thickness)

1st lift (Node A) 2nd lift (Node B) Maximum concrete temperature CT

Minimum crack index ic*

Maximum concrete temperature CT

Minimum crack index ic* Material

CT (ºC)

Age (days) ic Age

(days)CT (ºC)

Age (days) ic Age

(days)1/10-low-activation

concrete 38.5 1.6 1.67 13.0 38.5 1.6 1.76 15.0

1/30-low-activation concrete 52.8 1.4 0.84 11.0 52.8 1.4 1.06 13.0

*Crack index ic=ftk(t) /σt(t), ftk(t): tensile strength, σt(t): thermal stress, t:age

The results of the thermal stress analyses are summarized in Table 9. The table presents that the maximum temperature of the concrete and the minimum crack index of A and B. The ratio of the tensile strength to the thermal stress is defined as “crack index” (JSCE, 2002). The value of the minimum crack index to prevent the crack generation is recommended to be 1.45-1.75. The probability of the crack generation is estimated to be 5% for the crack index of 1.75, to be 25% for the crack index of 1.45, and also to be 85% for the crack index of 1.0. The minimum crack index of the 1/10-low-activation concrete is 1.76, and that for the 1/30-low-activation concrete is 1.07. This calls for further improvement regarding on the property of the massive concrete for the 1/30-low-activation concrete. The 1/30-low-activation concrete is, however, applicable for a thin wall and a pre-cast panel. The results of the analyses conduct to the necessity to reduce the value of coefficient concerning temperature rise speed α for the application of the 1/30-low-activation concrete to the massive concrete structure. In order to improve the crack index of the 1/30-low-activation concrete, another work has been performed for the low activation concrete with the low activation admixture obtained by the screening test (Kimura, 2006).

4. CONCLUSIONS In order to reduce the residual radionuclides in a concrete shield around a reactor, we have identified several low-activation raw materials by performing a screening test of neutron irradiation. For a biological shielding wall, we proposed two types of low-activation concrete, namely, “1/10-low-activation concrete” and “1/30-low-activation concrete”. The reduction rate of the residual radioactivity, in ∑D/C unit, for the 1/10-low-activation concrete is designed to be 1/10 compared to the andesite concrete which is considered to be “ordinary concrete”. That for the 1/30-low-activation concrete is designed to be 1/30. The results of this feasibility study are summarized as follows, 1) The 1/10-low-activation concrete which is composed of the low-heated Portland cement and high-

purity limestone aggregates can be utilized for a structure member of a biological shielding wall. 2) The 1/30-low-activation concrete which is composed of the white cement and high-purity

limestone aggregates is also applicable to that. This concrete, however, requires further improvement on the high-heat problems to apply to massive concrete of the biological shielding wall.

5. REFERENCES AIJ (2001). “State of The Art Report on Mass Concrete. ”, Architectural Institute of Japan. Fujikura, Y. et al. (2006). “Feasibility Study on Low-Activation Concrete as Massive Concrete”,

Preprints 2006 Annual Mtg., Japan Society of Civil Engineers, September, [preparing]. Hasegawa, A. et al. (2006). “Development of Low-Activation Design Method for Reduction of

Radioactive Waste below Clearance Level”, Innovative and Viable Nuclear Energy Technology (IVNET) Development Project, Ministry of Economy, Trade and Industry, Japan.

22 CMT - 181


CMT -

IAEA (2004) “Application of the concepts of exclusion, exemption and clearance”, Safety standards series No. RS-G-1.7, International Atomic Energy Agency.

JSCE (2002). “STANDARD SPECIFICATION FOR CONCRETE STRUCTURES-2002, Materials and Construction. ”, Japan Society of Civil Engineers.

Kimura, K. et al. (2006). “Development of Low-Activation Design Method for Reduction of Radioactive Waste below Clearance Level (12) –Low-Activation Admixture–.” Preprints 2006 Annual Mtg., Atomic Energy Society of Japan.

Kinno, M. et al. (2002). “Raw Materials for Low-Activation Concrete Neutron Shield”, Journal of Nuclear Science and Technology, Vol.39, No.12, pp.1275-1280.

Kinno, M. (2004). “The Present Activities on Low-activation.” Japan Concrete Institute, Concrete Journal, Vol.42, No.6, pp.3-10.

ACKNOWLEDGMENT This work is supported by a grant-in-aid of Innovative and Viable Nuclear Technology (IVNET) development project of Ministry of Economy, Trade and Industry, Japan.

22 CMT - 182


CMT -

DEVELOPMENT OF LOW ACTIVATION CONCRETE FOR NUCLEAR PLANT -PROPOSAL OF VARIOUS TYPES OF LOW ACTIVATION CONCRETE-

Ken-ichi Kimura1, Hirokazu NISHIDA1, Yusuke FUJIKURA1, Norichika KATAYOSE1, Masaharu KINNO1, Akira HASEGAWA2, Mikio UEMATSU3, Katsumi HAYASHI4,and Mikihiro NAKATA5

ABSTRACT : Concrete is very valuable and inexpensive material, however it can be changed to be expensive and hard to deal with in use of a nuclear plant after long operation. One of the counter plans for the above is to use low activation concrete instead of ordinary concrete, which can be below clearance level in decommissioning even after the operation. In this paper, firstly fifty raw materials for the concrete were evaluated by radiochemical analyses, which conducted the quantities of dominant trace elements for the activation in certain condition. As results of above investigation, three kinds of aggregates (fused alumina ceramics, silica sand and limestone) and two kinds of cements (high alumina cement and white cement) were selected as low activation raw materials. Secondly six types of low activation concrete including two mortal based on the investigation were proposed and described their mix proportion design with target characteristic, considering the application of the proposed concrete to the certain portion member in nuclear plants, with assignments for further improvements.

KEYWORDS: Low activation concrete, limestone aggregate, silica sand, quartz aggregate, fused silica aggregate, high alumina cement, white cement, clearance level.

1. INTRODUCTION Concrete enveloping a nuclear reactor remains residual radioactivity after decommissioning. The disposal of such concrete is very costly and requires strict supervision. From this point of view, we have developed new concrete that retains little residual radioactivity, that is, “low-activation” concrete (Kinno, 2002a). One national funded project of comprehensive development for the low activation concrete has just started recently (Hasegawa, 2006). The goal of this comprehensive project is to reduce radioactive concrete to be a half and is especially to design the whole concrete structure below clearance level on decommissioning. That will contribute much toward solving safety and economical problems of the nuclear power plant. The indispensable items of the development could be as follows; investigation of residual radionuclides in construction materials around a reactor, improvement and development of low-activation materials, and establishment of low-activation design method for reduction of radioactive waste below clearance level. To achieve above comprehensive project, a feasible study (Kinno, 2005) was performed for two basic researches as follows,

Proportioning of low-activation raw materials and improvement for low-activation cement We proposed six kinds of mixing design of low-activation raw materials to make low-activation concrete, which reduction ratios to the andesite concrete are from 1/300 to 1/10 in ∑Di/Ci unit, where Di is the concentration of the radionuclide i and Ci is the clearance level of the radionuclide i. Three kinds of improvement plans for low activation cement were also proposed.

1Fujita Corporation Technology Development Division 2 Tohoku University 3 Toshiba Corporation 4 Hitachi, Ltd. 5 Mitsubishi Heavy Industries, Ltd.

23 CMT - 183


CMT -

Investigation of method for development of low-activation concrete The methods for development of low-activation concrete and mortal which reduction ratio to the andesite concrete were from 1/300 to 1/10, were investigated to apply to the reactor shielding wall of the boiling water reactor (BWR) and to the biological shielding wall of the pressurized water reactor(PWR). The results indicated that the applications of low-activation concrete were economically reasonable taking into consideration of construction cost, and reduced the quantity of the radioactive waste steeply and the disposal cost drastically.

Based on the results of the above feasible study, we describe outline of the results related concrete and raw materials for the concrete in this paper. Thus, firstly fundamental technical background is introduced as background of the project, and secondly the investigation of the potential low activation materials is addressed. Finally, six types of low activation concrete are proposed and these mix proportion design are described with target characteristic and assignments, considering the application of the proposed concrete to the certain portion of the member in nuclear plants. 2. CRAITERIA OF CLIARANCE LEVEL FOR RADIOACTIVE CONCRETE WASTE As well known, concrete is very valuable and inexpensive material, and is used in variety of the application such as, building, infrastructure, and so on. Concrete is also used in the nuclear plant as a structural member and shielding wall against radiation and radioactivity. Certain radiation, typically neutron, may make concrete radioactive under certain condition, and this phenomena is called “activation”, and this concrete can be called “activated concrete” (Price, 1958). In this case, the concrete can be changed to be expensive and hard to deal with, and the activated concrete may be required strict supervision as a radioactive waste at the end of the operation. The increase of the expense for the radioactive waste can include the cost for demolition (decommission), treatment of the radioactive waste, setting of the waste under the ground and maintenance. The cost for the setting of the radioactive concrete waste was estimated 65 times to 2800 times expensive than construction cost for ordinary concrete, which were calculated by the estimated three leveled radioactive waste and total estimated cost for two type of reactors in the intermediate report by Ministry of International Trade and Industry (MITI, 1999). Most of the concrete used in the nuclear plant, however, is not such radioactive waste, which is required strict supervision. Table 1 gives an idea of the quantity of the waste for the typical nuclear plant (MITI, 1999). This table shows that most of the waste from nuclear plants is concrete and most of the concrete waste is not radioactive waste. Nevertheless, concrete radioactive waste is existed still 4000 ton for the BWR (1.1GW level). On the other hand, 7000 ton is categorized as the waste which is not necessary to treat as the radioactive waste, in another word “below clearance level” in this case. So, there is no concrete radioactive waste, when all concrete used in nuclear plants are made by enough low activate concrete below clearance level even after the operation.

BWR (1.1GW level) PWR (1.1GW level) GCR (1.1GW level) Criteria for waste metal concrete total metal concrete total metal concrete totalLow level

radioactive waste 0.9 0.4 1.3 0.4 0.2 0.6 0.3 1.8 2.2

Not necessary to treat as radioactive waste 2.1 0.7 2.8 0.3 0.8 1.2 0.6 3.6 4.2

Not radioactive waste 0.8 48.7 49.5 3.4 44.3 47.7 1.0 11.9 12.9

total 3.8 49.8 53.6 4.1 45.4 49.5 1.9 17.3 19.2

Table 1. Estimated waste of the typical nuclear plants after operation

[Unit: 107 kg]

23 CMT - 184


CMT -

Figure 1. Configuration of the JRR-4 core

Clearance level has been discussed to determine the line between regular waste and low level radioactive waste (LLW) reasonably. International Atomic Energy Agency (IAEA) has provided international guideline for the LLW, specifying “basic safety standard” for LLW in the report “TECDOC-855” (IAEA, 1996) in 1996, based on the principle of basic safety for the protection against the ionizing radiation and for the safety of radiation sources(UN,1996). TECDOC-855 defines criteria to define trivial radiation level which poses negligible risk based on increased cancer risk of one per million per year, which translates to 100 µSv/y, comparing to 1000 µSv/y for individual radiation allowance recommended by International Committee for Radiation Protection. So, TECDOC-855 recommended 10 µSv/y as clearance level criteria. Japan Nuclear Safety Commission (JNSC) reported clearance level from exposure estimation performed using food-chain scenarios localized Japanese lifestyle (STA, 1999), based on above IAEA 10 µSv/y criteria (describing “CL-Japan1999”). On the other hand, after above first report from IAEA, they reevaluated TECDOC-855 and issued RS-G-1.7 for clearance level (IAEA, 2004, describing “CL-IAEA2004”). JNSC also reevaluated exposure estimation using RS-G-1.7 and regulated in 2004. 3. INVESTIGATION FOR LOW ACTIVATE MATERIALS Based on previous works (Kinno, 2002b), (Kimura, 1994), (Kinno, 2000), 45 raw materials for concrete were selected for candidates of low activation materials for low activation concrete. Aggregates (30 kinds of limestone, 6 kinds of fused alumina aggregates, and 4 kinds of quartz aggregates including silica sand) and cements (a white cement and 5 kinds of high alumina cements) were investigated for the candidates. Geo standard samples (JR-1, JA-1 and JB-1), standard ordinary concrete, and ordinary Portland cement were also selected as comparable samples for above materials. Standard ordinary concrete was executed by Andesite aggregates, which was representative aggregate for the concrete in Japan, and ordinary Portland cement. Low activation concrete has been developing to achieve the concrete which activity is below clearance level in use of most portions in the nuclear plants. Low activation, however, can have several meanings depending on purpose, term, materials, and so on. So, this paper focuses on the concrete and its raw materials in the radioactive waste for nuclear plants after long operation. Therefore typical conditions for the discussion on radioactive concrete waste in this paper are 40 years operation and 6 years cooling (maintaining after stop of the operation for 6 years). Based on above conditions, radioactive nuclides in the concrete should be limited (typically 60Co, 134Cs and 152Eu), and therefore trace elements for the investigation were selected Co, Cs, Eu, Fe and Sc, which is represent of rare earth elements. 3.1 Radiochemical analysis The trace elements dominated to the activation for concrete materials were evaluated by radiochemical analyses (Price, 1958), as follows, 1. Collecting certain concrete materials 2. Crashing materials to certain size

(typically under 1mm) 3. Packing above crashed samples for 0.1 to

1g with special treatment for irradiation 4. Irradiation by thermal neutron in the

reactor core of JRR-4 (shown in Figure 1) Thermal neutron flux: 5.3E13 cm-2sec-1

Irradiation time: 20 minutes 5. Cooling for 66 to 87 days 6. Measurement of gamma spectrum for irradiated samples by Ge detector 7. Evaluation of quantity of the trace element for each sample.

23 CMT - 185


CMT -

3.2 Results for aggregates

From above radiochemical analyses, quantities of the Co, Cs, Eu, Fe and Sc for 50 samples were evaluated. Figure 2 shows the distribution of the evaluated quantities of Eu and Co in typical aggregate samples (limestone aggregates, fused alumina aggregates and quartz aggregates) with those of other aggregates from previous works (Kinno, 2002). This figure indicates above selected samples for this work have little quantities of Eu and Co comparing to those in ordinary aggregates (the distribution of Eu and Co for ordinary aggregates is located around the center of figure, and enlarged figure is also shown at the upper right in Figure 2). Figure 2 also has index curves of Σ(D/C) for the activation. The Σ(D/C) in this figure is defined by equation (1) as follows, Σ(D/C) = D152Eu/C152Eu + D154Eu/C154Eu + D60Co/C60Co (1)

Di: Concentration of radionuclide of 152Eu, 154Eu and 60Co induced under 5.0E5 n cm-2sec-1 thermal neutron flux yield

Ci: Clearance level referring CL-Japan1999 for this calculation, which are 0.4 for 152Eu, 154Eu and 60Co (STA, 1999).

Evaluated limestone samples in the figure are located withinΣ(D/C)=1.0 line, and fused alumina ceramics are located withinΣ(D/C)=0.1, in comparison with other aggregates againstΣ(D/C)=1.0 curve. Therefore limestone, quartz, and fused alumina were selected low activation materials. Table 2 shows the Σ5D/C ratio of the evaluated some aggregates to the ordinary aggregate, which is average of the Geo-standard samples JR-1 (Rhyolite), JB-1 (Basalt) and JA-1 (Andesite). Σ5D/C is defined by equation (2), as follows, Σ5D/C = D55Fe/C55Fe + D60Co/C60Co + D134Cs/C134Cs + D152Eu/C152Eu + D154Eu/C154Eu (2)

0.001

0.01

0.1

1

10

100

1000

10000

0.0001 0.001 0.01 0.1 1 10 100 1000

Eu (ppm)

Co

(ppm

)

Dunite(1)

Quartzite(5)

Alumina-ceramics(3)

Limestone(12)

Serpentite(9)

Magnetite(2)

Ferro-manganese(3)

Apatite(1)Syenite(1)

Bauxite(1)

Phosphate(1)Mullite(1)Baryte(2)

Colemanite(1)

Σ(D/C) = 1.0

Aggregate

Σ(D/C) = 0.11,000

1,000

10,000

This work,Alumina-ceramics(3)

This work, Silica sand

This work, Limestone Rhyolite(2)

Trachyte(4)

Silica stone(5)

Sandstone(38)Granite(5)

Sand(20)

Basalt(14)

Diorite(9)Porphyrite(5)Gravel(35)

Porphyry(4)

Crushed stone(120)

Andesite(34)

Figure 2. Distribution of quantities for Eu and Co in aggregates with enlargement of ordinary aggregates (upper right)

Ordinary aggregates

Enlargement of the distribution for Ordinary aggregates

23 CMT - 186


CMT -

0.001

0.01

0.1

1

10

100

0.001 0.01 0.1 1Eu (ppm)

Co

(ppm

)

OOPE401GSR-6KH

JLs-1

Σ(D/C ) = 1.0

CCH

L imestoneΣ(D/C ) = 0.1

○　This work

Figure 3. Distribution of quantities for Eu and Co in limestone aggregates

Table 2. Σ5D/C ratio for evaluated aggregates

*The ratios is to the average of JR-1, JB-1, and JA-1, which is assumed the average aggregate.

Di: Concentration of radionuclide of 55Fe, 60Co, 134Cs, 152Eu, and 154Eu induced under 5.0E5 n cm-2sec-1 thermal neutron flux yield

Ci: Clearance level referring CL-Japan1999 for this calculation, which are 0.4 for 152Eu, 154Eu, 60Co, 0.5 for 134Cs, and 3000 for 55Fe (STA, 1999).

Figure 2 and Table 2 indicate that selected aggregated in this work are very useful for the low activation materials (the Σ(D/C) ratio of fused alumina, quartz sand and limestone to the ordinary aggregates are about 1/1000, 1/200, and 1/20, respectively). Figure 3 also shows the distribution of the quantities of Eu and Co for selected limestone aggregates with others by previous works. This figure indicates that the most of the limestone including the ones of this works, are within the curve of Σ(D/C) = 1.0, but a few of them are outside of the curve. So this result conducts that all limestone is not necessary low activated material, therefore it is necessary to

Aggregate Σ5D/C ratio* Fused alumina aggregate CA 0.00068 Fused alumina aggregate JA 0.0026 Fused alumina aggregate JB 0.0021 Fused alumina aggregate CF 0.0014 Fused alumina aggregate EA 0.00050 Quartz sand JT 0.0057 Quartz sand JA 0.65 Quartz sand AF 0.0049 Quartz aggregate IA 0.00076 Limestone aggregate FO 0.014 Limestone aggregate OK 0.038 Limestone aggregate HK 0.020 Limestone aggregate FA 0.0067 Limestone aggregate FK 0.023 Limestone aggregate TH 0.0085 Limestone aggregate AH 0.0054 Limestone aggregate SB 0.011 Limestone aggregate KT 0.011

Figure 4. Distribution of quantities for Eu and Co in cement

0.001

0.01

0.1

1

10

100

1000

10000

0.0001 0.001 0.01 0.1 1 10 100 1000

Eu (ppm)

Co

(ppm

)

High-alum ina cement(2)

W hite Portland cem ent(3)

Moderate-heat Portland cem ent(3)O rdinary Portland cem ent(13)

Fly -ash Type B cement(3)Low -alkali cem ent(2)

Blast-fu rnace Type B cement(4)

Low-alum ina cem ent(2)

∑(D/C ) = 1.0

Cem ent

∑(D /C ) = 0.11,000

10,000

1,000

Fly-ash(5)Gypsum (5)

Silica fum e(5)

This w ork, High-aluminacement(4)

This work, W hite cem ent

23 CMT - 187


CMT -

Table 3. Σ5D/C ratio for evaluated cements

Table 4. Supplement condition for low activation materials

*Di: Concentration of radionuclide i, Ci: Clearance level of radionuclide i, cited from IAEA-RS-G1.7, assuming inner part of the biological shield in a 1,100 MW BWR, 40 year of operation, and 6 years of cooling. **Average aggregate is assumed 1.0 for aggregates, and ordinary Portland cement is assumed 1.0 for cements.

*The ratio is to the ordinary Portland cement, which is assumed the average cement

chose right limestone with certain investigation. 3.3 Results for cements Figure 4 shows the distribution of the quantities for Eu and Co in selected cements for this work with other cements by previous works. Σ(D/C) curves, which can be calculated by equation (1), are also set for the index of the activation in the figure. Table 3 shows the ratio of Σ5D/C, which is calculated by equation (2), for the evaluated some cements to the ordinary Portland cement manufactured by A factory in Japan. Figure 4 and Table 3 indicate that selected aggregates in this work are very useful for the low activation materials (the Σ(D/C) ratio of high alumina cement, and white cement to the ordinary Portland cement are about 1/20 - 1/50 and 1/3, respectively). 4. PROPOSAL OF LOW ACTIVATION CONCRETE The investigation in the section 3 is submitted possible low activation materials for the low activation concrete. Based on the investigation, various types of low activation concrete are described in this section. 4.1 Supplement conditions for the low activation aggregates and cements The investigation is conducted three aggregates and two cements as potential low activation materials for low activation concrete from the points of view of the activation by radiochemical analyses. Obviously, it is very important to investigate other conditions in actual use of the above materials for concrete. Table 4 shows the results of the investigation related to the supplement for the above materials such as relative cost and ability of the supplement. This investigation was performed to the company for each aggregate and cements listed the above investigation in section 3, by inquiries, site visit investigations and actual orders of the certain quantity. The ΣDi/Ci ratios to the assumed average materials (the average of the geo standard samples for aggregates and ordinary Portland cement for cements) are also

Cement Σ5D/C ratio* High alumina cement FR 0.045 High alumina cement EA 0.045 High alumina cement JA 0.0040 High alumina cement JB 0.056 High alumina cement JC 0.021 White cement S 0.35 Low heated Portland cement T 1.0

Material ΣDi/Ci* ratio Relative Cost** Ability for supplements Point characteristic

Fused alumina aggregate 1/400 – 1/1500 20- 40 Good High density and

hardness

Quartzite (Silica) aggregate 1/200 - 1/1200 5 - 10 Good Hardness

Limestone aggregate 1/30 – 1/200 1.5 – 3 Good (partially N.G) Require washing

High alumina cement 1/20 – 1/50 30 Good Long term durability,

Thixotropy

White cement 1/3 3 Good Heat generation

23 CMT - 188


CMT -

Table 5. Mix proportion design for proposed six low activation concrete

* Andesite concrete is assumed 1.0.

shown in the figure. The ΣDi/Ci for each material in this figure is recalculated under the new clearance level (CL-IAEA2004, IAEA, 2004) focusing actual nuclear plants. The point characteristics for each material are also addressed in the table. Relative cost is shown as a relative comparison for the actual expense to the ordinary materials (the expense of average aggregate is used for the comparison of low activation aggregates and the expense of ordinary Portland cement is used for the comparison of low activation cements) taking into account other conditions from the above investigation. This investigation confirmed realization of the use of low activation aggregates and cements listed above for the low activation concrete. 4.2 Mix proportion design Above investigation leads some potential low activation concrete. Table 5 summarizes the mix designs for six types of proposed low activation concrete with target of characteristics. In this table, ΣDi/Ci ratio, which is the ratio of ΣDi/Ci for each designed low activation concrete to that for the Andesite concrete (Andesite aggregate and ordinary Portland cement), is also addressed. In order to apply low activation concrete for the certain portion of the nuclear plants effectively, not only all of the aspects of the examination results of the concrete, such as compressive strength, shrinkages, and fresh concrete properties are not sufficient. Other conditions, such as cost, durability, low activation, and so on, should be taken into account. Therefore above low activation design should be very complex and difficult. In this sub section, six types of proposed concrete including two types of mortal are addressed from the point of view of materials, characteristics, assignments and target application in followings. Concrete A Materials: Fused alumina ceramics for coarse and fine aggregates, high alumina cement, and

admixture for anti-shrinkage Characteristics: Ultra low activation (ΣDi/Ci ratio is expected 1/300), high density, good property for

heat resistance and high thermal conductivity Assignment: Expensive, high drying shrinkage, Long term durability, thixotropy, heat generation

and difficulty for execution work Target application: Reactor shielding wall for BWR, inner portion of shielding wall for PWR, and

partial spot for high neutron yield Concrete B

Target of characteristics Coarse

aggregate Fine aggregate Cement ΣDi/Ci ratio* Density (g/cm3)

A Fused alumina ceramics

Fused alumina ceramics High alumina cement 1/300 – 1/400 3.0

B Quartzite Silica sand High alumina cement 1/150 -1/200 2.4

C Limestone Limestone White cement 1/30-1/50 2.3

D Limestone Limestone Low heated cement 1/10-1/30 2.3

E Silica sand High alumina cement 1/150 2.3

F Silica sand + Limestone powder White cement 1/25 2.1

23 CMT - 189


CMT -

Materials: Quartz (partially fused alumina) for coarse aggregate, silica sand (partially fused alumina) for fine aggregate, high alumina cement, and admixture for anti-shrinkage

Characteristics: Very low activation (ΣDi/Ci ratio is expected 1/150). Assignment: Expensive, high drying shrinkage, Long term durability, thixotropy, heat generation,

difficulty for execution work and not for use in high temperature. Target application: Reactor shielding wall for BWR, inner portion of shielding wall for PWR, and

partial spot for high neutron yield Concrete C Materials: Low activation limestone aggregate for coarse and fine aggregates, and white cement Characteristics: Low activation (ΣDi/Ci ratio is expected 1/30 to 1/50) and could be moderate type as

balance of cost and performance. Assignment: Further improvement of white cement for low activation, and heat generation during

execution work Target application: Biological shielding wall for BWR, and outer portion of shielding wall for

PWR. Concrete D Materials: Low activation limestone aggregate for coarse and fine aggregates, and low heated

cement Characteristics: Relatively low activation (ΣDi/Ci ratio is expected 1/10 to 1/30), and inexpensive. Assignment: Further improvement of low heated cement for low activation Target application: Biological shielding wall for BWR, and outer portion of shielding wall for

PWR Mortal E Materials: Fused alumina ceramics and silica sand for fine aggregate, high alumina cement, and

admixture for anti-shrinkage Characteristics: Very low activation (ΣDi/Ci ratio is expected 1/300) and good property of filling Assignment: High drying shrinkage, Long term durability, thixotropy, heat generation, difficulty

for execution work and not for use in high temperature Target application: Reactor shielding wall for BWR, inner portion of shielding wall for PWR,

partial spot for high neutron yield and infilled mortal Mortal F Materials: Silica sand and limestone powder for fine aggregate, white cement, and admixture for

anti-shrinkage Characteristics: Low activation (ΣDi/Ci ratio is expected 1/25), and good property of filling Assignment: Further improvement of white cement for low activation, and heat generation during

execution work Target application: Biological shielding wall for BWR, outer portion of shielding wall for PWR,

and infilled mortal. 4.3 Remarks for Further improvement Most of the assignments described in the sub section 4.2 for proposed six types of low activation concrete are caused from cements. The assignments for the concrete in use of high alumina cement are typically high drying shrinkage, long term durability, thixotropy, and heat generation. Those for the concrete in use of white cements are heat generation and further improvement for low activation and the assignment for the concrete in use of low heated cement is further improvement of low activation.

23 CMT - 190


CMT -

In order to achieve further low activation cement for white cement and low heated cement, raw materials in manufacture stage should be carefully selected, especially for the contents of Co and Eu in the cement. As mentioned in section 3, Eu and Co are dominant materials for activation under the condition described in this paper. Heat generation during execution work may cause unacceptable crack, so it is also very important for the decrease the heart generation. We have developed to reduce the heat generation for the concrete with white cement by adding low activation admixture. This could be one of the ways to improve heat generation for the concrete with white cements and alumina cements. Long term durability is another problem for use of high alumina cements, and may improve by the mix proportion with low W/C (the ratio of water to cement) under 40% (Scrivener, 1999). 5. CONCLUSION Recent comprehensive project of low activation design method for reduction of radioactive waste below clearance level has started from 2004, as a feasibly study, which has four main objects such as, investigation of major construction materials and low activation materials, determination of representative part and establishment of calculation and evaluation methods, proportioning of low activation raw materials to make low activation concrete, and investigation of method for low activation concrete and mortal. In this paper, the results of the feasibly study related concrete and its raw material were mainly discussed. As fundamental issues, criteria of the clearance level for radioactive concrete waste were addressed as a background of the comprehensive development. 50 raw concrete materials potentially for the low activation concrete were investigated by radiochemical analyses, in order to estimate the quantities of trace elements dominated to the activation in radioactive concrete waste in certain condition. Limestone aggregates, quartz including silica sand, and fused alumina ceramics were selected as low activation aggregates, and high alumina cements and white cement were selected as low activation cements, which were similar results as previous works. Based on above investigations, six types of low activation concrete were proposed, which were Concrete A (fused alumina ceramics aggregates and high alumina cement), Concrete B (quartz aggregates including silica sand and high alumina cement), Concrete C (limestone aggregates and white cement), Concrete D (limestone aggregates and low heated cement), Mortal E (fused alumina ceramics aggregates including silica sand and high alumina cement) and Mortal F (silica sand including limestone powder and white cement). The mix proportion designs were listed with characteristics, assignments, and target application in the nuclear plants for each proposed concrete. These assignments of the proposed concrete also conducted to the necessary points of further improvement for cements, which were discussed in this paper. ACKNOWLEDGMENT This work is supported by a grant-in-aid of Innovative and Viable Nuclear Technology (IVNET) development project of Ministry of Economy, Trade and Industry, Japan. REFERENCES Hasegawa, A. et al. (2006). “Development of Low-Activation Design Method for Reduction of

Radioactive Waste below Clearance Level”, Innovative and Viable Nuclear Energy Technology (IVNET) Development Project, Ministry of Economy, Trade and Industry, Japan (in Japanese).

23 CMT - 191


CMT -

IAEA (1999) “Clearance levels for radionuclides in solid materials”, IAEA-TECDOC-855, International Atomic Energy Agency

IAEA (2004) “Application of the concepts of exclusion, exemption and clearance”, Safety standards series No. RS-G-1.7, International Atomic Energy Agency

Kimura, K., Ishikawa, T., Kinno, M., and Nakamura, T. (1994). “Compilation of neutron activation cross section and trace element content of concrete for induced radioactivities.”, Proceedings of 8th International Conference on Radiation Sheilding, vol. 1, pp.35-42, ANS inc, Arlington, USA..

Kinno, M., Kimura, K. and Nakamura, T. (2000). “Ultra-low-activation limestone for neutron irradiation” ANS Radiation Protection and Shielding Conference, pp.673-678, Spokane, USA.

Kinno, M. (2002a). “The Present Activities on Low-activation.” Japan Concrete Institute, Concrete Journal, Vol.42, No.6, pp.3-10 (in Japanese).

Kinno, M., Kimura, K. and Nakamura, T. (2002b). “Raw Materials for Low-Activation Concrete Neutron Shields,” Journal of Nuclear Science and Technology, Vol. 39, No.12, 1275-1280.

Kinno,M. et al. (2005). “Development of Low-Activation Design Method for Reduction of Radioactive Waste below Clearance Level (1) –General–. ” Atomic Energy Society of Japan. (in Japanese)

MITI(1999), “Intermediate report for the decommissioning of commercial nuclear reactor”, General energy research committee, 1999. 5.18, Ministry of International Trade and Industry (in Japanese)

Price, W.J.(1958), “Nuclear radiation detection.” New York, McGraw-Hill Book Company, Inc. Scrivener, K. L., Cabiron, J., Letourneux, R., (1999) “High-performance concretes from calcium

aluminate cements”, Cement and Concrete Research 29, pp.1215-1223. STA (1999) “Japanese clearance levels for radionuclides in solid materials”, Science and Technology

Agency (in Japanese) UN (1996) “International Basic Safety Standards for Protection against Ionizing Radiation and for the

Safety of Radiation Sources”, FOOD AND AGRICULTURE ORGANIZATION OF THE UNITEDNATIONS, INTERNATIONAL ATOMIC ENERGY AGENCY, INTERNATIONAL LABOUR ORGANISATION, OECD NUCLEAR ENERGY AGENCY, PAN AMERICAN HEALTH ORGANIZATION, WORLD HEALTH ORGANIZATION, Safety Series No. 115, IAEA, Vienna (1996).

23 CMT - 192


CMT -

APPLICABILITY OF ELASTO-VISCO-PLASTIC MODEL FOR TIME-DEPENDENT TENSILE BEHAVIOR OF CONCRETE

Koji Matsumoto1, Yasuhiko Sato2 and Tamon Ueda3

ABSTRACT : The numerical model, which is for time-dependent tensile behavior of concrete, was proposed and verified its applicability in this paper. The model has elastic spring, plastic slider and dashpot element and visco-plastic component, in which dashpot is connected to plastic slider, and elastic component is connected in series. Introducing the viscosity reduction model, viscosity is reduced with crack opening. The model succeeded to numerically express the enhancement of tensile strength and rapid stress release in post-peak range with higher strain rates. However, through the comparison with the CEB model code, it became clear that the effect of strain rate on concrete strength in the model is too oversensitive.

KEYWORDS: Time-dependent, Elasto-visco-plastic model, Mesoscopic

1. INTRODUCTION In order to develop the performance based design scheme, we have to predict deterioration degree of structures and structural materials during their service lives. Hence, many studies on prediction of long term chracteristics of concrete under various deterioration factors have been conducted. As a result, for instance, chloride ion transfer and carbonation proces became predictable. On the other hand, structural performance under fatigue loads, which is a typical mechanical deterioration of concrete structure, is arrested by S-N curve in current design codes. Additionally, it is assumed that damage under fatigue loads is predicted by liner cumulative damage law. However, these methods cannot predict structural state at arbitrary point in time and that under combined action with another deterioration factors. Therefore, current design codes for fatigue loads cannot play important role in the performance based design concept. In order to clarify deterioration mechanism of concrete structures, nonliner analysis such as finite element method can be a powerful tool, but of course, it requires constitutive model which is applicable to fatigue loads. Besides, in order to develop constitutive law, fracture mechanism of concrete material itself have to be clarified. Nagai et al. analytically clarified that macroscopic failure of concrete is a combination of local tensile and shear fractures [1]. Even though it seems to be a crashing in compression, local cracking phenomenon actually governs the failure of concretes. Therefore, modeling of failure and softening under uni-axial tensile loads is important to clarify the mechanism of concrete failure. In this paper, applicability of the elasto-visco-plastic model to time-dependent problem under uni-axial tensile loads is discussed.

2. MODELING

2.1 Basic Concept As is well known, concrete under uni-axial tension is cracked in perpendicular direction to loading axis. Therefore, under uni-axial tensile loads, concrete has two deformational components, one is elongation in cracking zone and the other is that in un-cracked zone which are characterized by aggregate bridging effect and material elasticity, respectively. So then, time effect on each component should be different. Consequently, in this study, deformation of cracked and un-cracked zone is

1 Doctoral student, Hokkaido University, Japan, [email protected] 2 Associate professor, Hokkaido University, Japan, [email protected] 3 Professor, Hokkaido University, Japan, [email protected]

24 CMT - 193


CMT -

separately modeled and total deformation is described as a summation of each component as shown in Fig.1.

2.2 Mechanical Model Figure2 shows the mechanical model in this study. In this model, visco-plastic component, in which dashpot is connected to plastic slider in parallel, is connected to elastic spring in series. Elastic spring, which is a reversible deformation, represents the deformation of un-cracked zone mentioned before. Contrary, visco-plastic component, which is an irreversible deformation, represents the deformation of cracked zone. In this model, dashpot represents time-dependent deformation caused by crack development because it is connected with plastic slider. The authors have already reported that concrete under time-dependent loading has two deformational components: one is caused by cracks and the other is caused by consolidation [2]. And crack induced deformation gives damage while consolidation induced deformation does not affect any damages. In this paper, since dashpot is not connected to elastic spring, consolidation induced deformational components is not taken into account. It is known that consolidation component becomes much smaller in case of higher stress levels.

2.3 Governing Equations (1) Elastic spring Mechanical behavior of the elastic spring is perfect elastic. Stress-strain relationship of elastic spring is given as following equation.

ee kεσ = where, σe: stress of elastic spring, k: elastic modulus of the spring, εe: strain of elastic spring In this study, elastic modulus is determined as k=25,000 (MPa), which is general Young’s modulus of ordinary concrete. (2) Plastic slider Plastic slider itself does not have time effects, so it should be accord to static behavior and can be formulated based on the mechanical model for static loads. Fig.4 shows stress-strain

(1)

Elastic (for un-cracked zone)

Visco-plastic (for cracked zone)

Figure 2. Mechanical model in this study

Un-cracked zone

Un-cracked zone

Cracked zone

Uni-axial tension

Total deformation

Figure 1. Basic concept – Total deformation is a summation of cracked and un-cracked zone

24 CMT - 194


CMT -

relationship of the plastic slider. Before stress has not reached to tensile strength ft, plastic slider never generates strains because crack does not happen under the lower stress than tensile strength. Once stress reaches to tensile strength, stress is released with softening branch. When strain reaches to a strain εu, which is a limit strain to transfer any tensile stresses between crack surfaces, tensile stress becomes zero. Here, unloading and reloading path are not taken into account. The tensile strain at zero stress εu should be changed with different specimen length. Therefore, it had better to take account into fracture energy concept to determine, but in this paper, it is simply given as constants εu=750μ. Tensile strength of plastic slider ft should accord to the tensile strength of concrete in case of no time effects. Therefore, ft=2.8 (MPa) was determined from the experimental results by Suzuki et al., in which monotonic tensile loading under very lower strain rate was conducted [3]. Stress-strain relationship of the plastic slider is mathematically given as following equation.

( )( )

( )⎪⎪⎩

⎪⎪⎨

⎧

≥=

<<⎟⎟⎠

⎞⎜⎜⎝

⎛−=

<=

upp

upu

ptp

tpp

f

f

εεσ

εεεε

σ

σε

0

01

0 max

where, εp: strain of plastic slider, σp: stress of plastic slider, σpmax: maximum stress of plastic slider in the past, ft: tensile strength of plastic slider, εu: limit strain to transfer stresses (3) Dashpot It is assumed that stress generated by dashpot is proportional to strain rate. The stress can be calculated by the following equation.

dtd

c pv

εσ =

where, σv: stress of dashpot, εp: strain of dashpot, c: viscosity of dashpot, t: passing time Viscosity of dashpot c indicates viscous property between crack surfaces. It should be reduced with crack opening because viscosity never exists when crack surfaces are completely separated away. Therefore, viscosity c is determined by liner reduction model as shown in Fig.4 and then it is given as following equation.

(2)

(3)

Initial viscosity, c0

Viscosity, c

Strain of dashpot, εp

εu

Figure 4. Viscosity reduction model

Tensile strength, ft

Stress, σp

Strain, εp εu Tensile Compression

Figure 3. Stress-strain relationship of the plastic slider

[A]

[B]

[C]

24 CMT - 195


CMT -

( )( )⎪

⎩

⎪⎨

⎧

≥=

<⎟⎟⎠

⎞⎜⎜⎝

⎛−=

up

upu

p

c

cc

εε

εεεε

　　　　　　

　　

0

10

where, the initial viscosity c0 is determined as 50,000 (MPa*sec) so that computed strength accords to strength enhancement with changing strain rates in the experiment conducted by Suzuki et al. [3] Additionally, total stress and total strain is related to each component as follows.

pe

vpe

εεε

σσσσ

+=

+==

2.4 Calculation Flow Digitizing the Eq.(3) by difference method, following equation is obtained.

( ) ( )t

tttc pp

v Δ

Δ−−=

εεσ

where, εp(t): Strain at t, εp(t-Δt): Strain at t-Δt, Δt: Time increment in one step Plastic slider has three kinds of strain paths, [A], [B] and [C] as shown in Fig.4. Solving the simultaneous equations composed of Eq.(1), (2), (5) and (3)’ for each strain path, total stress is related to total strain by following equations. For strain path [A], tp f<maxσ

εσ k=

For strain path [B], up εε <<0

( )

tcf

k

ttt

ct

cff

k

u

t

pu

tt

Δ+−

Δ−Δ

−Δ

+−⋅=

ε

εεεε

σ

For strain path [C], up εε ≥

( )

tck

ttt

ct

c

kp

Δ+

Δ−Δ

−Δ⋅=

εεσ

Figure5 shows the calculation flow of the model. Firstly, total strain and time is given as input value. Then, assuming that plastic slider goes on strain path [A], total stress and stress-strain of each component are calculated using Eq.(6). If stress calculated by Eq.(6) satisfies

(5)

(3)’

(6)

(7)

(8)

Given total strain ε and time t

Calculate total stress using Eq.(6)

tp f<maxσ



START

up εε <<0

END

Calculation finished?

Yes

No

No

Yes

No

Yes

Go to next step

Figure 5. Calculation flow of the model

(4)

24 CMT - 196


CMT -

the condition of strain path [A] (σpmax<ft), it can be deemed that the assumed strain path was correct. Contrary, if it does not satisfy the condition (σpmax<ft), the assumption that plastic slider goes on strain path [A] is not correct. Then whether or not the stress calculated by Eq.(7) satisfies the condition of strain path [B] is examined. If the re-calculated strain does not satisfy the condition of strain path [B] (0<εp<εu), stress is determined by Eq.(8) because remaining strain path [C] is correct.

3. MODEL VERIFICATION

3.1 Parametric Analysis In order to verify the proposed model, parametric analysis was conducted. Table 1 shows strain rate used in the analysis. There are four cases in total. Displacement controlled tensile monotonic loads are applied. Figure6 shows stress-strain diagrams calculated by the model. It shows that, with higher strain rates, strength becomes higher and stress rapidly yields to post-peak range. This tendencies accord to the experimental observation conducted by Fujikake et al. [4]. 3.2 Comparison with the CEB Model Code Strength enhancement of concrete for higher strain rate, both compression and tension, is presented by the CEB Model Code. In tension, the dynamic increase factor, which is ratio of tensile strength under higher strain rate to static strength, is given by

Table 1. Strain rate in the parametric analysis

No. Strain rate (mic/sec)

Case1 1

Case2 10

Case3 20

Case4 30

200 400 600 800 1000

1

2

3

4

0Tensile strain (mic)

Tens

ile s

tress

(MPa

)

Case1

Case2

Case3

Case4

Figure 6. Computed stress-strain diagrams

10 100 10001

5

10

Strain rate (mic/sec)

Dyn

amic

incr

ease

fact

or CEB Model Code Proposed Model

Figure 7. Dynamic increase factor with strain rates – CEB model code and proposed model

24 CMT - 197


CMT -

⎪⎪

⎩

⎪⎪

⎨

⎧

⎟⎟⎠

⎞⎜⎜⎝

⎛

⎟⎟⎠

⎞⎜⎜⎝

⎛

=

×>

×≤

(mic/sec)103for

(mic/sec)103for

73/1

7016.1

ε

ε

εεβ

εε

δ

&

&

&

&

&

&

s

s

ts

t

ff

where, ft: tensile strength at ε& , fts: static tensile strength at sε& , ft/fts: tensile strength dynamic increase factor, ε& :strain rate in the rage of 3 to 3x108 (mic/sec), 3=sε& (mic/sec), 33.211.7log −= δβ ,

)/610/(1 coc ff ′′+=δ , 10=′cof (MPa) Figure7 shows the relationships between dynamic increase factor and strain rate calculated by the proposed model and the CEB model. According to the CEB model code, strength enhancement in logarithmic scale and strain rate becomes linearly. However, strength given by the proposed model is much larger than it when strain rate is greater than 100 mic/sec. It means, the time effect on strength enhancement caused by dashpot in proposed model is too oversensitive for such a higher strain rate. 4. CONCLUSIONS In this paper, applicability of elasto-visco-plastic model, in which elastic strain and visco-plastic strain components are connected in series, for concrete under tensile loads was verified. As a result, followings were concluded. 1) When concrete has different strain rates under tensile loads, strength and stress-strain relationship

become larger and more brittle with higher strain rates, and smaller and more ductile with lower strain rates. These tendencies can be expressed by elasto-visco-plastic model proposed in this paper.

2) According to the CEB model code, logarithm of tensile strength is proportional to logarithm of

strain rate. However, tensile strength computed by proposed model in this paper became much larger than it. That is, time effect on strength is too oversensitive.

5. REFERENCES Kohei Nagai, Yasuhiko Sato and Tamon Ueda (2004). “Mesoscopic Simulation of Fracture of Mortar

and Concrete by 2D RBSM.” Journal of Advanced Concrete Technology, Vol.2, No.3, pp.359-374, October

Koji Matsumoto, Yasuhiko Sato and Akihiro Tateishi (2004). “A Study on Prediction Method of Concrete Deformation under Compressive Fatigue Loading.” Proceedings of the First Internatinal Conference of Asian Concrete Federation, Vol.1, pp.399-408, October

Masahiro Suzuki, Hirotaka Kawano, Hiroshi Watanabe and Yoshiki Tanaka (1999). “Effect of Strain Rate on Tensile Strength of Concrete.” Proceedings of JCI, Vol.21, No.2, pp.649-654 (in Japanese)

Kazunori Fujikake, Katsutoshi Uebayashi, Tomonori Ohno and Katsuhiko Emori (2001). “Study on Dynamic Tensile Softening Characteristic of Concrete material under High Strain-Rates.” Proceedings of JSCE, No.669/V-50, pp.125-134, February (in Japanese)

L. Javier Malvar and C. Allen Ross (1998). “Review of Strain Rate Effects for Concrete in Tension.” ACI Materials Journal, Vol.95, No.6, November-December, pp.735-739

(9)

24 CMT - 198


CMT -

EFFECTS OF LOW SURFACE TENSION WATER CURING ON SURFACE ABSORPTION AND DEGREE OF HYDRATION

Komsan Maleesee1 and Amnouy Panitkulpong2

ABSTRACT : The reaction between cement and mixing water causes difference of the hardened paste. In the case of high strength concrete, the mix proportion does not have enough water to complete the hydration. The penetrating water from outside is important to continue the hydration reaction. However, in the low water-cement ratio (W/C) and large specimen, the curing water cannot completely penetrate to fill pores formed during the hydration reaction thus causing different degree of hydration and degree of self-desiccation in the surface layer and in the bulk of the specimen1)-3). For this reason, influence of dynamic character of high strength concrete is considered in these experiments. Additionally, possibility of applying the same method of curing and testing on the normal strength concrete and on the high strength concrete will be discussed. In this paper, effects of low surface tension water curing on surface absorption and degree of hydration were experimentally investigated. The results of experiment were expressed by measuring the amount of water penetrated into the hardened cement paste at various distances from the surfaces. The measurement was compared with chemical shrinkage, degree of hydration, and absorption, SEM was used to investigate the hydration products of these specimens at any types of curing. Effects of penetrative curing water on degree of hydration were investigated. In a large-size specimen with a low water-cement ratio (W/C), both the mixing water and penetrative water at standard water curing are not sufficient to complete the hydration of the entire specimen. Therefore, curing conditions were studied to improve the properties of the specimens at any distances from the surface during hydration process. When W/C is low, specimens cured with water containing AE-admixture, which has low surface tension, was able to improve the degree of hydration more than those subjected to the standard water curing. KEYWORDS : Water penetration, Hydration, Chemical shrinkage, Self-desiccation, Low surface tension water curing, Absorption

1. INTRODUCTION

Proper curing is essential to achieve a discontinuous pore structure, which is the discontinuity in capillary pore network that formed during the hydration in concretes. With W/C lower than 0.45, the discontinuous pore structure will be formed easily. The discontinuous pore structure is important for a durable concrete, as it will limits both water and ion ingress into the concrete structures4). Based on the measurement of permeability, Powers has concluded that capillary pore discontinuity is a function of both W/C and degree of hydration: a higher W/C requiring longer hydration time to achieve a discontinuous capillary pore structure5). He, therefore, suggested to cure concretes until this discontinuity is achieved, as further “saturated” curing would result in vain unless any additional water flows into the concrete6). Self-desiccation takes place when the pores created by the hydration are not supplied with water from the surrounding environment during the curing process1), and empty pores are created within the microstructure8). In construction, there are several problems encountered nowadays to receive a good quality of cure concrete, for instance, construction time, project cost as well as knowledge. In order to fulfill these problems, the AE-water curing may becomes a good solution for helping curing in-site water curing and prefabricated samples, since it easily prepares a variety of materials such as beam, column, slab, etc.

1 Dr.E, Civil Engineering, King Mongkut Institute of Technology Ladkrabang, Thailand, [email protected] 2 Associate Professor, Civil Engineering, King Mongkut Institute of Technology Ladkrabang, [email protected]

25 CMT - 199


CMT -

2. MATERIALS

An ordinary Portland cement with characteristics as shown in Table 1 was used. For the compressive strength test of mortar, the crushed sand (density: 2.58 g/cm3, absorption: 2.28 %) was used as a fine aggregate. An air entraining and high range water-reducing agent were used for specimens with a W/C of 0.2, 0.25 and 0.3. A Silicone Antifoaming Agent was added to reduce the air content in all type of pastes. The air content of pastes was decreased less than 0.1%. Alkyl-ether based air-entraining agent of 1% added with water, in accordance with JIS A6204 of, was used in AE-water curing to reduce surface tension of curing water. This amount is quite larger than the case of concrete mixture. The surface tension of water for AE-water curing was approx. 38 mN/m which is nearly one half of that used in standard water curing. It was measured by Du Nouy surface-tension meter at a room temperature of 22oC and 50 % RH.

3 EXPERIMENTAL PROCEDURES

3.1 Test method for chemical shrinkage Mass of a glass vessel was measured with accuracy of 0.1g. This value was referred to as M1(g). The cement paste with a W/C of 0.2, 0.3, 0.4 and 0.6 were placed in the glass vessel with a thickness of approx. 10mm in height, and then the total mass of cement paste and the vessel were measured with an accuracy of 0.1g. This value was referred to as M2(g). The vessel was filled slowly with water in order to prevent the disturbance of the interface, and then a measuring pipette was inserted perpendicularly

Mineral Composition (mass%) Cement

Fineness

(cm2/g)

Density

(g/cm3) C3S C2S C3A C4AF CaSO4

Ig. Loss

(mass%)

OPC 3290 3.16 62.4 13.5 8.3 8.5 2.4 2.01

Figure 1. Measuring method of chemical shrinkage

Cement paste

Rubble plug

Measuring pipette

water

Water level

50 mm

50 mm

300 mm

25 mm

Cutting the specimen in the six parts to measure penetrated water

Curing conditions

Water, AE-water, 10 MPa-water, Wrapping

Figure 2. Measuring method of the penetrative curing water in hardened cement paste

Acrylic pipe

Table 1. Physical properties and mineral composition of cement

Paraffin wrap

25 CMT - 200


CMT -

into the sample vessel through a rubber plug. Subsequently water was added from the upper inlet of the measuring pipette by a funnel or washing bottle to a measuring range, and then sealed the upper parts of measuring pipette with a paraffin wrap and this specimen was stored in the curing room as shown in Fig.1. The initial water level in the pipette was read with an accuracy of 0.05 ml. This value was referred to as H0 (ml) when starting the measurement. Subsequently the values of water level in the pipette were read during hydration. These values were referred to as Hn (ml). The chemical shrinkage of hardened cement paste can be calculated using Eqs.(1) and (2). 3.2 Test method for water penetration in hardened cement paste In this test method, the masses of each hardened cement pastes, as prepared at a room temperature 22oC, were measured by M21(g). Subsequently, the specimens were stored for 24 hours in an oven at a temperature of 105oC and their masses were measured again (M105). Finally, the specimens were stored for 24 hours in an electric furnace at a temperature of 950oC and then their masses were measured (M950). Cement pastes with a W/C of 0.25, 0.40 and 0.6 were placed in the acrylic pipe with a diameter of 25 mm and with a paste height of 300 mm. The specimens were cured after one day using four different types of curing: the standard water curing (Water curing), the low surface tension water curing (AE-water curing) and the sealed curing(Wrapping). The penetrated water was determined after 3, 7 and 28 days by cutting the specimen in six parts as shown in Fig.2 and their masses were measured. In a specimen with a W/C of 0.4 and 0.6, bleeding generally occurs. However, it will be ignored at this stage because we used double mixing method for mixing cement paste to decrease the bleeding. 3.3 Curing conditions 1) The standard water curing (Water curing): The Water curing with a surface tension approx. 70 mN/m was used as the standard water curing. 2) The low surface tension water curing (AE-water curing): Alkyl-ether based air-entraining agent (MA101, MA202, MA775) of 0.1~2.0% were added with water to measure surface tension of AE-water as shown in Fig.3. The surface tension of AE-water which added MA101 is lowest so it was used in this thesis. Furthermore, different amounts of cement (0, 1.0%, 3.33% and 10.0 % of water)

30

35

40

45

50

0.0 0.5 1.0 1.5 2.0Amount of admixture mixed with water (%)

Surf

ace

tens

ion

(mN

/m)

C=0 (W × wt%)C=1.00 (W × wt%)C=3.33 (W × wt%)C=10.0 (W × wt%)

MA101

30354045505560657075


Surf

ace

tens

ion

(mN

/m)

MA101

MA202

MA775

Figure 3. Surface tension of curing water at different type of AE mixed

Figure 4. Surface tension of AE-water curing (MA101) at different amount of cement mixed

25 CMT - 201


CMT -

were added in AE-water to examine an appropriate amount of AE admixture as shown in Fig.4. At the point of MA101 1% added with water, its surface tension did not change although it was added with different amounts of cement. For this reason, MA101 1% was added with water to be used in AE-water curing. Its surface tension was approx. 38 mN/m as shown in Fig5. 3) Sealed curing (Wrapping): plastic sheets were used to prevent the water movement through the surfaces.

4. CALCULATION

In this research, air content was reduced by adding Silicone Antifoaming Agent. Thus, in all cases of calculation, the air content was neglected. The chemical shrinkage ratio of cement paste as described in 3.1 is given by Eq.(1)1) .

0 100 (%)nhyn

p

H HS

V−

= × (1)

where Shyn is chemical shrinkage ratio at an age n, H0 (ml) and Hn (ml) are water level at the start and at the age n and Vp is the volume of cement paste as given by Eq.(2)1).

2 1( ) (( / ) / 1/ )( )

/ 1w c

pM M W C D D

V mlW C

− × +=

+ (2)

where M2 is the mass of cement paste and glass vessel, M1 is the mass of glass vessel, and Dc is density of cement = 3.16 g/cm3. The total volume of the paste per unit mass of cement (Vp/c) is given by Eq.(3) 2).

3/ 1/ ( / ) /p c c wV D w c D cm= + /g (3)

where Dw is the density of water = 1 g/cm3.

Figure 5. Surface tension energy of AE-water curing (MA101)

Selected point that used in this paper

01020304050607080


Surfa

ce te

nsio

n (m

N/m

)

25 CMT - 202


CMT -

The volume of evaporable water per unit mass of cement (Vew/c) is obtained by the amount of gel water and capillary water that can be evaporated at 105oC from the experiment (2).b, and given by Eq.(4)

2),3).

321 105/

950

( ) / /wew c

M M DV cm gM

° °

°

−= (4)

where M21 is the mass of specimen at room temperature, M105

o and M 950o are the mass of the

specimen after heating for 24 hours at 105oC and 950oC, respectively. The non-evaporable water contains nearly all chemically combined water. This water has a vapor pressure lower than that of the ambient atmosphere and the quantity of such water is in fact a continuous function of the ambient relative humidity. The volume of non-evaporable water per unit mass of cement (Vnew/c) can be determined by the loss upon drying at temperatures from 105oC to 950oC. It can be expressed by Eq.(5) 2),3).

3105 950

/950

/nnew c cm gM MV V

M° °

°

−= × (5)

Degree of hydration (DH) is obtained by the proportionality between the amount of non-evaporable water and the solid volume of the cement paste where the former volume can be used as a measure of the quantity of the degree of hydration. It can be expressed by Eq.(6).

105 950

950( ) 100 / 23 %M MDH

Mα ° °

°

−= × (6)

where 23% is the content of non-evaporable water at the complete hydration2),3).


5.1 Chemical shrinkage and degree of hydration The chemical and physical changes that occur during hydration are accompanied by a reduction in absolute volume. The combined volume of the liquid and solid components after hydration is less than the initial volumes of water and anhydrous cement. The chemical shrinkage for cement paste in different water-cement ratio is shown in Fig.6. The change of chemical shrinkage for cement paste with W/C of 0.2 and 0.3 was very large at the early stage, but it decreased after approx. four or five days and became blunted after 28 days due to the absence of enough water to react with cement. The change of chemical shrinkage of cement paste with W/C of 0.4 and 0.6 lasted slowly after 28 days until the hydration completed. The degree of hydration (DH) for pastes with different water-cement ratios are shown in Fig.7 and Fig.8. DH is calculated by chemical shrinkage as shown in Fig.7 and DH can be also calculated from the quantity of non-evaporable water at that time as shown in Fig.8. DH for specimen with W/C of 0.2 and 0.3 increased very

0

2

4

6

8

10

1 10 100Hydration time (days )

S hyn

(%)

W/C = 0.6W/C = 0.4W/C = 0.3W/C = 0.2

Water curing

Figure 6. Relation between chemical shrinkage and hydration time

25 CMT - 203


CMT -

slowly after four or five days and finally becomes blunted after 28 days because of the lower content of water than that necessary to continue the hydration. The cement pastes with water-cement ratio of 0.4 and 0.6 seemed to have enough water to continue the hydration reaction even after 28 days.

0

20

40

60

80

100


DH

(%)

W/C = 0.6W/C = 0.4W/C = 0.3W/C = 0.2

Water curing

Figure 7. Relation between degree of hydration and hydration time (by chemical shrinkage)

0

20

40

60

80

100

1 10 100Hydration time (days)

DH

(%)

W/C = 0.6W/C = 0.4W/C = 0.3W/C = 0.2

Water curing

Figure 8. Relation between degree of hydration and hydration time (by non-evaporable water)

Figure 9. Relation between Self-desiccation and hydration time

Figure 10. Relation between DH and distance from surface at 28 days (Wrapping)

505560657075808590

0 50 100 150 200 250 300Distance from surface (mm)

DH

(%)

w/c 0.25w/c 0.40w/c 0.60 Wrapping

0369

1215182124


DS

(%)

W/C = 0.6W/C = 0.4W/C = 0.3W/C = 0.2

Water curing

25 CMT - 204


CMT -

5.2 Effect of W/C on degree of self-desiccation The degree of self-desiccation of specimens with different W/C is shown in Fig.9. It was calculated from the hardened cement pastes with thickness of about 5 mm. The specimen with W/C of 0.2 shows larger range of self-desiccation during the hydration while that with W/C of 0.6 shows small range of self-desiccation. The change in DH became blunted after 28 days for the specimens with W/C of 0.2 and 0.3 because the near-surface-area became hardened and impenetrable for the water from outside. However, the change in DS still gradually increased because the inside of specimens there remained a little water to react with cement and the empty pores were formed. Likewise, in the specimen with W/C of 0.6, the near-surface-area became hardened during the hydration, and the water penetration to fill pores was partly obstructed. Therefore, it is thought that DS is gradually increased until 12% after 90 days as shown in Fig.9. The hydration increases the gel content and, in mature and dense paste, the capillaries may become segmented by the gels and interconnected solely by the gel pores. The absence of continuous capillaries may be obtained by a combination of suitable W/C and moist curing of sufficiently long time. Thus, if a specimen with a low W/C is not given more water from the outside, the larger range of self-desiccation during hydration may occur.

5.3 Effect of distance with any types of curing on degree of hydration

The content of non-evaporable water relative to that in a fully hydrated paste of the same cement was used as a measure of the degree of the hydration according to Powers-Brownyard model. The degree of hydration calculated from Eq.(6) for any distance of the specimen with water-cement ratios of 0.25,

505560657075808590


DH

(%)

w/c 0.25w/c 0.40w/c 0.60

505560657075808590


DH

(%)

w/c 0.25w/c 0.40w/c 0.60

Figure 11. Relation between DH and distance from surface at 28 days (Water curing)

Figure 12. Relation between DH and distance from surface at 28 days (AE-water curing)

Water curing

AE-water curing

25 CMT - 205


CMT -

0.40 and 0.60 after 28 days are shown in Figs.10-12. The different degrees of hydration, especially in the specimen with a low water-cement ratio, were found for the three types of curing used in this investigation. It could be seen that for the specimen with W/C of 0.25 treated by AE-water curing, the hydration progressed more than those subjected to the standard water curing and sealed curing. However, for the specimen under AE-water curing, the degree of the hydration was about 65 % at a distance smaller than 175 mm, and then slightly decreased at larger distances of specimen as shown in Fig.12. The degree of hydration in the specimens with water-cement ratio of 0.40 and 0.60 was not so affected by the types of curing.

5.4 Effect of curing on hydration products of hardened cement paste investigated be SEM As discussed before about the effects of water penetration into the low W/C specimens, the hydration reaction, non-

evaporable water, and strength increase when water from outside penetrates into the specimens because of the penetrated water filled pores. May be it was made some products of hydration in the pores. Therefore, effect of curing (AE-water curing, Water curing) on hydration products of hardened cement paste with water-cement ratio of 0.25 after 28 days were investigated by Scanning Electron Microscope (SEM) as shown in Figs.13-14. The pores were concentrated investigation of hydration products by the SEM. The different types of curing have effects with amounts and characteristics of pores as shown in Figs.13-14. The pores were filled with some product of hydration for the specimens curing with AE-water curing as clearly shown in Fig.14, but in case of Water curing the empty pores were occurred as shown in Fig.13.

6. CONCLUSIONS

1) For the hardened cement pastes with low W/C, such as W/C of 0.20, 0.25 and 0.30, it was found that the mixing water was unable to fill pores formed as a result of hydration reaction, and to penetrate into the central part of hardened cement pastes. A larger self-desiccation occurs in the middle part of specimens and the progress of hydration became slower.

2) At any distance from surface to middle part of the specimen with a low W/C, different degree of hydration and self-desiccation due to the lack of hydration were shown.

3) The low surface tension of AE agent, the AE curing water was able to penetrate well into the specimen, in the distance less than 175 mm from the surface area, more than that with the standard water curing, especially at a low W/C.

100 μm

Figure 13. A picture of specimen with W/C of 0.25 curing by Water

25 CMT - 206


CMT -

4) The high-strength concrete with low W/C that treated by the standard water curing clearly showed difference in degree of hydration and self-desiccation in the bulk and in the surface layer. Consequently, it is not suitable to apply the same curing condition as applied to the normal concrete because of the difference in surface tension. The degree of the penetration of water into the specimen and degree of self-desiccation were different by the curing water. Further investigations to find the appropriate curing condition as standard for the high-strength concrete, are needed. Nevertheless in this paper, it was found that low surface tension water curing was able to develop hydration reaction more than that with the standard curing, especially at a low W/C ratio.

7. REFERENCES

Tazawa, E. (1998). “Autogenous Shrinkage of Concrete.” Proceeding of the International Workshop organized by JCI, pp.3-67.

Taylor, H.F.W. (1997). “Cement chemistry.” 2nd edition, pp. 227-255. Neville, A.M. (1963). “Properties of Concrete.” pp. 5-53. Bentz, D.P. (2002). “Influence of Curing Conditions on Water Loss and Hydration in Cement Pastes

with and without Fly Ash Substitution.” National Institute of Standards and Technology, NISTIR 6886.

Powers, T.C. (1959). “Capillary Continuity or Discontinuity in Cement Paste.” PCA Bulletin, No. 10, pp. 2-12.

Powers, T.C. (1974). “A Discussion of Cement Hydration in Relation to the Curing of Concrete.” Proc. Of the Highway Research Board, 27, pp.178-188.

Bentz, D.P., Garboczi, E.J. (1991). “Percolation of Phases in a Three-Dimensional Cement Paste Microstructure Model.” Cement and Concrete Research, Vol.21, pp. 325-344.

Power, T.C. (1935). “Absorption of water by Portland cement paste during the hardening process.” Industrial and Engineering Chemistry, Vol. 27, pp. 790-794.

Maleesee, K., Kasai, T. (2001). “Effect of Penetration of Curing Water on Self-desiccation and Strength of Cementitious Materials.” Cement Science and Concrete Technology by JCA, Vol.55, pp. 109-115.

Maleesee, K., Kasai, T. (2002). “Effect of Penetrative Conditions. of Water for Curing on Strength of Concrete.” Proceeding of the first fib Congress 2002, Osaka, Japan, pp. 289-296.

Maleesee, K., Panitkulpong, A., Kasai, T. (2003). “Effect of Penetrative Conditions of Water for Curing on Properties of Hardened Cement Paste.” Concrete Research and Technology by JCI, Vol.25, pp. 581-586.

100 μm

Figure 14. A picture of specimen with W/C of 0.25 curing by AE-water

25 CMT - 207


CMT -

Maleesee, K. and Kasai, T. (2004). “Influences of Penetrative Curing on Properties of Cementitious Materials.” Journal of Materials, Concrete Structures and Pavements, JSCE, No. 767 / V-64, pp. 301-312.

25 CMT - 208


CMT -

INFLUENCE OF KINDS OF ORGANIC ADMIXTURES ON SHRINKAGE OF MORTAR

Toyoharu Nawa1 and Mari Masanaga2

ABSTRACT: Recently durability of concrete has drawn attention from many researchers.

Cracks at the early ages of concrete due to shrinkage usually cause the decreasing in durability. In

this study, we investigated the effect of organic admixtures on the autogenous and drying

shrinkage of mortar. Four different kinds of organic admixture were used; three kinds of polymer

introduced acrylate and one kind of polyethyleneimine based admixtures. Strain gauge cover with

silicon and plastic are used to measure autogenous shrinkage and drying shrinkage, respectively.

For the internal relative humidity, ceramic sensors are applied. The specimens were sealed for 7

days at 20˚C and continued until 28 days at 20˚C, at 60% of relative humidity. The results

indicated the relation between the surface tension and the autogenous shrinkage. The autogenous

shrinkage increased as the surface tension increased. However, the relation was strongly

influenced by kinds of organic admixture. For the drying shrinkage, there was no relation between

the surface tension and the drying shrinkage. This means that only surface tension is not enough to

explain autogenous mechanism. The adsorption and strength studies indicated that polymers

having more hydrophobic groups adsorbs strongly on the hydrated cement particles, resulting in

the lower strength development. Further, comparisons between the sizes of absorbed polymers and

the pore sizes in mortar might imply that the absorbed hydrophobic groups of polymer influence

the autogenous shrinkage and drying shrinkage as well as the surface tension of pore water.

KEYWORDS: Shrinkage Reducing Admixture, Surface Tension, Chemical Structure,

Autogenous Shrinkage, Drying Shrinkage, Compressive Strength, Adsorption

1. INTRODUCTION

Concrete is one of the vital materials for our infrastructure and is widely used in construction all over the world. The durability of concrete is defined as the ability of the material to remain serviceable for at least the required lifetime of the structure. Accordingly, its durability is essential in preserving the infrastructure of society. Deterioration of concrete is induced by penetration of trigger substances such as chlorides and carbon dioxide gas into concrete through its surfaces. Cracking in concrete due to autogenous and drying shrinkage accelerates such penetration, aggravating deterioration, so it has been a matter of great concern when focusing on maintaining durable structure. Therefore more effective drying shrinkage-reducing admixtures and methods of reducing shrinkage are urgently required. 1 Professor of Socio-Environmental Engineering at Hokkaido University, Japan 2 Division of solid waste, resources and geoenvironmental engineering, Graduate School of Engineering, Hokkaido

University, North 13 West 8, Kita-ku, Sapporo, Japan 060-8628

26 CMT - 209


CMT -

Several drying shrinkage mechanisms for concrete such as capillary tension, surface energy, disjoining pressure, movement of interlayer water and others have been proposed. However it seems that these mechanisms actually would combine each other to cause drying shrinkage [1]. Many types of shrinkage-reducing admixture have been developed according to the capillary tension theory [2-5]. Most of the commercial shrinkage-reducing admixtures consist of basically polyoxyalkyleneglycols or their alkyl ethers. But it is well known that the shrinkage-reducing admixture generally retards hydration of cement, and reduces compressive strength. Therefore, further improvement in shrinkage-reducing admixtures is anticipated. In this paper, the authors synthesized four polymers with different chemical compositions as shrinkage reducing agent (SRA). These polymers were used to investigate the effects of their surface tension and chemical structure on the shrinkage behavior of mortar, the hydration of cement and the strength development of mortar.

2. EXPERIMENTAL

2.1 Material and mix proportion Mortars were prepared using ordinary Portland cement (OPC) and land sand used in this study. The mix proportion of mortar is shown in Table 1. The cement has a Blaine’s fineness of 351m2/kg and a

Bogue phase composition of 61.0%C3S, 13.5%C2S, 8.3%C3A, and 8.8%C4AF. The land sand has a density of 2,710 kg/m3, absorption of 1.74% and fineness modulus of 3.06.

Table 1 Mix propor t ions of mor tarRatio of Aggregate W C S 0 hit Air

W/C to Mortar (kg/m3) (kg/m

3) (kg/m

3) kinds dosage(%) kinds dosage(%) mortar flow (%)

30HU7 HU7 0.1830HU630KC4 KC430KC5 KC530KC6 KC630KC5' KC5 0.00530MOE MOE 0.20

50N - -50HU7 HU7 0.0350HU650KC4 KC450KC5 KC550KC6 KC650KC5' KC5 0.00550MOE MOE 0.20

0.50 0.47 649 1255325

200±10 5.0±1.0

HU6 0.170.50

- -

HU6 0.020.50

numberSP SRA

0.30 0.47 860 1255

- -

258

KC4, KC5 and KC6 compose of copolymer of acylate and methacrylate. On the other hand, MOE, newly developed SRA, is a hydrophilic polymer having hydrophobic groups and carboxyl groups which can adsorb onto cement particles. The flowability were controlled by using two superplasticizer (SP); one (HU7) is conventional superplasticizer which have polymetacrylic polymers in main chain

Table 1. Mix proportions of mortar

26 CMT - 210


CMT -

and polyethylene oxide groups in the graft chains, and the other (HU6) is the trially synthesized one having the alkyl groups to the end of main chain. The air control admixture (AE-1) was used to control the air content.

2.2 Preparation of mortar and experimental procedure

The air content of mortar was adjusted to 7 ± 1 volume%, when necessary, by using the air control admixtures. Mixing of mortar was carried out according to JIS R5201 “Physical Testing Method for Cement” by using a Hobart type mortar mixer. Water, containing SRA and other admixtures, and OPC

were mixed for 30 seconds at slow speed (rotation:140±5 rpm, revolution:62±5 rpm), then land sand was added over 30 seconds while mixing at slow speed. After that, the mixing speed was changed to

medium speed (rotation:285±10 rpm, revolution:125±10 rpm) and the mortar was mixed for 30 seconds. Then, mixing was stopped for 15 seconds, during this interval any mortar on the side of the bowl was quickly scraped down into the batch. After that, more mixing for an additional 60 seconds at medium speed was performed, and the mortar was obtained for evaluation.

Cylindrical specimens of 100 x 50 mm diameter were used for the autogenous shrinkage and drying shrinkage measurement. The fresh mortar was cast into a cylindrical mold. A strain gauge and thermocouples were equipped with a digital computer system [6]. The strain gauge and thermocouples were vertically set in the center of a cylindrical mold 50 mm in diameter and 100 mm in height, placed. A polytetrafluoroethylene sheet was placed between the cement system and the mold to reduce the friction between them. The top surface of the sample were sealed after placing to prevent moisture dissipation. Then, immediately after mixing, both shrinkage strain and temperature changes were continuously monitored. Measurement of drying shrinkage was also performed using the same specimen for measurement of autogenous shrinkage. The specimens were demoulded at age of 7 days and stored in controlled environment of 20˚C and 65% of relative humidity. The change in internal relative humidity due to self-desiccation and moisture evaporation is closely related to its autogenous shrinkage and drying shrinkage, respectively. Thus the internal relative humidity in mortar was measured using a ceramic sensor [7]. Further, we measured the combined water in hydrates and the compressive strength of mortar to confirm the effect of SRA on hydration of cement. Cement paste was immersed in acetone to stop the hydration and dried for 6 hours at 40˚C. The amount of combined water was calculated by mass loss at 1000˚C. The compressive strength tests for mortar were carried out on cylindrical specimens at 3,7,14 and 28days.

After stirring 400g of cement and 800g of water for 30 minutes, and filtrating with 5C filter paper, the cement filtrate (A) was obtained. The surface tension was measured at 0.01, 0.05, 0.5, 1.0 and 5.0

26 CMT - 211


CMT -

mass% of shrinkage reducing agent in the cement filtrate (A) at 20±3˚C by the duNouy method using a platinum ring. 100g of cement was added to 200g of the aqueous solution containing 0.2 mass% of SRA and stirred. Taking portions of the filtrate after 1 minute and 60 minutes stirring, and filtrating them, the cement filtrate (B) was obtained. The surface tension of the cement filtrate (B) was again measured by the duNouy method mentioned above. The adsorption of shrinkage reducing agent on cement particles was measured by total organic carbon (TOC) measurement system. Samples for TOC measurement were obtained as follow. 100g of ordinary portland cement and 200g of the aqueous solution containing 0.2 mass% of SRA were stirred at 700 rpm. About 20g of mixture was extracted at 5, 20 and 60 minutes stirring after adding OPC, and placed for 5 minutes, then the supernatant solution of mixture was filtrated with filter paper and chromat filter (Chromatdisk 25A).


3.1 Effect of polymer on shrinkage behavior

Figure 1 shows the total (autogenous and drying) shrinkage strain behavior of mortars containing SPs and SRAs to an age of 28 days. A shrinkage-reducing effect of SRAs is recognized in mortars with a W/C of 30%. In contrast, for mortar of W/C of 50 %, the significant difference could not detect until the age of 7 days.

Figure 2 shows the relationship between the autogenous shrinkage of mortar containing SPs and SRAs and the surface tension of cement filtrate containing these polymers. From Fig. 2 (a), the shrinkage reductions associated with the decreases in the surface tension due to the addition of SPs and SRAs are recognized in mortars with W/C of 30%. However, it was also found that such a shrinkage-reducing effect of SP and SRA polymer varies depending on the type of polymers: the plots of mortars

Fig.1 Effect of SP and SRA on shrinkage of mortar until the age of 28 days

Fig. 2 Effects of SP and SRA on shr inkage of mor tar for 28

(a) W/C=30% (b) W/C=50%

-1400

-1200

-1000

-800

-600

-400

-200

0

0 7 14 21 28Age(day)

Shrinka

ge(×

10

-6)

30HU7 30HU630KC4 30KC530KC5' 30KC630MOE

-1400

-1200

-1000

-800

-600

-400

-200

0

200

0 7 14 21 28

Age(day)

Shrinka

ge(×

10

-6)

50N 50HU7

50HU6 50KC4

50KC5 50KC5'

50KC6 50MOE

(b) W/C=50%

Figure 1. Effects of SP and SRA on shrinkage of mortar for 28

26 CMT - 212


CMT -

containing HU7, KC4, KC5, and KC6 are nearly on the same straight line, whereas those of mortars containing HU6 and MOE are apart from this line. It was therefore found that the influence of surface tension on the autogenous shrinkage-reducing effect of polymers varies depending on their chemical structure. From Figure 2 (b), mortars with W/C of 50% containing SPs and SRAs also show a correlation between the autogenous expansion and surface tension of cement filtrates containing these polymers, but it is not as significant as with 30% W/C. Further, it can be found that the influence of surface tension on the autogenous expansion is depending on the type of SP and SRA polymers. It should be noted that, among these SRAs having both hydrophobic and hydrophilic groups, MOE has the hydrophobic groups in all side chains, whereas KC4, KC5, and KC6 have both hydrophobic and hydrophilic groups in their side chains. This may suggest that the balance of hydrophobic and hydrophilic group in side chains of polymers may influence their effect of reducing the autogenous shrinkage of mortar.

-800

-700

-600

-500

-400

-300

-200

-100

35 40 45 50 55 60 65 70 75

Surface Tension(dyn/cm)

Shrinka

ge(×

10

-6)

30HU7 30HU6

30KC4 30KC5

30KC5' 30KC6

30MOE

Fig. 3 Relat ion between surface tension and shr inkage of mor tar added SP and SRA at age of 7

(b) W/C=50%(a) W/C=30%

-100

0

100

200

35 40 45 50 55 60 65 70 75


Shrinka

ge(×

10

-6)

50N 50HU7 50HU6

50KC4 50KC5 50KC5'

50KC6 50MOE

Figure 3 shows the relationship between the drying shrinkage from the age of 7 days and surface tension for mortar containing SPs and SRAs. As shown in Fig. 3 (a), in mortars with a W/C of 30%, the surface tension has no effect on drying shrinkage. In contract, for mortar with W/C of 50%, the addition of SRAs reduces the drying shrinkage by around 10%, as shown in Fig. 3(b). Further, there is a relatively good correlation between the surface tension and the drying shrinkage in contrast to the case of 30% W/C. Furthermore, it is obvious that the effect of MOE on the reduction of drying shrinkage was almost the same as those of mortars containing KC series polymers. Comparing with the results shown in Fig.2, it was therefore confirmed that the tendency on the effect of the chemical structure of SRAs on drying shrinkage differs from that on autogenous shrinkage.

Figure 2. Relation between surface yension and shrinkage of mortar added SP and SRA at age of 7

26 CMT - 213


CMT -

-1100

-1000

-900

-800

35 40 45 50 55 60 65 70 75


Shrinka

ge(×

10

-6)

50N 50HU7 50HU6

50KC4 50KC5 50KC5'

50KC6 50MOE

Fig. 4 Rela t ion between sur face tension and shrinkage of mor tar added SP and SRA at age of 28 days

(b) W/C=50%(a) W/C=30%

-700

-600

-500

35 40 45 50 55 60 65 70 75


Shrinka

ge(×

10

-6)

30HU7 30HU6

30KC4 30KC5

30KC5' 30KC6

30MOE

3.2 Effect of SRAs on the hydration of cement and compressive strength development It is well known that conventional SPs and SRAs affect the hydration of cement and retard the early-age strength development. Thus, the effects of SPs and SRAs on the hydration of cement and the strength development are investigated in this section. The change in bound water of mortar added with SPs and SRAs is measured to estimate the effect of newly synthesized SP and SRAs on the hydration of cement. The results are shown in Fig. 4. The bound water in mortars with W/C of 30% containing SRAs such as KC5 and MOE is lower than that of the mortar without SRA by around 2% at all ages. In contrast, mortar with a W/C of 50% shows no such hydration retardation until 3 days although it shows the retardation of around 2% at 7 days.

0

2

4

6

8

10

12

14

16

18

1day 3day 7day

Rat

io o

f B

ondi

ng

Wat

er(

%)

30HU7 30KC5 30MOE

0

2

4

6

8

10

12

14

16

18

1day 3day 7day

Rat

io o

f B

ondi

ng

Wat

er(

%)

50HU7 50KC5 50MOE

(a) W/C=30% (b) W/C=50%

Fig. 8 E ffects of SP and SRA on r a t io of bonding water of cement paste

This can be explained by taking into account the fact that SP polymers can retard cement hydration. The dosage of polymers is higher for paste with W/C of 30% than that for paste with W/C of 50%.

Figure 3. Relation between surface tension and shrinkage of mortar added SP and SRA at age of 28

Figure 4. Effects of SP and SRA on bonding water of cement paste

26 CMT - 214


CMT -

Accordingly, the retardations of cement hydration with different W/Cs can primarily be also attributed to the difference in the SP dosage. Figure 5 shows the compressive strength of mortar containing SPs and SRAs. The strength of mortars containing SRAs are lower than that of mortar without SRAs. In particular, polymers having more hydrophobic groups, such as KC6 and MOE, led to greater strength losses. As previous mentioned, the MOE showed the same effect on the retardation of hydration, and accordingly it seems that the effect of SRAs on strength development was not directly related to the retardation of hydration. The strength of cement mixes depends on the interaction between hydrated cement particles. On the other hand, it is well known that the adsorbed polymer produced the interaction such as steric repulsion or bridging attraction. Hence it can be expected that the ratio of hydrophobic groups in polymer molecule might have an influence on the interparticle forces of cement particles, and this can be attributed to the strength development. The amounts of adsorption on cement particles also varied. The adsorption percentages of KC4, KC5, and KC6 were as low as 0%, 3.7%, and 7.7%, respectively, after an hour, whereas that of MOE reached 90.0%. After 7 hours, the values for KC4, KC5, and KC6 increased to 4.6%, 18.9%, and 36.1%, respectively. These results imply that polymers having more hydrophobic groups, such as KC6 and MOE, adsorbs strongly on the hydrated cement particles, resulting in the lower strength development.

3.3 Effect of SRA on internal relative humidity According to Kelvin theory, the relative humidity can be expressed as a function of pore radius and surface tension of pore solution as follows:

srRT

Mpp 12ln0 ρ

γ−= (1)

(a) W/C=30% (b) W/C=50%

Fig. 7 E ffects of SP and SRA on compressive strength of mor tar

0

10

20

30

40

50

60

70

80

90

100

50N 50HU7 50HU6 50KC4 50KC5 50KC5' 50KC6 50MOE

Com

press

ive S

trengt

h(N

/m

m2 )

3day 7day 28day

0

10

20

30

40

50

60

70

80

90

100

30HU7 30HU6 30KC4 30KC5 30KC5' 30KC6 30MOE

Com

press

ive S

trengt

h(N

/m

m2 )

3day 7day 28day

Figure 5. Effect of SP and SRA on compressive strength of mortar

26 CMT - 215


CMT -

where p is a partial pressure of water vapor, p0 is a saturated water vapor pressure, γ is a surface tension of liquid like water, M is a molecular weight of water, ρ is a density of liquid like water, R is gas constant, T is absolute temperature, and rs is a curvature radius of interface.

According to Equation (1), the relative humidity in mortar with SRA becomes higher because SRA can reduce the surface tension of pore solution. Figure 6 shows the time dependency of internal relative humidity of mortars containing SRAs until the age of 28 days. Figure 7 shows the relationship between the surface tension and the internal relative humidity of mortar containing SRAs at 7 days. There is no clear effect of SRA on the internal relative humidity in spite of W/Cs. This implies that there is other influencing factor on autogenous and drying shrinkage except for the surface tension of pore solution.

55

60

65

70

75

80

85

90

95

100

0 7 14 21 28

Age(day)

R.H

.(%)

30HU730HU630KC430KC530KC5'30KC630MOE

55

60

65

70

75

80

85

90

95

100

0 7 14 21 28

Age(day)

R.H

.(%)

50N50HU750HU650KC450KC550KC5'50KC650MOE

(a) W/C=30%

Fig. 5 E ffects of SP and SRA on ela t ive humidity of mor tar for 28 days

(b) W/C=50%

90

92

94

96

98

100

35 40 45 50 55 60 65 70 75

Surface Tention(dyn/cm)

R.H

.(%)

30HU7 30HU6

30KC4 30KC5

30KC5' 30KC6

30MOE

78

80

82

84

86

88

90

92

35 40 45 50 55 60 65 70 75

Surface Tention(dyn/cm)

R.H

.(%)

50N 50HU7 50HU6

50KC4 50KC5 50KC5'

50KC6 50MOE

(a) W/C=30% (b) W/C=50%

Fig. 6 Relat ion between sur face tension and rela t ive humidity of mor tar added SP and SRA at

Figure 6. Effects of SP and SRA on relative humidity of mortar for 28

Figure 7. Relation between surface tension and relative humidity of mortar added SP and SRA at age

26 CMT - 216


CMT -

3.4 Discussion on shrinkage-reducing mechanism of SRAs From Laplace equation (1), the curvature radii of pore in equilibrium can be calculated with various humidity levels. For example, at the surface tension of water, 75 dyn/cm2 the pore radii corresponding to the relative humilities of 90%, 80%, and 60% are approximately 10, 5, and 2 nm, respectively. On the other hand, the side chain lengths of SRAs used in this study after they are adsorbed on cement particles are around 4 nm for KC4, KC5, and KC6 and 2 nm for MOE when their chain structures are stretched. It is therefore presumed that with a relative humidity of around 80% or less, the size of pores filled with water may be smaller than that of the stretched chains of SRAs. This requests that the autogenous shrinkage and drying shrinkage are dependent on other factors such as interaction between polymers and between polymer and pore walls, as well as the surface tension of pore water. In this study, surface tension showed strong effects on the autogenous shrinkage with a high relative humidity, but no clear effect on the drying shrinkage with a low relative humidity, especially at the W/C of 30%. This agrees with the above-mentioned working hypothesis. From these discussions, it is concluded that the interaction between the absorbed polymers might influence the shrinkage, in particular the drying shrinkage.

4. CONCLUSIONS

In this study, the effect of types of polymer on the autogenous shrinkage and drying shrinkage of mortar was investigated using four different kinds of synthesized polymer to develop the shrinkage reducing agent. The shrinkage study indicated there is the good relation between surface tension and autogenous shrinkage, but the relation varied depending on t`e type of polymer. On the other hands, for drying shrinkage, no clear relationship was obtained between the surface tension and the shrinkage when the W/C was as low as 30%. Newly synthesized polymer did not retard significantly the cement hydration. The strength of cement mixes depends on the addition of polymer. The ratio of hydrophobic groups in polymer molecule might have an influence on the interparticle forces of cement particles, and this can be attributed to the strength development. Comparing the pore radius in equilibrium to the size of polymer, it could be concluded that the autogenous shrinkage and drying shrinkage are dependent on other factors such as interaction between polymers and between polymer and pore walls, as well as the surface tension of pore water.

5. REFERENCES

Wittmann, F. H. (1976). “On the Action of Capillary Pressure in Fresh Concrete.” Cement and ConcreteResearch, Vol. 6, pp.49-56.

Tomita R. et al. (1983). “Drying Shrinkage of Concrete Using Cement Shrinkage Reducing Agent.” Cement Associationof Japan (CAJ) Review, pp.198-199.

Shah, S.P et al. (1992). “Effects of Shrinkage-Reducing Admixtures on Restrained Shrinkage Cracking

26 CMT - 217


CMT -

of Concrete.” ACI Material Journal, Vol.89, N0.2, pp.289-295. Berke, N. S. et al. (1997). “New Developments in Shrinkage-Reducing Admixtures.” Proceedings of

Fifth CANMET/ACI International Conference on Superplasticizer and Other Chemical Admixtures in Concrete, Rome, Italy, ACI SP-173, pp.971-998.

Nami, C.K. et al. (1998), “Shrinkage-Reducing Admixture.” Concrete International, Vol.20, No.4, pp31-37.

Nawa, T., Horita, T. and Ohnuma, H. (2002). “A Study on Measurement System for Autogenous Shrinkage of Cement Mixes.” Edited by R. K. Dhir, M. D. Newlands and T. A. Harrison, Thomas Telford, pp.281-290.

Nawa, T. and Horita, T. (2005). “A Mechanism of Autogeneous Shrinkage of Cementitious Materials.” Proceedings of International Congress ‘Global construction: ultimate concrete opportunities’, Dundee, pp.425-434.

26 CMT - 218


CMT -

DEVELOPMENT OF CREEP AND SHRINKAGE PREDICTION FOR MALAYSIAN CONCRETE

Wahid Omar1, Tan Pui Lai2, Khoo Hui Kiang3 and Roslina Omar4

ABSTRACT: This paper presents the study on development of creep and shrinkage prediction for normal strength concrete in Malaysia. The current design estimation practiced in the local industry for creep and shrinkage estimation is mainly based on BS 8110, or any other foreign standard codes available. It is a well known fact that creep and shrinkage is influenced by various factors, namely the constituent materials, age at loading, ambient temperature and relative humidity of the surrounding. Therefore suitability and accuracy of the prediction models developed in temperate countries for tropical concrete in Malaysia is questionable. In this study, laboratory testing on creep and shrinkage of normal strength concrete comprising of G20, G30 and G40 were conducted to determine the local time-dependent deformation. Five existing creep and shrinkage prediction models were then assessed as it pertains to the experimental results obtained from the laboratory testing to determine the accuracy and precision of each model. The five models assessed were from the Eurocode 2, ACI 209 Code Model, Bazant B3 Model, CEB-FIP 1990 Code Model, and the Australian Standard 3600. Based on statistical analysis conducted, AS 3600 code model is found to provide the best prediction of creep for local concrete. As for shrinkage, B3 model is found to offer the best prediction. With the experimental result and statistical analysis, creep and shrinkage modification factors for application on Malaysian concrete is proposed. The CEB-FIP 1990 Model is preferred to the AS 3600 model code because AS 3600 predicts creep by interpreting graphs rather than formulas, thus compromising the precision of creep estimation value. Therefore the creep modification factor is proposed to CEB-FIP 1990 Model whereas the shrinkage modification factor is proposed to B3 Model.

KEYWORDS: Creep, shrinkage, normal strength concrete, tropical climate, prediction models

1. INTRODUCTION

Creep is defined as the time-dependent deformation resulting from sustained load. Generally, creep is divided into basic and drying. Creep that occurs under the influence of moisture loss is referred as drying creep whereas basic creep occurs under the condition where there is no moisture loss. Shrinkage on the other hand is the time-dependent deformation that occurs in the absence of applied load. It is caused by loss of water mainly due to evaporation, hydration of cement and carbonation. Creep and shrinkage deformation are crucial for durability and serviceability analysis of prestressed members, high rise buildings and long span members. They contribute to the increase in deflection and curvature of beams, cracking, loss of prestress and redistribution of stresses in structures. Often, damaged structures are either shut down or undergo extensive repairs long before the end of their intended design life, resulting in significant economic consequences.

Due to the detrimental effects on structures, these time-dependent deformation has been the subject of much research during recent years. It includes both experimental works to gain insight into the physical phenomena as well as mathematical modeling. Even though the significant effect of temperature and surrounding moisture on creep rate is a well known fact, little research is conducted for concrete in the

1 Deputy Dean and Associate Professor, Faculty of Civil Engineering, Universiti Teknologi Malaysia, Malaysia. 2 Postgraduate Research Student, Faculty of Civil Engineering , Universiti Teknologi Malaysia, Malaysia. 3 MEng, Faculty of Civil Engineering , Universiti Teknologi Malaysia, Malaysia. 4 Postgraduate Research Student, Faculty of Civil Engineering, Universiti Teknologi Malaysia, Malaysia.

27 CMT - 219


CMT -

tropical countries. Most of the research was conducted on concrete under the influence of temperate climate and prediction models were empirically derived using curve fitting based on the test results.

In practice, local engineers predict this time-dependent deformation by referring to recommendations in standard codes or any other prediction models available. These models were mainly developed in temperate countries whereas creep and shrinkage of concrete in tropical countries such as Malaysia is expected to have different magnitudes. The degree of differences of values between the prediction models to the actual deformation under tropical climate is never significantly verified. Hence, it is difficult for local design engineers to predict the time-dependent deformation with confidence.

The main objective of this study is to propose modification factors to the creep and shrinkage prediction for the application on local normal strength concrete (NSC). Besides the modification factors, this paper also presents the creep and shrinkage laboratory testing results for concrete G20, G30 and G40. Comparison of the experimental data to prediction values by five existing prediction models were carried out in order to determine the accuracy of each model for local concrete. The five models assessed were from the Eurocode 2, ACI 209 Code Model, Bazant B3 Model, CEB-FIP 1990 Code Model, and the Australian Standard 3600.

2. LABORATORY TESTING

2.1 Materials and Mix Proportions The concrete tested in this study consists of NSC of G20, G30 and G40. The mix proportions of these mixes are presented in Table 1. The concrete are cast using graded 20mm crushed granite as the coarse aggregate and river sand as the fine aggregate.

Table 1. Materials and mix proportions (Omar et al., 2006)

Unit Weight (kg/m3) Concrete Grade w/c OPC Coarse Fines Water G20 0.68 302 1085 818 205 G30 0.57 360 1052 793 205 G40 0.49 418 1036 750 205

2.2 Testing Procedures The laboratory work was conducted in the Materials and Structures Laboratory in Universiti Teknologi Malaysia. The experimental work consists of concrete compressive strength test, elastic modulus, creep and shrinkage testing. All the testing specimens were moist cured using wet burlap for a duration of 7 or 28 days, depending on the age of creep and shrinkage testing. The compressive strength test was conducted on both 150mm cubes and 150mm x 300mm cylinders cast from the same batch of concrete, cured under the same conditions and tested at the same age for comparison purpose.

Creep and shrinkage were tested on 100mmx300mm cylinders and 100x100x500mm prisms, respectively. The creep test was carried out according to the standard method specified by ASTM C512-87 whereas shrinkage test was according to ASTM C157-92. Both tests were carried out in control room under controlled environment with temperature of 27 ± 2oC and relative humidity (RH) of 50 ± 4%. For each set of creep testing, a total of twelve specimens were cast with three specimens each for elastic modulus, compressive strength and creep testing. The remaining three specimens are kept unloaded as control. The measurement of total deformation after loading was performed using a mechanical Demec gauge of 200mm gauge length at four circumferential positions of the specimens.

27 CMT - 220


CMT -

The shrinkage was measured on partially embedded gauge studs fixed on the centre point of the two ends to obtain the axial deformation of specimens (Omar et al., 2006). 2.3 Analysis Methods Three steps of analysis are involved in order to determine the best prediction model and to propose modification factors for tropical concrete. The first step consists of evaluation of experimental data and comparing the experimental results to the prediction values of the five models studied. Comparison of the experimental data and model prediction were conducted through Residuals Analysis. Residual is the difference between the experimental and model prediction values. It identifies over-prediction or under-prediction of a particular model based on experimental results. Positive value of mean residual obtained indicates overprediction of a model whereas negative value indicates underprediction. Table 3 and 5 present the summary of Residuals Analysis results for creep and shrinkage, respectively. The result of this analysis however is limited to identifying if the prediction model is conservative or otherwise and does not distinguish the best prediction model.

Therefore the second step of analysis involves the ranking of prediction model. Error Percentage Method and Residuals Squared analysis were used to determine the best prediction model for local concrete. The Error Percentage is obtained based on Equation (1) with the smallest error percentage indicating the best fit model. The Residuals Squared on the other hand, is obtained based on the summation of the residuals squared as shown in Equation (2). The model with the smallest value indicates the best prediction model. Combination of these two methods gives an overall ranking of the five prediction models.

Error Percentage = 100Re xValuealExperiment

sidual

(1)

Residuals Squared = [ ]∑j

iisidual 2)Re

(2) In step three, modification factors for local concrete is developed. Using correlation method, modification factors are proposed to the best fit model for a more accurate creep and shrinkage prediction. The recommended values from the best prediction model are plotted against the experimental results and the modification factors are taken as the gradient of the best fit line for the data points.


3.1 Concrete Properties The concrete compressive strength at 7 and 28 days are presented in Table 2. Besides compressive strength results, the modulus of elasticity of concrete at 28 days and the concrete density are also presented. Based on the results presented, it is observed that a consistent strength gain from 7 day to 28 day for both cube and cylinder is achieved.

Table 2. Concrete compressive strength and modulus of elasticity (Omar et al., 2006)

Compressive Strength (N/mm2)

Cube Cylinder Concrete Grade 7 day 28 day 7 day 28 day

Elastic Modulus (kN/mm2)

Density (kg/m3)

G20 23.39 25.56 18.69 22.44 29.9 2368 G30 31.53 36.21 29.33 31.33 31.8 2398 G40 34.39 40.62 28.80 35.48 32.4 2411

27 CMT - 221


CMT -

3.2 Creep Analysis The creep strains measured from experimental works were expressed in terms of creep coefficient for analysis purposes. Creep coefficient is defined as creep strain as a fraction of the elastic strain. Figure 1 and 2 presents the results of creep coefficient tested under controlled condition for age at loading of 7 days and 28 days, respectively. With reference to Figure 1 and 2, it is observed that the creep coefficient is lower as concrete strength increases. This pattern can be explained as higher strength concrete has lower water-cement ratio in the mix design. Thus fewer pores exist in the mature cement and this subsequently increases the rigidity of the solid matrix, decreasing the creep deformation (Smadi, 1987). The creep result obtained is in agreement with the trends obtained by other work conducted on concrete time-dependent deformation (Marzouk, 1991).

The actual creep coefficients were compared to the prediction models through Residuals Analysis. The results of the analysis for creep are presented in Table 3. The best prediction model for creep of local concrete is then obtained through Error Percentage and Residuals Squared analysis. Results of the analysis are presented in Table 4. Based on the analysis, it is found that the AS 3600 Code Model provides the best prediction for creep of Malaysian concrete. This is due to the fact that AS 3600 Code Model provides for concrete under tropical and near coastal climatic condition, which is the nearest to Malaysian environment. However, AS 3600 model does not provide correction factors in its equation and the predicted values are obtained using graphs. Therefore CEB-FIP 1990 Model Code is preferred for correlation analysis in order to propose for modification factors.

0.00

1.00

2.00

3.00

4.00

5.00

6.00

0 50 100 150 200

Time after loading (days)

Cre

ep C

oeffi

cien

t

G20-7daysG30-7days

G40-7days

Figure 1. Creep coefficient of concrete G20, G30 and G40, loaded at 7 days

0.00

1.00

2.00

3.00

4.00

5.00

6.00

0 50 100 150 200 250Time after loading (days)

Cre

ep C

oeff

icie

nt

G20-28daysG30-28daysG40-28days

Figure 2. Creep coefficient of concrete G20, G30 and G40, loaded at 28 days

27 CMT - 222


CMT -

Table 3. Creep coefficient mean residual for specimens loaded at 7 and 28 days (Khoo, 2006)

* UP = Under Predict; OP = Over Predict

Table 4. Overall creep coefficient prediction models ranking (Khoo, 2006)

Prediction Model EC2 ACI209 CEB-FIP B3 AS3600 Age at

loading Concrete

grade error (%) R2

error (%) R2

error (%) R2

error (%) R2

error (%) R2

G20 3 4 5 5 2 2 4 3 1 1 G30 2 2 4 4 1 1 5 5 3 3

7 da

ys

G40 2 2 4 1 3 3 5 5 1 4 G20 4 4 5 5 3 3 1 1 2 2 G30 4 4 5 5 3 3 2 2 1 1

28 d

ays

G40 2 1 4 4 1 2 5 5 3 3 Sum 34 51 27 43 25

Ranking 3 5 2 4 1 3.3 Shrinkage Analysis As for shrinkage strain, the experimental results for the age of drying at 7 and 28 days are presented in Figure 3 and 4, respectively. Similar to the creep strain pattern, shrinkage is observed to be lower for higher concrete strength. The reduction in shrinkage strain as the strength of concrete increase is because of the finer pore structure as well as the reduced volume of evaporable water in the concrete.Comparison of the measured shrinkage to the models prediction conducted through Residuals Analysis produces results as given in Table 5. Based on the Error Percentage and Residuals Squared analysis, the B3 Model was found to be the best prediction model for shrinkage. It was followed by CEB-FIP 90, EC 2, ACI-209 and AS 3600. The result of the best prediction model ranking is shown in Table 6.

0

100

200

300

400

500

600

0 50 100 150 200Time after drying (days)

Shrin

kage

Str

ain

(mic

ron)

G20-7days

G30-7days

G40-7days

Figure 3. Shrinkage strain of G20, G30 and G40 concrete, with age at drying of 7 days

age of loading = 7 days age of loading = 28 days G20 G30 G40 G20 G30 G40

Prediction Models

mean remark mean remark mean remark mean remark mean remark mean remark EC2 -0.541 UP -0.063 UP 0.300 OP -1.475 UP -0.347 UP 0.085 OP ACI209 -1.095 UP -0.323 UP 0.358 OP -1.697 UP -0.372 UP 0.339 OP CEB-FIP -0.522 UP -0.047 UP 0.315 OP -1.460 UP -0.335 UP 0.097 OP B3 0.554 OP 0.805 OP 1.308 OP -0.627 UP 0.145 OP 0.742 OP AS3600 -0.206 UP -0.292 UP 0.026 OP -0.796 UP -0.093 UP 0.677 OP

27 CMT - 223


CMT -

050

100150200250300350400450500

0 50 100 150 200Time after drying (days)

Shrin

kage

Str

ain

(mic

ron)

G20-28days

G30-28days

G40-28days

Figure 4. Shrinkage strain of G20, G30 and G40 concrete, with age at drying of 28 days

Table 5. Creep coefficient mean residual for specimens loaded at 7 and 28 days (Khoo, 2006)

age of loading = 7 days age of loading = 28 days G20 G30 G40 G20 G30 G40

Prediction Models

mean remark mean remark mean remark mean remark mean remark mean remark EC2 64 OP 104 OP 87 OP 89 OP 47 OP 132 OP ACI209 106 OP 163 OP 203 OP 130 OP 93 OP 224 OP CEB-FIP 90 -104 UP -40 UP -53 UP -123 UP -96 UP -6 UP B3 -43 UP 44 OP 23 OP -18 UP -22 UP 69 OP AS3600 81 OP 171 OP 177 OP 105 OP 110 OP 222 OP * UP = Under Predict; OP = Over Predict

Table 6. Overall shrinkage prediction models ranking (Khoo, 2006)

Prediction Model Ranking

EC2 ACI209 CEB-FIP90 B3 AS3600

age at loading

concrete grade

error (%) R2

error (%) R2

error (%) R2

error (%) R2

error (%) R2

G20 3 3 5 4 2 6 1 1 4 5 G30 3 3 4 4 1 1 2 2 5 5

7 da

ys

G40 3 3 5 5 2 2 1 1 4 4 G20 3 2 4 4 2 5 1 1 5 3 G30 2 2 4 4 3 3 1 1 5 5

28 d

ays

G40 3 3 4 5 1 1 2 2 5 4 Sum 33 52 29 16 54

Ranking 3 4 2 1 5 3.4 Modification Factors for Creep After the best creep prediction model for local concrete is identified, modification factor to better predict tropical concrete is determined based on experimental results. Figure 5 presents an example of the correlation of creep coefficient for CEB-FIP 90 model and the experimental result for concrete G30 with age at loading for both 7 and 28 days. The gradient of the linear function obtained is suggested as the modification factor, α. A modification factor is introduced to each concrete grade, as shown in Table 7. Therefore, the final creep coefficient, φ (t,to) for the modified CEB-FIP 1990 Model Code is given as follows:

27 CMT - 224


CMT -

( ) ( ) ( )ocoo tttt ,**, βφαφ = (3) where, α : Modification factor for creep coefficient φo : Notional creep coefficient βc(t,to) : Coefficient describing creep development with time after loading

y = 1.0179xR2 = 0.9978

y = 1.1859xR2 = 0.9752

0

1

2

3

4

0 1 2 3 4CEB-FIP 90 creep coefficient

Expe

rimen

tal c

reep

coe

ffic

ient

G30 (7 days)

G30 (28 days)

Figure 5. Linear correlation of experimental creep data and CEB-FIP 90 prediction for G30 concrete

Table 7. Creep coefficient multiplication factors, α for CEB-FIP 1990 Model Code

Age at loading Concrete Grade

7 days 28 days G20 1.1677 1.5962 G30 1.0179 1.1859 G40 0.8691 0.9617

In order to evaluate the level of accuracy of the modified CEB-FIP 90 Model to the original CEB-FIP 90 model, the Residuals Squared and Error Percentage analysis were conducted. After the modification factor is introduced to the CEB-FIP 90 Model, the residual squared for concrete G20, G30 and G40, loaded at 7 days dropped from 51.38 to 0.60, 0.44 to 0.11 and 21.75 to 1.36, respectively. As for the age of loading at 28 days, the residual squared value for concrete G20, G30 and G40 dropped from 396.56 to 0.49, 17.21 to 1.06 and 3.91 to 2.78, respectively (Khoo, 2006). In addition to that, the range of error also reduced significant when the modified CEB-FIP 90 model was used to predict creep coefficient for Malaysian concrete. For concrete with age at loading of 7 days, the error percentage of the modified value dropped from 14.09% to 1.44% for G20, 1.89% to 1.01% for G30 and 13.61% to 3.61% for G40 concrete. At the age of loading of 28 days, the error percentage dropped from 37.65% to 1.23% for G20, 13.92% to 4.38% for G30 and 8.53% to 6.99% for G40 concrete. The range of error of the modified CEB-FIP 90 Model Code is within 2.4% to 20.0%. These statistics show that the new modified prediction values are very close to the experimental results. Most of the new prediction figures fall into 95% confidence of the experimental data.

3.5 Modification Factors for Shrinkage

An example of the correlation between experimental results to prediction by B3 to obtain modification factor for shrinkage strain for concrete G30 is shown in Figure 6. The slope of the straight line is suggested as the modification factor, which is summarized in Table 8.

Therefore, the final shrinkage equation for the modified B3 model is given as follows:

27 CMT - 225


CMT -

)(***),( tSktt hshosh ∞= εβε (4)

where, β : Modification factor ε∞ : Ultimate shrinkage strain kh : Cross section shape factor S(t) : Time function for shrinkage

y = 1.0618xR2 = 0.9851

y = 0.8607xR2 = 0.9959

0

100

200

300

400

500

0 100 200 300 400 500B3 shrinkage strain

expe

rimen

tal s

hrin

kage

str

ain

G30 (7 days)

G30 (28 days)

Figure 6. Linear correlation of experimental shrinkage data and B3 prediction for G30 concrete

Table 8. Shrinkage multiplication factor, β for B3 Model

Age of drying Concrete grade 7 days 28 days G20 1.1368 1.0709 G30 0.8607 1.0618 G40 0.9171 0.7816

After the modification factor, β is applied to B3 model, the shrinkage strains predicted by the modified model is again analysed using Residuals Squared and Error Percentage method. At the age of drying of 7 days, the residual squared of B3 Model shrinkage with inclusion of modification factor for concrete G20, G30 and G40 dropped by 86.9%, 98.7% and 78.6%, respectively as compared to before the modification. As for the age of drying at 28 days, the residual squared values dropped by 46.3%, 75.1% and 95.2% for each G20, G30 and G40 concrete.

The error percentage for shrinkage of the modified B3 value dropped from 9.7% to 5.1% for concrete G20, 15.3% to 2.3% for G30 and 8.9% to 5.4% for G40. At the age of drying of 28 days, the error percentage for G30 concrete dropped from 8.30% to 3.9% and from 39.5% to 12.1% for G40 concrete. In contrary to the statistical pattern obtained, a slight increase is observed for concrete G20. The increment from 8.3% to 8.9% however is not significant. The range of error for the modified B3 Model is within 10.5 to 23.9%. Most of the new prediction values fall within the 95% confidence of the experimental data.

4. CONCLUSIONS

Based on the study conducted, the following conclusions can be drawn:

i. The creep prediction by AS 3600 was found to be the best prediction for local creep. However, the prediction values are obtained through graph reading in which the results obtained is subjected to error. Therefore the CEB-FIP 1990 Model Code was chosen to be modified for application of local concrete instead.

27 CMT - 226


CMT -

ii. The B3 Model was found to give the best prediction for shrinkage in local concrete and modification factors were proposed to B3 Model for better prediction of local concrete.

iii. The modified CEB-FIP 1990 Model Code is recommended for creep prediction of tropical concrete and the modified B3 Model is recommended for shrinkage prediction. This is due to the substantial reduction in the statistical values with the inclusion of modification factor, representing a better prediction model for local concrete strain.

5. REFERENCES

ACI Committee 209 (1971). “Prediction of creep, shrinkage and temperature effects in concrete structures.” SP-277, ACI, Detroit.

Annual Book of ASTM (1992). “Standard test method for creep of concrete in compression.” Detroit: ASTM C 512-87.

Annual Book of ASTM (1992). “Standard test method for length change of hardened hydraulic-cement mortar and concrete.” Detroit: ASTM C 157-91.

Australian Standard (2001). “Concrete Structures.” Sydney: AS 3600 – 2001. Bazant, Z. P. and Baweja, S. (2000). “Creep and shrinkage prediction model for analysis and design

for concrete structures: Model B3.” American Concrete Institute, Farmington Hills, Michigan. Comite Euro-International du Beton (1993). “CEB-FIP Model Code 1990”. Thomas Telford Service

Ltd. London. European Committee for Standardization (2002). “Eurocode 2: Design of concrete structures – Part 1:

General rules and rules for buildings.” Brussels. prEN 1992-1-1. Khoo, H. K. (2006). “Development of creep and shrinkage prediction model for Malaysian normal

strength concrete.” Universiti Teknologi Malaysia: Master of Engineering Thesis. Marzouk, H. (1991) “Creep of high strength concrete and normal strength concrete.” Magazine of

Concrete Research, 43, No.155. pp. 121 – 126. Omar, W., Tan, P. L. and Omar, R. (2006). “A study on creep and shrinkage deformation of Malaysian

concrete.” NRMCA International Concrete Convention 2006, Kuala Lumpur. Smadi, M.M., Slate, F.O. and Nilson, A. H. (1987). “Shrinkage and creep of high, medium and low

strength concretes, including overloads.” ACI Mat. J., May-June, v.84, n.3, pp. 224-234.

27 CMT - 227


CMT -

PREDICTION OF THERMAL PROPERTIES OF CONCRETE

P. Choktaweekarn1 and S. Tangtermsirikul2

ABSTRACT : Thermal properties of concrete are the important factors for calculating thermal stress in mass concrete. In this study, some thermal properties such as specific heat, thermal conductivity and thermal expansion coefficient (CTE) at various ages were studied. Pastes were prepared to study effect of water and fly ash content. Mortars were prepared to study effect of aggregate type and aggregate content and concrete mixtures were prepared for studying effect of paste and water content. It was found that specific heat depended largely on the amount of free water content. It reduced with decrease in the amount of free water content then specific heat reduced with age. The use of fly ash tended to reduce the specific heat in long term due to pozzolanic reaction. The specific heat of mortars were lower than that of cement pastes. In case of CTE, the mixtures with aggregate had lower CTE than the neat cement paste. Fly ash reduced CTE especially at early age. Water to binder ratio and paste content in the tested ranges have little effect on CTE. CTE of concrete slightly increases with age. It was found that the thermal conductivity of all samples decreased with age. Models for predicting thermal properties of paste and concrete were proposed as time and mix proportion dependent functions. The models were verified with various experimental results and the verification results were satisfactory. KEYWORDS: Thermal conductivity, Specific heat, Thermal expansion coefficient

1. INTRODUCTION

In massive concrete structures such as dams and mat foundations, temperature gradients occur inside the structures due to heat of hydration. To avoid cracking due to heat of hydration, one approach is to control the hydration heat of concrete by reducing cement content in the concrete mixture since the components of cement are responsible for the generation of heat. This reduction in cement content can be achieved by the proper use of a good pozzolanic material such as fly ash to replace cement. To predict the thermal cracking of concrete, the quantitative evaluation of heat evolution during hardening as well as the thermal properties and related mechanical properties at early age are necessary to be investigated. Thermal properties of concrete depend on thermal properties of ingredients in concrete such as cementitious materials, water, air and aggregates and their proportion as well as their time-dependent change. Table 1 shows the thermal properties of each ingredients of concrete. ACI Committee 211 (1994) reported that thermal conductivity varies with density of concrete. Heavier aggregate results in higher thermal conductivity. Specific heat of concrete is slightly affected by the mineralogical character of the aggregate, but is considerably increased by an increase in the moisture content of the concrete (Neville, 1995). The coefficient of thermal expansion of concrete is a result of coefficient of cement paste and aggregate since they are the main constituents of the concrete. In this study, specific heat and thermal conductivity of paste and mortar and coefficient of thermal expansion of paste, mortar and concrete were studied and predicted.

2. EXPERIMENTAL PROGRAM

2.1 Mix Proportion and Materials

The mix proportions of the tested paste and mortar are shown in Table 2. The mix proportions of the tested concrete are shown in Table 3. Chemical compositions and physical properties of the cement and fly ash used in the tests are given in Table 4. Mixes 1 to 8 and 10 of pastes and mortars were used in the study of specific heat. Mixes 1 and 4 of pastes were used in the study of thermal conductivity. In the study of CTE, pastes, mortars and concrete were produced and tested. Mixes 1 to 4 and 7 to 13 1 Doctoral Student, Sirindhorn International Institute of Technology, Thammasat University, Pathumthani, Thailand 2 Prof. Dr, Sirindhorn International Institute of Technology, Thammasat University, Pathumthani, Thailand

28 CMT - 228


CMT -

of pastes and mortars were used in the CTE experiment. Mixes 14 to 17 of concrete were used in CTE experiment. Limestone was used as coarse aggregate and natural river sand was used as fine aggregate. The ratio of water to binder (w/b) and replacement ratio of cement by fly ash f/(c+f)) were varied. The tests were conducted at various ages (1, 3, 7, and 28 days). The specimens were wrapped using aluminum foil immediately after casting in order to prevent the evaporation of water and to simulate the physical condition of the specimens to be similar to that inside the mass concrete (no moisture loss or gain). The specimens were kept in the room temperature (30 + 2 °C) and in seal-curing condition until the test date.

Table 1. Thermal coefficients of the ingredients of concrete

Thermal Coefficients Limestone Quartz Sand

Water Cement Fly Ash Hydrated Product*

Specific Heat (Kcal/Kg°C) (ASHRAE, 1997)

0.23 0.21 1.00 0.18 0.17 0.15

Heat Conductivities (Kcal/m.day.°C) (ASHRAE, 1997)

20.50 7.50 12.44 0.62 1.16 23.5

Thermal Expansion Coefficient (micron/°C)

10.4

(Klieger,

1994)

.4.5

(Klieger,

1994)

- 14.4

(super civil

CD, 2005)

6.45

(Mangutova1,

2004)

20.0

Stiffness (MPa) x 104 (Neekhra, 2004)

14.27 8.64

* The value was obtained from back analysis

Table 2. Mix proportions of the tested pastes and mortars.

Mix No. Mixture Type w/b f/(c+f) s/(c+f) g/(c+f) 1 w25r0 0.25 0 0 0 2 w25r3 0.25 0.3 0 0 3 w25r5 0.25 0.5 0 0 4 w40r0 0.40 0 0 0 5 w40r3 0.40 0.3 0 0 6 w40r5 0.40 0.5 0 0 7 w40s1 0.40 0 1 0 8 w40s3 0.40 0 3 0 9 w40g1 0.40 0 0 1

10 w40g3 0.40 0 0 3 11 w35r0 0.35 0 0 0 12 w35r3 0.35 0.3 0 0 13 w35r5 0.35 0.5 0 0

Remarks: w: water, f: fly ash, c: cement, s: sand and g: crushed limestone sand.

Table 3. Mix proportions of the tested concrete. Mix No. Mixture Code γ w/b f/(c+f)

14 γ1.2w4 1.2 0.40 0 15 γ1.4w4 1.4 0.40 0 16 γ1.2w5 1.2 0.50 0

17 γ1.4w5 1.4 0.50 0 Remarks: γ : the ratio of the volume of paste to volume of void in aggregate phase. River sand and limestone are used as fine and coarse aggregate, respectively.

28 CMT - 229


CMT -

Container

Specimen

Insulating material

Hot water

Thermocouples

Data logger 1.27 cm 0.5 cm

1st specimen 2nd specimen

Table 4. Chemical composition and physical properties of Portland cement type I and fly ash.

Chemical Compositions Cement Type I Fly Ash (% by weight) SiO2 20.99 45.88 Al2O3 5.18 26.20 Fe2O3 3.20 10.94 CaO 64.63 8.28 MgO 1.30 2.83 SO3 2.61 1.04 Na2O 0.04 0.90 K2O 0.40 2.78 TiO2 0.25 0.51 P2O5 0.05 0.10 LOI 1.17 0.17 Free Lime 0.75 0.18 Specific Gravity 3.15 1.85 Blaine Fineness (cm2/g) 3190 3460

2.2 Specimen Preparation and Test Procedure

2.2.1 Specific Heat

All samples were cast in PVC pipes having 1-inch in diameter and 2- inch in length. Two specimens were cast for each type of mixture. For the first one, thermocouple was placed at the center of the specimen while for the second specimen, thermocouple was placed near the surface (0.5 cm. from the surface of the specimens). The positions of the thermocouple are shown in Figure 1a. This effort was firstly for taking into account the non-uniform temperature that might occur within the specimen. However, it was found later from the test that the differences were negligible. The specific heat, obtained from the tests, was then the average value of these two specimens. The apparatus and its setting for testing specific heat are shown in Figure 1b. (a) (b)

Figure 1. (a) The positions of the thermocouples, (b) Schematic of apparatus for testing the specific heat values

In the tests, the specimens were submerged in hot water in an insulated container. Temperature of the hot water and the specimens were recorded, using data logger, at every 30 seconds. The calculations of specific heat are shown in Eq. (1) to Eq. (3).

splossw QQQ =− (1)

spspsplosswwwww ΔTcmΔTcmΔTcm =− (2)

( )spsp

losswwwsp ΔTm

ΔTΔTcmc −= (3)

28 CMT - 230


CMT -

where Qw is the heat reduced in water (kcal), Qloss is the heat loss from water to the environment (kcal), Qsp is the heat intake in specimens (kcal). mw and msp are the mass of water and specimen, respectively (kg). cw and csp are the specific heat of water and specimen, respectively (kcal/kg/°C). ΔTw is the temperature reduction of water (°C), ΔTsp is the temperature increase of specimen (°C), and ΔTloss is the temperature reduction of water due to loss of heat to the environment. In order to measure temperature loss (ΔTloss), the system with water only was used to record the temperature change of the water due to loss of heat into the environment. The temperature was also recorded by using data logger at every 30 seconds. This calibration was operated for 3 cycles and the average value was used.

2.2.2 Thermal Conductivity

The apparatus for testing the thermal conductivity is shown in Figure 2. The test method of this equipment is based on British standard (BS 874), which is called steady state method. The size of the specimens is 30x30x5 cm.

2.2.3 Coefficient of Thermal Expansion

Prism specimens with dimension of 25x25x285 mm were used for cement-fly ash pastes and mortars and those with dimension of 75x75x285 mm were used for concrete. Two specimens were tested at 1, 3, 7 and 28 days of age. Thermocouple was placed at the center of each specimen to measure the specimen temperature. The temperature range used in the experiment was adopted from TI-B 101 (Danish Technological Institute Building Technology, 1994). The temperature range recommended in the standard TI-B 101 is 5 ºC to 30 ºC. The specimens were tested by cooling them down in the refrigerator to reduce temperature of the specimens from room temperature (about 30 + 2 ºC) to 10 ºC, then moving them out of the refrigerator for heating up to room temperature. The steps of cooling down and heating up are shown in Figure 1. It has been reported that for a given concrete mixture, the magnitude of thermal expansion or contraction of concrete in normal range of temperature, including the range occurred in mass concrete, is the same for each unit temperature change. In other word, CTE is constant in that range (Klieger, 1994).

For every 5 ºC change of temperature, the length change was measured using the length comparator. The thermal expansion coefficient is calculated from Eq. (4).

Figure 2. Apparatus for Testing Thermal Conductivity

Figure 3. Changes in temperature of the tested specimens

Room Temperature (30 + 2 ºC) 25 ºC 20 ºC 15 ºC 10 ºC

28 CMT - 231


CMT -

ΔΤ=

εCTE (4)

where CTE is the thermal expansion coefficient (micron/ ºC) , ε is the strain due to temperature change and ΔΤ is the temperature change (ºC).

3. THERMAL PROPERTIES MODEL

For simplicity at this stage, all hydrated and pozzolanic reaction products are assumed to have the same thermal properties. Thermal properties of hydrated and pozzolanic products obtained from back analysis as shown in Table1. The effect of air was not considered in the specific heat and CTE models. In case of thermal conductivity model the volume of air remains unchanged during the hydration.

3.1 Specific Heat Model

The specific heat of concrete was assumed to be computed based on the weight fraction of the ingredients including the hydrated and pozzolanic products and their individual specific heat as shown in Eq.(5). As the reaction proceeds, the amount of free water in concrete reduces with an increase in the amount of reacted product and concrete converts from a fresh state to plastic and hardened states. So the specific heat of concrete decreases with age. As coarse aggregate and fine aggregate are inert, their weight fractions and specific heat remain constant throughout the reaction. At the same time the volume of non-reacted cementitious material (e.g. cement and fly ash) and free water reduces and the amount of reacted product increases with time. Considering these factors, the following equations were proposed to determine the specific heat of concrete during the reaction process. The amount of hydrated product is composed of amount of hydrated cement, reacted fly ash, water consumed by hydration and pozzolanic reaction process which can be determined as shown in Eq.(6). The amount of unhydrated cement is calculated based on the average degree of reaction of cement from Eq.(7). The amount of non-reacted fly ash at age t is calculated from Eq.(8). The amount of free water is defined as the volume of water excluding water consumed by the hydration and pozzolanic reaction and can be computed using Eq.(9). The details of free water determination are adopted from Tangtermsirikul et al (2002). The details of degree of hydration are not provided in this paper but elsewhere (Tangtermsirikul, 2002 and Nipatsat, 2000). The specific heat of hydrated product obtained from back analysis was about 0.15 kcal/kg/°C

hphpfaufacucwfwssgg ctwctwctwctwcwcwtc )()()()()( +++++= (5)

))()()((0.1)( twtwtwwwtw ufaucwfreesghp ++++−= (6)

0)100

)(1()( c

hyuc w

ttw

α−= (7)

0)100

)(1()( fa

pozufa w

ttw

α−= (8)

( ) ( ) ( )twtwwtw wgelwhpwwfree −−= 0 (9) where c(t) is the specific heat of concrete at the time considered (kcal/kg/°C). wg, and ws are the weight ratio of gravel, and sand per unit weight of concrete, respectively. wfw(t), wuc(t), wufa(t), and whp(t) are the weight ratio of free water, unhydrated cement, non-reacted fly ash, and the hydrated and pozzolanic products, respectively, at the time considered. cg, cs, cw, cc, cfa, and chp are the values of specific heat of coarse aggregate, sand, water, cement, fly ash, and the hydrated and pozzolanic

28 CMT - 232


CMT -

( ) ( ) ( )( ) ggsspp

gggsssppp

EnEntEnECTEnECTEntEtCTEn

tCTE++

++=

products, respectively (kcal/kg/°C). wc0, wfa0, and ww0 are the weight of cement, fly ash, and water per unit weight of concrete at the time of mixing ( at t = 0 ). αhy(t), and αpoz(t) are the average degree of hydration, and the degree of pozzolanic reaction of paste at the considered age, respectively (%). t is the considered age (day). wwfree(t), wwhp(t), wwgel(t) are the weight ratio at the considered age of free water, water consumed by hydration and pozzolanic reactions and gel water, respectively.

3.2 Thermal Conductivity Model

By using the same concept as mentioned in the specific heat model, Eq. (10) was proposed for estimating the value of heat conductivity of concrete. The heat conductivity of concrete is assumed to be computed based on the volumetric ratio of the ingredients including the hydrated product and their individual heat conductivity. The volumetric fraction of each ingredient in concrete at considered age can be calculate by using the same concept as shown in Eq. (5) to Eq. (8). After conducting back analysis the thermal conductivity of hydrated product was found to be 23.5 kcal/m.day°C where z(t): conductivity of concrete at any age (kcal/m.day°C), zg, zs, zw, zc, zfa, zra, zhp: thermal conductivity of gravel, sand, cement, fly ash, retained air and the hydrated product, respectively (kcal/m.day°C). nfw(t), nuc(t), nufa(t), nhp(t): volumetric ratio of free water, unhydrated cement, unhydrated fly ash and the hydrated product, respectively, at the time considered. ng, ns, nra: volumetric ratio of gravel, sand and air, respectively.

3.3 Thermal Expansion Coefficient Model

Since CTE of concrete is the result of CTE, stiffness and volume fraction of all ingredients of concrete. Aggregate seems to be the main factor that affects CTE of concrete because the aggregate occupies most of the concrete volume. The use of higher CTE aggregate gives higher CTE of concrete. In this study, the existing CTE model for composite material (Autar, 1997) was adopted to calculate CTE of mortar and concrete. The model shown in Eq. (11) was proposed as a function of the volumetric fraction, stiffness and CTE of each ingredient in concrete.

(11)

where CTE(t) is the coefficient of thermal expansion of mortar or concrete at the considered age (micron/°C). CTEp (t), CTEs, and CTEg are the values of coefficient of thermal expansion of paste, fine aggregate and coarse aggregate, respectively (micron/°C). np, is the volumetric ratio of paste Ep(t), Es, and Eg are the stiffness of paste, fine aggregate, and coarse aggregate, respectively (MPa) The stiffness values of fine and coarse aggregates were obtained from a previous study (Neekhra, 2004) as shown in Table 1. The stiffness of paste was adopted from the study of Yomeyama et al (1993). CTE of paste varies according to CTE and volume fraction of non-reacted cementinious materials and hydrated product. Due to the differences of CTE of cement, fly ash and hydrated product, when the amount of each ingredient changes with time according to the reactions, CTE of paste also changes time-dependently. The proposed equation for estimating the value of CTE of pastes is shown in Eq. (12). CTE of hydrated product obtained from back analysis was about 20 micron/°C.

hphpfaufacucp CTEtncCTEtnbCTEtnatCTE )()()()( ×+×+×= (12)

(10) ( ) ( ) ( ) ( ) ( ) hphprarafaufacucwwfreessgg ztnznztnztnztnznzntz ++++++=

28 CMT - 233


CMT -

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0 5 10 15 20 25 30Age of paste (days)

Spec

ific

Hea

t (kc

al/k

g/o C)

.

w/b = 0.25; r = 0 (Test)w/b = 0.25; r = 0.3 (Test)w/b = 0.25; r = 0.5 (Test)w/b = 0.25; r = 0 (Model)w/b = 0.25; r = 0.3 (Model)w/b = 0.25; r = 0.5 (Model)

w25r0 (Tested)w25r3 (Tested)w25r5 (Tested)w25r0 (Model)w25r3 (Model)w25r5 (Model)

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0 5 10 15 20 25 30

Age of Paste (days)

Spec

ific

Hea

t (kc

al/k

g/o C)

. w/b = 0.25; r = 0 (Test)

w/b = 0.40; r = 0 (Test)w/b = 0.25; r = 0 (Model)w/b = 0.40; r = 0 (Model)

w25r0 (Tested)w40r0 (Tested)w25r0 (Model)w40r0 (Model)

where CTEc, CTEfa, and CTEhp are the values of coefficient of thermal expansion of cement, fly ash, and the hydrated and pozzolanic reaction products, respectively (micron/°C). nuc(t), nufa(t), and nhp(t) are the volumetric ratio, at the considered age, of unhydrated cement, non-reacted fly ash, and the hydrated and pozzolanic reaction products, respectively. The constants a, b and c are derived to be equal to 0.284, 1.230 and 1.499, respectively.

4. EXPERIMENTAL RESULTS AND MODEL VERIFICATION

4.1 Specific Heat

From the experimental results, it was found that the specific heat of concrete decreased with a decrease in the amount of free water in concrete. The effect of water to binder ratio is shown in Figure 4a. Pastes with lower free water content (w/b = 0.25) yielded lower specific heat than those of higher free water content (w/b = 0.40). The specific heat of water is much higher than that of cement, so lower w/b gives lower specific heat. The effect of fly ash content is shown in Figure 4b. Specific heat of pastes with fly ash continues decreasing in long term when compared to that of the cement paste. Moreover, specific heat of water is the highest among those of all ingredients of concrete. As the results, the specific heat of fly ash-cement paste has similar time-dependent tendency as the amount of free water content. From the experimental results, it was found that the specific heat of concrete decreased with a decrease in the amount of free water in concrete.

(a) (b)

Figure 4 Comparisons between test and predicted results of specific heat (a) cement paste with w/b = 0.25 and 0.40, (b) cement paste with fly ash replacement ratio of 0, 0.3 and 0.5, and w/b = 0.25

The effects of fine aggregate content are shown in Figure 5. The specific heat of cement paste is higher than that of mortars because of its higher amount of free water content. Moreover, both sand and crushed limestone sand have lower specific heat than water, so the specific heat of mortars are lower than that of the cement paste and the mixtures with more aggregate content yields smaller specific heat. The proposed equations can be nearly quantitatively used to predict the specific heat of the tested pastes and mortars.

4.2 Thermal Conductivity

The experimental results are shown in Figure 6. The air content of the samples w25r0 and w40r0 is equal to 3 and 2.1 percent respectively. It was founded that the thermal conductivity of all samples decreases with the increasing of age. This may be explained that the thermal conductivity of water shown in Table 1 is high then as the hydration proceed the amount of binder decrease so the heat conductivity of pastes decrease. By using Eq.(10), it was found that the model could predict the heat conductivity of the tested pastes within the acceptable range.

28 CMT - 234


CMT -

0

5

10

15

20

25

0 5 10 15 20 25 30

Age of Paste (days)

Ther

mal

Con

duct

ivity

(Kca

l/m.d

ay°C

)

.

w25r0 (Tested)

w25r0 (Model)

0

5

10

15

20

25

0 5 10 15 20 25 30

Age of Paste (days)

Ther

mal

Con

duct

ivity

(Kca

l/m.d

ay°C

)

.

w40r0 (Tested)

w40r0 (Model)

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0 5 10 15 20 25 30Age of Mortar (days)

Spec

ific

Hea

t (kc

al/k

g/o C)

.

w/b = 0.40; s = 0 (Test)w/b = 0.40; s = 1 (Test)w/b = 0.40; s = 3 (Test)w/b = 0.40; s = 0 (Model)w/b = 0.40; s = 1 (Model)w/b = 0.40; s = 3 (Model)

w40r0 (Tested)w40s1 (Tested)w40s3 (Tested)w40r0 (Model)w40s1 (Model)w40s3 (Model)

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0 5 10 15 20 25 30Age of Mortar (days)

Spec

ific

Hea

t (kc

al/k

g/o C)

.

w/b = 0.40; g = 0 (Test)w/b = 0.40; g = 3 (Test)w/b = 0.40; g = 0 (Model)w/b = 0.40; g = 3 (Model)

w40g0 (Tested)w40g3 (Tested)w40r0 (Model)w40g3 (Model)

(a) (b) Figure 5 Comparisons between test and predicted results of specific heat of mortars (a) river sand mortars

with sand to binder ratio of 0, 1 and 3, and w/b = 0.40 (b) crushed limestone sand mortar with sand to binder ratio of 0 and 3, and w/b = 0.40.

(a) (b)

Figure 6 Comparisons between test and predicted results of thermal conductivity (a) cement paste with w/b = 0.25. (b) cement paste with w/b = 0.40.

4.3 Thermal Expansion Coefficient

By using Eq. (12) and CTE of the concrete ingredients shown in Table 1, the verifications with the experimental results of CTE model of paste are shown in Figure 7. It is shown that the proposed CTE model is satisfactory for predicting the CTE of the tested cement-fly ash pastes. The model and the test results show that pastes with higher fly ash replacement ratio (f/(c+f) = 0.5) have lower CTE than those with lower fly ash replacement ratio (f/(c+f) = 0.3) at early age but have higher CTE at later age. The model also shows that CTE of paste is time-dependent and increases with age. By using Eq. (11) and the properties shown in Table 1, the verifications of CTE model with the test results of mortar and concrete are shown in Figures 8 and 9. It is shown in the figures that the proposed CTE model is satisfactory to predict the CTE of the tested mortar and concrete. The model and the test results show that CTE of concrete depends largely on CTE of aggregate and also slightly increases with age. It was found that the prediction was satisfactory especially for concrete the difference is less than 1 micron/°C.

28 CMT - 235


CMT -

02468

10121416182022

0 5 10 15 20 25 30

Age of paste (days)

CTE

(mic

ron/

o C

)

.

w25r0 (Tested)w25r0 (Model)w35r0 (Tested)w35r0 (Model)w40r0 (Tested)w40r0 (Model)

02468

10121416182022

0 5 10 15 20 25 30

Age of paste (days)

CTE

(mic

ron/

o C

)

.

w35r0 (Tested)w35r0 (Model)w35r3 (Tested)w35r3 (Model)w35r5 (Tested)w35r5 (Model)

02468

10121416182022

0 5 10 15 20 25 30

Age of mortar (days)

CTE

(mic

ron/

o C

)

.

w40r0 (Tested) w40r0 (Model) w4s1 (Tested)

w4s1 (Model) w4s3 (Tested) w4s3 (Model)

02468

10121416182022

0 5 10 15 20 25 30

Age of mortar (days)

CTE

(mic

ron/

o C

)

.

w40r0 (Tested) w40r0 (Model) w4g1 (Tested)w4g1 (Model) w4g3 (Tested) w4g3 (Model)

2

3

4

5

6

7

8

9

0 5 10 15 20 25 30Age of concrete (days)

CTE

(mic

ron/

o C

)

.

g 1.4w4 (Tested)

g 1.4w4 (Model)

g 1.4w5 (Tested)

g 1.4w5 (Model)

γ

γ

γ

γ

2

3

4

5

6

7

8

9

0 5 10 15 20 25 30Age of concrete (days)

CTE

(mic

ron/

o C

)

.

g 1.2w5 (Tested)

g 1.2w5 (Model)

g 1.4w5 (Tested)

g 1.4w5 (Model)

γ

γ

γ

γ

(a) (b)

Figure 7. Comparison between test and predicted results of CTE of pastes (a) cement- paste with w/b = 0.25, 0.35 and 0.40. (b) cement-fly ash paste with fly ash replacement ratio of 0, 0.3 and 0.5, and w/b =

0.35.

(a) (b)

Figure 8. Comparison between test and predicted results of CTE of mortars with sand to binder ratio of 0,1 and 3, and w/b = 0.40 (a) river sand mortars. (b) crushed limestone sand mortar.

(a) (b)

Figure 9. Comparison between test and predicted results of CTE of concrete (a) γ of 1.4, w/b = 0.40 and 0.5. (b) γ of 1.2 and 1.4, w/b = 0.5

5. CONCLUSION

The conclusions based on the measurement are as follow: 1. It can be concluded from the experimental results that the specific heat depends largely on the amount of free water content in the specimens. The replacement of cement by fly ash yields high specific heat at young age but continues to decrease in long term due to pozzolanic reaction. The proposed equations can be nearly quantitatively used to predict the specific heat of the tested pastes, mortars and no-fine concrete.

28 CMT - 236


CMT -

2. Thermal conductivity of all samples decreases with age. The proposed equations can be nearly quantitatively used to predict the thermal conductivity of the tested pastes. 3. Thermal expansion coefficient of concrete is a time-dependent property which increases with age. Thermal expansion coefficient of cement – fly ash paste increases since the first day and still increases in long term due to continuing pozzolanic reaction. The use of aggregates reduces the thermal expansion coefficient of concrete. The use of higher CTE aggregate gives higher CTE of concrete. In regard of mix proportion, aggregate seems to be the main factor that affects CTE of concrete because it occupies most of the concrete volume. CTE model of concrete was verified to be satisfactory for predicting the CTE of cement-fly ash paste. CTE model of concrete was also proved to be satisfactory to predict the authors’ test results of CTE of mortar and concrete.

6. REFERENCES

ACI Committee 211 (1988). “Standard Practice for selecting Proportions for Normal, Heavy weight, and Mass Concrete.” Manual of Concrete Practice, Part 1, pp. 34.

Neville, A. M.( 1995). “Properties of Concrete. Fourth Edition.” Longman Scientific and Technical Autar K. Kaw ( 1997). “Mechanics of Composite Materials.” CRC Press Boca Raton New York. Danish Technological Institute Building Technology (1994). “Test Method: Expansion Coefficient of

concrete.” Klieger P. and Lamond J. F. (1994). “Signifinance of Tests and Properties of Concrete and Concrete-

Making Materials.” ASTM Publication, USA. Mangutova1 B., Angjuševa1 B., Miloševski1D, Fidanevska1 E., Bossert J., Miloševski1 M. (2004).

“Utilization of Fly Ash and Waste Glass in Production of Glass Ceramics Composites.” Bulletin of the Chemists and Technologists of Macedonia, Vol. 23, No. 2, pp. 157–162.

Nipatsat N. and Tangtermsirikul S. (2000). “Compressive strength prediction model for fly ash concrete.” Thammasat International Journal of Science and Technology 11, pp. 1-7.

Super Civil CD (2005). “Coefficient of Thermal Expansion.” http://www.supercivilcd.com/THERMAL.htm, version 2 Released - MAY.

Tangtermsirikul S. & Saengsoy W. (2002). “Simulation of Free Water Content of Paste with Fly Ash.” Research and Development Journal of the Engineering Institute of Thailand.

The American Society of Heat and Refrigerating Engineers Fundamentals Handbook. (1997). Neekhra S. (2004). “A New Mineralogical Approach to Predict the Coefficient of Thermal Expansion

of Aggregate and Concrete.” Master Thesis. Texas A&M University, US. Yomeyama K, Tanzil G and Toyama M. (1993). “Effective tensile young’s modulus of early aged

concrete through compressive loading.” The 47th Annual Meeting of Japan Cement Association, pp 380-5 (in Japanese).

28 CMT - 237


CMT -

PREDICTION OF COMPRESSIVE STRENGTH OF CONCRETE USING SUPPORT VECTOR MACHINE

Doo Kie Kim1, Jong Jae Lee 2, Seong Kyu Chang3, Sang Kil Chan 4, and Jong In Kim5

ABSTRACT : The compressive strength of concrete is commonly used criterion in producing concrete. However, the tests on the compressive strength are complicated and time-consuming. More importantly, it is too late to make improvement even if the test result does not satisfy the required strength, since the test is usually performed at the 28th day after the placement of concrete at the construction site. Therefore, an accurate and practical strength estimation method before the placement of concrete is highly desirable. In this study, the estimation of the compressive strength of concrete is performed by support vector machine for regression (SVMR) on the basis of concrete mix proportions. Applications of SVMR in the compressive strength estimation of concrete are carried out using the mix proportion data and the actual test results of a ready-mixed concrete company. Then the performance of SVMR is compared with that of the artificial neural network (ANN). It shows that the present method is very efficient and practical in terms of the estimation capability and the computational time.

KEYWORDS: Concrete Strength, Strength Prediction, Support Vector Machine for Regression (SVMR), Concrete Mix Proportions.

1. INTRODUCTION

Concrete is one of the most widely used construction materials in the world. Traditionally, concrete has been fabricated from a few well-defined components: cement, water, fine and coarse aggregates, etc. In concrete mix design and quality control, the strength of concrete is a very important property. Many properties of concrete such as elastic modulus, water-tightness or impermeability, resistance to weathering agents, etc. are directly related to the strength. The strength parameters of concrete include compressive, tensile, flexural, shear, bond strength, and so on. However most of concrete elements are designed on the basis of the compressive strength of the material. The mixture design of concrete targets its 28-day compressive strength which is based on a standard uni-axial compression test and is accepted conventionally as a general index of concrete strength. Generally, concrete testing procedures require special equipment and are time-consuming. Furthermore, experimental errors are inevitable. A typical test performed 28 days after concrete placement may be too late to make improvements if the test results do not satisfy the required criterion. Therefore, accurate and realistic strength estimation before the placement of concrete is desirable. Over a period of many years, researchers have proposed various methods for predicting concrete strength. Firstly, conventional methods for predicting 28-day compressive strength of concrete are basically based upon statistical analyses, by which many linear and nonlinear regression equations have been constructed to model such prediction problems (Snell et al., 1989; Popovics, 1998). Such traditional prediction models have been developed with a fixed equation form based on a limited number of data and parameters. Secondly, a standard multi-layer feed-forward neural network with a back propagation algorithm (artificial neural network, ANN) has been used to predict the compressive strength of concrete (Lai and Serra, 1997; Yeh, 1998; Oh et al., 1999; Ni and Wang, 2000; Lee, 2003, Rajasekaran and Lee, 2003; Kim et al., 2004). Various types of data have been used as the input to the 1 Professor. Dept. of Civil and environ. Eng., Kunsan National University, Kunsan, Korea. 2 Research Associate Professor, Smart Infra-Structure Technology Center, Korea Adv. Inst. of Sci. and Tech., Daejeon, Korea 3 Graduate student. Dept. of Civil and environ. Eng., Kunsan National University, Kunsan, Korea. 4 Graduate student. Dept. of Civil and environ. Eng., Kunsan National University, Kunsan, Korea. 5 Professor. Dept. of Civil Eng., Taegu University, Taegu, Korea.

29 CMT - 238


CMT -

neural network: mix proportions, temperature and humidity history, measurement data such as slump, air content, concrete temperature, etc. ANN has an advantage of that it can effectively consider various inputs without using complicated equations, in contrast to conventional regression analyses. Also, it can easily adapt to new data through a re-training process. However, ANN needs more efforts to determine the architecture of network and more computational time in training the network. Lastly, support vector machine (SVM) (Vapnik et al., 1995) has been proposed for pattern recognition such as text classification and image recognition (Joachims, 1998), and extended to regression analysis in various applications (Mukherjee, 1997; Mülller, 1997). In this study, support vector machine for regression (SVMR) is applied for predicting the compressive strength of concrete. Training and test patterns for SVMR are based on the actual mix proportions of a ready-mixed concrete company. The predicted results using SVMR are compared with those by using the conventional ANN. The present SVMR-based estimation has shown better performance in terms of estimation accuracy and computational costs than the ANN.

2. THEORETICAL BACKGROUNDS

2.1 Artificial neural network Figure 1 presents a simple architectural layout of the artificial neural network (ANN) with the conventional back-propagation algorithm, which consists of an input layer, a hidden layer, an output layer, and connections weights between them.

Output layer {k}

Hidden layer {j}

Input layer {i}

kjW

jiW

Input value (I)

Output value (O) Desired value (T) Forward signal

propagation

Backward error propagation

Figure 1. Structure of ANN

The corresponding architecture for back-propagation learning incorporates both the forward and the backward phases of the computations involved in learning process. The learning mechanism of this back-propagation network is a generalized delta rule that performs a gradient descent on the error space to minimize the total error between the actual calculated values and the desired ones of an output layer during modification of connection strengths. In other words, a least mean square procedure is carried out which finds the values of the connecting weights that minimize the error function by using a gradient descent method. The training is accomplished in an iterative process. The procedure of training is summarized as following steps: Step 1: Assign initial values to connection strengths jiW and kjW , and to biases jθ and kθ . Step 2: Input values pinet become activations on the input neurons in an input layer. Step 3: Training and testing patterns are prepared. In this study, water-cement ratio, fine aggregate percentage, unit water content, unit cement content, unit fine aggregate content, unit coarse aggregate content, admixtures, and slump values were used as input parameters in training and test patterns. The corresponding compressive strength of concrete is the output variable. 62 test patterns were selected exclusively from the training patterns of total 217 samples. That is, 155 and 62 samples were utilized as training patterns and test patterns.

29 CMT - 239


CMT -

Step 4: Calculate input values of a hidden layer j , pjnet , using the output values of an input layer i ,

piO , connection strength jiW , and biases jθ between an input layer i and a hidden layer j . Then, the output values of a hidden layer j , pjO , are derived from pjnet and activation function )(⋅f :

jpii

jipj OWnet θ+=∑ (1)

)( pjjpj netfO = (2)

where )(⋅f is an activation function, which is generally a sigmoid function :

)1/(1)( xexf −+= (3) Step 5: Calculate input values of an output layer k , pknet , using the output values of a hidden layer j ,

piO , connection strength kjW , and biases kθ between a hidden layer j and an output layer k . Then, the output values of an output layer k , pkO , are derived from pknet :

kj

pjkjpk OWnet θ+=∑ (4)

)( pkkpk netfO = (5)

Step 6: The error E between the calculated concrete strength value pkO and the desired concrete strength value kT of an output layer may be defined as

∑=

−=1

2)(21

kkpk TOE (6)

In the back-propagation network, the error at output neurons is propagated backward to hidden layer neurons, and then to input layer neurons modifying the connection weights and the biases between them by a generalized delta rule. The modification of the weights and the biases in a generalized delta rule is used through a gradient descent of the error. From hidden to output neurons

pjkkj OW ηδ=Δ and kkB ηδ=Δ (7)

where )()( pjpkkk netfOT ′−=δ and η = the learning rate. And from input to hidden neurons

pijji netW ηδ=Δ and jjB ηδ=Δ (8)

where )( pjkkjj netfW ′= δδ . Step 7: Repeat Steps 1 to 6 until error E goes below a target error. 2.2 Support vector machine for regression The support vector machine (SVM) can be applied to regression problem by introducing an alternative loss function. The loss function must be modified to include a distance measure. Figure 2 illustrates ε -insensitive loss function.

29 CMT - 240


CMT -

Figure 2. ε -insensitive loss function

Consider approximating the set of data,

),(,),,( 11 NN yy xx L , nRx∈ , Ry∈ (9)

With a linear function,

bxf +⋅= )()( xw . (10)

the optimal regression function is given by the minimum of the functional,

⎟⎟⎠

⎞⎜⎜⎝

⎛++= ∑∑

==

N

ii

N

iiC

1

*

1

2*

21),,( ξξξξ wwΦ , (11)

where C is a pre-specified value, and *, ξξ are slack variables representing upper and lower constraints on the outputs of the system. Using an ε -insensitive loss function (Figure 2),

ε

εε<−

⎩⎨⎧

−−=

yfotherwiseyf

foryL

)()(

0)(

xx (12)

the solution is given by,

⎪⎪

⎭

⎪⎪

⎬

⎫

⎪⎪

⎩

⎪⎪

⎨

⎧

+−−+

⋅−−−

=

∑

∑∑

=

∗

= =

∗∗

∗∗∗ N

iiiii

N

i

N

jjijjii

yyW

1

1 1

,,)()(

))()((21

max),(maxεαεα

αααααα

αααα

xx

, (13)

with constraints,

∑=

∗

∗

=−

=≤≤

=≤≤

N

iii

i

i

NiC

NiC

1

0)(

,,1,0

,,1,0

αα

α

α

L

L

. (14)

Solving Equation (13) with constraints Equation (14) determines the Lagrange multipliers, ∗

ii αα , , and the regression function is given by Equation (10), where

29 CMT - 241


CMT -

][

21

)(1

sr

i

N

iii

wb xx

xw

+⋅−=

−=∑=

∗αα. (15)

The Karush-Kuhn-Tucker (KKT) conditions that are satisfied by the solution are

Niii ,,1,0 L==∗αα . (16)

Therefore the support vectors are points where exactly one of the Lagrange multipliers is greater than zero. When 0=ε , we get the 1L loss function and the optimization problem is simplified,

i

N

ii

N

iji

N

jji y∑∑∑

== =

−⋅11 1

)(21min βββ

βxx . (17)

with constraints,

0

,,1,

1

=

=≤≤−

∑=

N

ii

i NiCC

β

β L

, (18)

And the regression function is given by Equations (9) and (15) A non-linear model is usually required to model data adequately. A non-linear mapping can be used to map the data into a high dimensional feature space where linear regression is performed. The kernel approach is again employed to address the curse of dimensionality. The non-linear support vector machine for regression solution, using an ε -insensitive loss function, Figure 2, is given by,

⎪⎪

⎭

⎪⎪

⎬

⎫

⎪⎪

⎩

⎪⎪

⎨

⎧

−−−

+−−

=

∑∑

∑

= =

∗∗

=

∗

∗∗∗ N

i

N

jjijjii

N

iiiii

K

yyW

1 1

1

,,),())((

21

)()(max),(max

xxαααα

εαεααα

αααα, (19)

with constraints,

∑=

∗

∗

=−

=≤≤

=≤≤

N

iii

i

i

NiC

NiC

1

0)(

,,1,0

,,1,0

αα

α

α

L

L

. (20)

Solving Equation (19) with constraints Equation (20) determines the Lagrange multipliers, ∗

ii αα , , and the regression function is given by

bKfSVs

iii +−=∑ ∗ ),()()( xxx αα (21)

where

29 CMT - 242


CMT -

)],(),([)(

21

),()(

isirSVs

ii

iSVs

ii

KKb

K

xxxx

xxxw

+−−=

−=⋅

∑

∑∗

∗

αα

αα

. (22)

If the Kernel contains a bias term, the bias can be accommodated within the Kernel function, and the regression function is given by,

∑ ∗−=SVs

iii Kf ),()()( xxx αα (23)

In this study, a radial basis function machine with convolution function given in Equation (23) is used as the kernel function.

)exp(),( 2

2

σi

iKxx

xx−

−= (24)

2.3 Estimation of concrete strength using support vector machine Concrete structures are generally required to have safety, strength, durability, and serviceability. In order to produce high-quality concrete to satisfy these needs, code information, specifications, and experience of experts in determining the concrete mix proportions play vital roles. The concrete used at construction sites is mostly produced in a ready-mixed concrete company according to specified concrete mix proportions. In general, slump tests are performed before the placing of concrete, but the compression tests of specimens are carried out at the 28th day after the placing. Therefore, it is difficult to estimate the compressive strength on construction sites. Ready-mixed concrete companies use their own mix proportions based on codes, previous experience, and experiment. In this study, the support vector machine for estimating the concrete compressive strength was incorporated using the actual mix proportion data provide by a concrete company. The material properties of concrete are shown in Table 1.

Table 1. Material properties of concrete

Normal Portland cement is used. The maximum size of aggregate is 25mm, the range of compressive strengths is from 9.8 to 39.2 MPa, and slump values are 5, 8, 10, 12, 15, 18, and 21cm. At first, nine parameters including water-cement ratio, fine aggregate percentage, unit water content, unit cement content, unit fine aggregate content, unit coarse aggregate content, admixtures, and slump are used as input set for SVMR, while the specified compressive strength is defined as the output to be estimated. All the input data are normalized to 0.1~0.9 to give an equal weighting factor before implementing the data to the network. To investigate the generalization capability of the SVMR, 62 test patterns are selected exclusively from the training patterns of total 217 samples; 155 and 62 samples are utilized as training patterns and test patterns for SVMR, respectively. Table 2 shows the samples of the data sets used for the training of SVMR.

Properties of material Experiment data Cement 3.14

Natural sand (s1) 2.59 Crushed sand (s2) 2.51

Specific gravity

Coarse aggregate 2.64 Natural sand (s1) 3.30 Crushed sand (s2) 2.25 Fineness

modulus Coarse aggregate 6.53 Admixtures Air-entraining

admixtures AE water-reducing (Standard)

29 CMT - 243


CMT -

Table 2. Samples of specified concrete mix proportions for training

Unit fine aggregate content (kN/m3) Specified

strength (MPa)

Slump (cm)

W/C weightratio (%)

Fine aggregate percentage

(%)

Unit water content (kN/m3)

Unit cement content (kN/m3) Natural

sand (s1) Crushed sand (s2)

Unit coarse aggregate content (kN/m3)

Admixture (%)

9.8 8 84.9 50.4 1.77 2.09 3.65 5.47 9.22 0.64 11.76 10 76.9 49.2 1.79 2.33 3.51 5.27 9.32 0.72 13.72 12 69.9 48.2 1.81 2.61 3.39 5.08 9.35 0.80 13.72 21 69.9 50.0 2.00 2.87 3.37 5.07 8.67 0.88 15.68 10 64.2 46.6 1.75 2.73 3.28 4.93 9.67 0.84 15.68 15 64.2 47.6 1.86 2.90 3.27 4.92 9.26 0.89 17.64 5 59.4 44.7 1.64 2.75 3.19 4.80 10.17 0.84 17.64 12 59.4 46.1 1.78 3.00 3.19 4.79 9.59 0.92 17.64 18 59.4 47.3 1.91 3.21 3.19 4.77 9.11 0.98 20.58 12 53.5 44.9 1.76 3.29 3.08 4.62 9.70 1.01 20.58 18 53.5 46.1 1.89 3.54 3.07 4.60 9.19 1.08 23.52 8 48.6 43.1 1.67 3.43 2.98 4.47 10.09 1.05 23.52 12 48.5 43.9 1.75 3.61 2.97 4.46 9.75 1.10 26.46 10 44.2 42.7 1.70 3.83 2.88 4.32 9.94 1.17 26.46 18 44.3 44.3 1.86 4.20 2.86 4.29 9.23 1.29 29.4 10 40.9 42.0 1.69 4.13 2.80 4.19 9.93 1.26 29.4 15 40.9 43.0 1.79 4.38 2.78 4.17 9.47 1.34 34.3 10 35.7 40.9 1.68 4.69 2.66 3.98 9.85 1.44 34.3 18 35.7 42.5 1.83 5.14 2.63 3.94 9.12 1.57 37.24 18 33.4 42.1 1.83 5.46 2.56 3.84 9.04 167 39.2 15 32.1 41.2 1.76 5.50 2.53 3.79 9.26 1.68

Table 3 and Figure 3 show the samples of the data sets used for the testing and estimation results of SVMR.

Table 3. Examples of estimation results for SVMR

Unit fine aggregate content

(kN/m3) Specified strength (MPa)

Slump (cm)

Water-cement Ratio (%)

Fine aggregate

percentage (%)

Unit water

content (kN/m3)

Unit cement content (kN/m3) Natural

sand(s1)Crushed sand(s2)

Unit coarse

aggregate content (kN/m3)

Admixture (%)

Estimated strength (MPa)

9.8 12 84.8 51.2 1.86 2.20 3.64 5.47 8.93 0.67 9.79 11.76 18 76.8 50.7 1.96 2.55 3.51 5.26 8.73 0.79 11.78 15.68 12 64.3 47.1 1.80 2.79 3.29 4.93 9.50 0.86 15.67 17.64 18 59.4 47.3 1.91 3.21 3.18 4.77 9.12 0.98 17.64 19.6 8 55.2 44.5 1.69 3.06 3.11 4.67 10.00 0.93 19.59 20.58 12 53.5 44.9 1.76 3.29 3.08 4.62 9.70 1.01 20.59 23.52 10 48.4 43.5 1.71 3.53 2.97 4.46 9.92 1.08 23.52 25.48 15 45.6 43.9 1.80 3.96 2.90 4.35 9.50 1.21 25.48 27.44 10 43.2 42.4 1.69 3.93 2.85 4.28 9.93 1.20 27.44 29.4 8 40.9 41.6 1.65 4.03 2.80 4.21 10.10 1.31 29.40 33.32 18 36.8 42.7 1.84 5.02 2.66 3.98 9.12 1.53 33.34 35.28 12 35 41.2 1.71 4.90 2.63 3.94 9.63 1.50 35.30 37.24 15 33.4 41.5 1.77 5.29 2.57 3.85 9.31 1.62 37.25 39.2 18 31.9 41.8 1.82 5.67 2.51 3.77 8.99 1.74 39.18

29 CMT - 244


CMT -

1 2 3 4 5 6 7 8 9 10 11 12 13 140

5

10

15

20

25

30

35

40concrete strength

number of data

stre

ngth

(MP

a)

Estimated strengthSpecified strength

Figure 3. Estimation result of SVMR

3. ESTIMATION RESULTS

In order to verify the availability of support vector machine for regression (SVMR), we compared the estimation results of the SVMR with those of the ANN. Training patterns and test patterns for SVMR and ANN are same data sets. The epochs of training for ANN were limited to 400. Table 4 shows the estimation errors of SVMR and ANN for all the test patterns and the RMS error defined as

( )∑=

−=N

iff

Ne

1

21 (25)

where N is the number of test patterns; f and f denote the actual and predicted concrete strength respectively. From the Table 4, it has been found that both the estimation errors and time-consuming of SVMR are less than those of ANN.

RMS : 0.0355, Time : 12.9(sec)

0 10 20 30 40 50 60-0.5

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

0.5

Test pattern No.

Estim

atio

n er

ror[M

Pa]

RMS : 0.0102, Time : 1.7(sec)

0 10 20 30 40 50 60-0.5

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

0.5

Test pattern No.

Estim

atio

n er

ror[M

Pa]

(a) ANN (b) SVMR

Figure 4. Estimation errors of SVMR and ANN

29 CMT - 245


CMT -

4. CONCLUSIONS

This paper presents a promising support vector machine algorithm for predicting the compressive strength of concrete based on its mix proportion parameters, such as water-cement ratio, fine aggregate percentage, unit water content, unit cement content, unit fine aggregate content, unit coarse aggregate, admixture, and slump. The validity of the proposed method was proven by comparing the predicted strengths using SVMR and ANN, respectively. From the comparison results, it has been found that the present SVMR method is very efficient in predicting the compressive strength of concrete in terms of estimation accuracy and computational time. Further, it is expected that the present SVMR method for predicting the compressive strength of concrete can contribute to the maintenance of concrete quality for optimal concrete mixtures. As the database containing mix proportions, specified and tested strengths is expanded over time, support vector machine using the training data obtained from this database will become more effective and the resulting predictions will become more reliable. In a future study, other important parameters that also affect concrete strength such as the uncertainty of concrete (i.e., the quality variation of aggregate and cement, measuring error, mixing conditions, etc.) and in-field conditions (i.e., delivery distance, curing conditions, etc.) and their co-relationships need to be collected and considered in support vector machine.

5. ACKNOWLEDGEMENTS

This work was supported by grant No. R01-2006-000-10610-0 (2006) from the Basic Research Program of the Korea Science & Engineering Foundation (KOSEF). The authors wish to express their gratitude for the financial support.

6. REFERENCES

Joachims, T. (1998). “Proceedings of the European Conference on Machine Learning.” Springer Berlin. New York.

Kim, J. I., Kim, D. K., Feng, M. Q., and Yazdani, F. (2004). “Application of Neural Networks for Estimation of Concrete Strength.” Journal of materials in Civil Engineering, ASCE, V. 16, No. 3, pp. 257-264.

Lai, S., and Serra, M. (1997). “Concrete Strength Prediction by Means of Neural Network.” Construction and Building Materials, V. 11, No. 2, pp. 93-98.

Lee, S.C. (2003). “Prediction of Concrete Strength Using Artificial Neural Networks.” Engineering Structures, V. 25, pp. 849-857.

Ni, H.G. and Wang, J.Z. (2000). “Prediction of Compressive Strength of Concrete by Neural Networks.” Cement and Concrete Research, V. 30, pp. 1245-1250.

Oh, J. W., Lee, I.W., Kim, J. T., and Lee, G. W. (1999). "Application of Neural Networks for Proportioning of Concrete Mixes." ACI Material Journal, V.96, No.1, pp.61-67.

Snell L.M., Van Roekel J., and Wallace N.D. (1989). “Predicting early concrete strength.” Concrete International, V. 11, No. 12, pp. 43–47.

Popovics S. (1998). “History of a mathematical model for strength development of Portland cement concrete.” ACI Materials Journal, V. 95, No. 5, pp. 593–600.

Rajasekaran, S. and Lee, S.C. (2003). “Prediction of Concrete Strength Using Serial Functional Network Model.” Structural Engineering and Mechanics, V. 16, No. 1, pp. 83-99.

Vapnik, V. (1995). “The Nature of Statistical Learning Theory.” Springer Berlin. New York. Yeh, I-C. (1998). “Modeling of Strength of High-Performance Concrete Using Artificial Neural

Networks.” Cement and Concrete Research, V. 28, No. 12, pp. 1797-1808. Mukherjee, S., Osuna, E. and F. Girosi. (1997). “Nonlinear Prediction of Chaotic Time Series using

Support Vector Machines”. To appear in Proc. of IEEE NNSP’97, Amelia Island, FL, 24-26 Sep. Mulller, K. R., Smola, A., Ratsch, G., Scholkopf, B., Kohlmorgen, J. and V. Vapnik. (1997).

“Predicting Time Series with Support Vector Machines.” Proceedings of ICANN'97, Lausanne.

29 CMT - 246


CMT -

AN EXPERIMENTAL STUDY ON THE EFFECTS OF VERTICAL VIBRATION DURING THE INITIAL CURING ON THE CONCRETE STRENGTH

Kwang-Soo Kim1, Sun-Kyu Park2, Kyoung-Bong Han3, Jun-Myung Park4, and Young-Jae Lee5

ABSTRACT : A bridge, a road structure, goes through construction for expansion, as time goes, on while it is still in use for many reasons such as society, economy, traffic environment, etc. Within the section under construction, the concrete structure is affected by the vibration resulting from the dynamic load of vehicles passing on the existing bridge when final concrete is placed at the connection between the existing and new concrete. There is currently no clear rules or standards in the design, however, to properly protect concrete thus in initial curing, for providing a certain level of strength, from outside vibration or impact. This study, therefore, applied vibration to concrete in initial curing, designated the vibration rate and period as the experimental variables, and evaluated their characteristics in effects on the compressive and bond strength of concrete.

KEYWORDS: Expansion construction , Vibration, Initial curing, Concrete, Vibration rate

1. INTRODUCTION

A bridge is designed and constructed so as to provide structural stability required by the design standard. A bridge, therefore, should sufficiently display the performance set at the design phase in its load carrying capacity and structural stability when assuming that there is little environmental change affecting the bridge, that the construction quality and maintenance is good, and that the bridge has not changed in the characteristics of the materials constituting the bridge. A bridge, however, goes through construction for expansion, as time goes on, during its use for many reasons such as changes in society, economy, traffic environment, etc. The concrete thus in initial curing is often affected by the surrounding vibration sources though it needs to be properly protected from external vibration or impact to provide specified strength. There is no clear relevant specification, however, in "Road Bridge Design Standards" or "Concrete Structure Design Standards." This study was conducted to provide the data required to establish construction measures to ensure required quality during placement of concrete and to minimize the effects of vibration, which results from the dynamic load of cars passing on the existing bridge when final concrete is placed at the connection between the existing and new base plates, on the structure by analyzing the effects of vibration on the concrete structure during curing. This study, therefore, investigated the changes in the characteristics of initially cured concrete of the expanded bridge that result from vehicles passing on the existing bridge when it is expanded during use. This study investigated the characteristics of thus vibration-applied concrete in its compressive and bond strength by dividing the experimental variables into vibration rate and period.

2. DOMESTIC AND OVERSEAS STUDIES

2.1 Domestic Studies In Korea, the subject has been studied since late 1980s. Y. W. Kwon, in his study, "Effects of Vibration on Initial Strength of Concrete," conducted experiments by applying impact to a mold with a

1 Department of Civil Engineering, University of Sungkyunkwan, Suwon, South Korea 2 Department of Civil Engineering, University of Sungkyunkwan, Suwon, South Korea 3 Department of Civil Engineering, University of Sungkyunkwan, Suwon, South Korea 4 Department of Civil Engineering, University of Sungkyunkwan, Suwon, South Korea 5 Department of Civil Engineering, National University of Sangju, Sangju, South Korea

30 CMT - 247


CMT -

steel hammer after arbitrarily specified time, and found that the compressive strength of concrete increases with the increase of the impact applied during the initial curing of concrete. S. G. Hong et al., in his study on the bending and fatigue behaviour of RC plate decks, studied the vibration effects resulting from cars passing through the existing bridge on the newly constructed bridge while the concrete is being cured when the connection construction method is selected for bridge width expansion. B. H. Oh et al. studied the vibration effects of nearby vehicles on the concrete in initial curing when the bridge is repaired or its width is expanded. J. G. Han, in his master's thesis, announced the strength characteristics of concrete found when transverse vibration is applied. J. S. Park, in his master's thesis, recently studied the mechanical effects of vertical vibration on concrete in initial curing. 2.2 Overseas Studies Bastian measured the strength of the concrete under vibration during the pile driving process and found that the strength was approx. 4% higher than that without such vibration. Howes continued to apply vibration to the test pieces mounted on a vibrating table once per hour for one week, and measured its density and compressive and tensile strength after 7 and 28 days of curing - the measurement showed no significant difference in the compressive strength by the changes of the vibration rate. Krell studied the effects of the vibration of the power generation facility of the existing concrete foundation on new concrete foundation in curing to be connected with the existing one. He measured the compressive and bond strength of the test pieces. In result of his test, the compressive strength was higher than that without any vibration and the test piece failed due to cut reinforcements without any bond failure between reinforcements and concrete when vibration is applied.

3. EFFECTS OF VIBRATION ON CONCRETE

The strength of concrete, the essential index determining the quality of concrete, is heavily affected by its composition, materials, placement, and curing. 3.1 Void and Compaction of Concrete The strength and physical characteristics of concrete are affected by its degree of compaction and the water-cement ratio: the strength is higher when the concrete is compacted by machines than by hands, and the strength increases with the decrease of the ratio though it rather decreases if the concrete is not sufficiently compacted. Initial vibration on concrete in curing may result in rather advantageous effects in the quality as the void and porosity decreases. The quality may be rather lowered, however, if vibration is applied during the hydration and hardening of concrete without compaction right after placement may rather degrade the quality. All chemical reactions are usually accelerated by physical catalytic activities such as agitation or initial vibration. The hydration of cement or concrete is also a type of hydraulic chemical reaction that increases the strength. In case of bleeding of concrete also, initial vibration or impact improves concrete in curing as the strength is increased because water is extracted from the foundation (i.e. lower part) during initial curing and thus the water-cement ratio approaches the lower limit. The initial vibration and impact, on the other hand, may be hazardous in weakening the concrete, increasing the porosity, and reducing the durability with high water-cement ratio in the upper part. 3.2 Setting of Concrete Concrete setting is different from hardening that provides measurable strength - setting is transition that occurs during hydration before hardening. Setting is a critical phenomenon from fluid to solid

30 CMT - 248


CMT -

whose period is measured by the penetration test. The period, shown in Figure 1 as the time from the initial to the final setting, shows if the unhardened cement paste can be mixed and placed. Cement hydration is active since the timing water is supplied into the concrete mixer and setting mainly depends on the hydration extent of SC3 ( 23CaOSiO ). The term of fluidity corresponds to the deactivation of SC3 ( 23CaOSiO ) and the setting is when the rapid hydration occurs upon the completion of the deactivation. The setting also rapidly raises the heat generation temperature and lowers the electric conductance and the sound transmission speed through the paste. The initial vibration applied then leads to catalytic effects that accelerates the concrete hydration resulting in hazards in quality control such as abnormal setting, reduced slump, increased unit quantity, and cracks. It should be noted that initial vibration is rather beneficial in increasing the concrete strength. Quick setting, in the mean time, is the worst type of setting where too much activity of AC3 rapidly sets the concrete and where even more mixing cannot enable the flow of water, cement, and aggregates. Any vibration applied then may result in hazardous effects in its chemical reaction but also in benefits in filling the void of the concrete.

Figure 1. Concrete setting and hardening Where, (a) limit of mechanical tamping (b) limit of hard tamping (c) the first tamping (d) the last setting

4. STANDARD ON ALLOWABLE VIBRATION

There are may types of concrete and standards for allowable vibration levels should be provided for each of them in view of the characteristics. Rather conservative standards are reviewed for application as a general guideline, however, because there are not such specific standards yet. ASCE has proposed a rather conservative standard for Portland cement in 1970. They regarded the term from 0 to 12 hours after placement as the most sensitive period with the vibration rate of 0.254 cm/sec. As the standard was controversial with too low values, a series of systematic experiments were conducted to review the allowable standard in relation to the “Seabrook Nuclear Station Project” in the USA. Hulshizer, in result, found that the term from 3 to 11 hours after concrete placement is the most sensitive and proposed the allowable vibration for concrete in curing as shown in Table 1.

30 CMT - 249


CMT -

Table 1. Vibration limit for curing concrete

Time after placement(hours) Peak particle velocity(cm/sec) 0 ~ 3 10.16

3 ~ 11 3.81 11 ~ 24 5.08 24 ~ 48 10.16 48 more 17.78

The Department of Transportation, the USA, has proposed the values much lower than those in Table 1 (quite conservative - see Table 2).

Table 2. Vibration limit for curing concrete(DOT, 1991)

Time after placement Peak particle velocity (cm/sec) 0 ~ 4 (hours) 5.08 4 ~ 24 (hours) 0.63 1 ~ 3 (days) 2.54 3 ~ 7 (days) 5.08

7 ~ 10 (days) 12.70 10 days more 25.40

In case of Germany, Vibratech has proposed Table 3 as allowable vibration for concrete structures in curing.

Table 3. Vibration limit for curing concrete(Vibratech, 1991)

Time after placement Peak particle velocity (cm/sec) 12 (hours) 0.635 24 (hours) 1.27 48 (hours) 2.54 7 (days) 6.35

14 (days) 10.16 28 (days) 12.70

In case of Korea, meanwhile, Korea Institute of Geoscience and Mineral Resources has corrected the most vibration-sensitive term to 3 to 24 hours based on their study result compared with the standard on allowable vibration of the Department of Transportation of the USA - see Table 4.

Table 4. Vibration limit for curing concrete(KIGAM, 1991)

Time after placement Peak particle velocity (cm/sec) 0 ~ 3 (hours) 5.08 3 ~ 24 (hours) 0.63 1 ~ 3 (days) 2.54 3 ~ 7 (days) 5.08

7 ~ 10 (days) 12.70 10 days more 25.40

5. VERTICAL VIBRATION EXPERIMENT

Tables 5 and 6 show the mix ratio of concrete and the material properties of the reinforcements used in this study for tests for bond strength - this study used the Portland cement and crushed gravels with the maximum size of 20 mm for aggregates. The design strength of the concrete was 26.48 MPa and the yield stress of the reinforcements was 392.27 MPa - Table 6 shows the specimen test result.

30 CMT - 250


CMT -

Table 5. Mixture proportion for concrete(kg/m3)

Cement Water Fine aggregate Coarse aggregate Fly ash AE W/C

337 191 750 993 46 0.77 45.3

Table 6. Properties of steel (MPa)

Bar Tensile strength Yield strength Elongation(%) SD40, D19 605.07 494.75 17.1

5.1 Test Piece Fabrication and Test Variables This study has fabricated total 120 test pieces, by the KSF specifications, for experiments on compressive and bond strength: 60 units (φ 100 mm×200 mm) for measuring compressive strength in consideration of the vibration speed and period, and 60 other units (150 mm×150 mm×150 mm) for measuring bond strength of vertical reinforcements also in consideration of the vibration speed and period. The experimental variables were vibration rate and the vibration period: the vibration rates were set as four types, 0.3, 0.45, 0.6, and 1.0 cm/sec, based on the literature study result and the vibration periods were divided into four cases, 3, 6, 12, and 24 hours, in view of the concrete setting time and other experimental conditions. The timing when vibration began to be applied was right after concrete placement and not included in the experimental variables. Figure 2 shows the test pieces.

V 0.45 - 12C

vibrating velocities(0.3, 0.45, 0.6, 1.0cm/sec)

vibrating time(3,6,12,24hours)

C : compression B : bond

V : vibrating N : no vibrating

Figure 2. Experimental variables 5.2 Vibrator This study used an MTS actuator to apply vertical vibration, and fabricated a 10 mm thick base plate with bolts (diameter: 015 mm) welded, to fix the actuator thereon with the 20 mm thick top vibration plate on the actuator. As shown in Figure 3, vertical vibration was directly applied during the time set to properly simulate the effects of the vertical vibration resulting from the vehicles driven on the top vibration plate.

30 CMT - 251


CMT -

Figure 3. Vibrating table


This study applied vibration for four kinds of periods (i.e. 3, 6, 12, and 24 hours) for each of the four vibration rates (i.e. 0.3, 0.45, 0.6, and 1.0 cm/sec). The experimental environment made it inevitable to separately place concrete because the number of units mounted on the vibrating table is limited though the mix ratios of concrete were the same for each vibration rate. This study tried to meet the experimental conditions set for each vibration rate and period, therefore, by showing the relative variation of strength to the control piece for each placement rather by focusing on the absolute values of the compressive and bond strength because variation in strength cannot be perfectly avoided whenever concrete is placed. The variation was shown as the ratio of the vibration-applied piece to that of the control piece, as shown in Equation (1).

100100(%) −×−

=PieceTestControlofStrength

PieceTestAppliedVibrationofStrengthR ……………………………. (1)

, where R is the strength variation: "0" means that the strength of a vibration-applied test piece is the same as that of the control piece without any variation while "+1" means that the strength has increased by 1% to that of the control piece.

6.1 Result of Compressive Strength Test Tables 7 and 8 and Figure 4 and 5 show the change of compressive strength to the vibration rate and period compared with that of the control test piece without any vibration applied.

Table 7. The fluctuation rate of compressive strength for loading time Vibrating velocities (cm/sec) Time of loading (hours) 0.3 0.45 0.6 1.0

0.0 0.0 0.0 0.0 0.0 3.0 -0.19 -1.04 4.99 -20.56 6.0 -1.78 12.83 -10.77 -16.73 12 0.0 20.34 -0.82 -24.79 24 3.61 7.02 16.31 -17.63

30 CMT - 252


CMT -

-30

-20

-10

0

10

20

30

0 5 10 15 20 25

Loading Tim e(hours)

R(%

)

0.3cm /sec 0.45cm /sec 0.6cm /sec 1.0cm /sec

Figure 4. The fluctuation rate of compressive strength for loading time

Table 8. The fluctuation rate of compressive strength for vibrating velocities

Time of loading (hours) Vibrating velocities (cm/sec) 3 6 12 24 0.3 -0.19 -1.78 0.0 3.61

0.45 -1.04 12.83 20.34 7.02 0.6 4.99 -10.77 -0.82 16.31 1.0 -20.56 -16.73 -24.79 -17.63

-30

-20

-10

0

10

20

30

0 0.2 0.4 0.6 0.8 1 1.2

Vibrating velocities(cm /sec)

R(%

)

3(hours) 6(hours) 12(hours) 24(hours)

Figure 5. The fluctuation rate of compressive strength for vibrating velocities

6.2 Discussion on Compressive Strength 1) The vibration rate of 0.3 cm/sec showed little effects on the compressive strength in that the change in strength was within 5% over the period of vibration applied during the curing of concrete. 2) The vibration rate of 0.45 cm/sec increased the compressive strength of the concrete in initial curing by maximum 20.34%. 3) The vibration rate of 0.6 cm/sec decreased the strength by 10.77% for the initial 6 hours of vibration, but increased it by maximum 16.73% for 24 hours of vibration.

30 CMT - 253


CMT -

4) The vibration rate of 1.0 cm/sec decreased the compressive strength regardless of the vibration period - maximum 25% of strength was decreased for 12 hours. 6.3 Result of Bond Strength Test Tables 9 and 10 and Figure 6 and 7 show the change of bond strength to the vibration rate and period compared with that of the control test piece without any vibration applied.

Table 9. The fluctuation rate of bond strength for loading time Vibrating velocities (cm/sec) Time of loading (hours) 0.3 0.45 0.6 1.0

0.0 0.0 0.0 0.0 0.0 3.0 7.61 9.13 15.6 -9.37 6.0 9.20 -0.23 4.34 -2.94 12 8.10 -7.89 -0.81 -14.23 24 22.21 8.46 6.40 11.48

-20

-10

0

10

20

30

0 5 10 15 20 25

Loading tim e(hours)

R(%

)

0.3(cm /sec) 0.45(cm /sec) 0.6(cm /sec) 1.0(cm /sec)

Figure 6. The fluctuation rate of bond strength for loading time

Table 10. The fluctuation rate of bond strength for vibrating velocities

Time of loading (hours) Vibrating velocities (cm/sec) 3 6 12 24

0.3 7.61 9.20 8.10 22.21 0.45 9.13 -0.23 -7.89 8.46 0.6 15.6 4.34 -0.81 6.40 1.0 -9.37 -2.94 -14.23 11.48

30 CMT - 254


CMT -

-30

-20

-10

0

10

20

30

0 0.2 0.4 0.6 0.8 1 1.2

Vibrating velocities(cm /sec)

R(%

)

3(hours) 6(hours) 12(hours) 24(hours)

Figure 7. The fluctuation rate of bond strength for vibrating velocities

6.4 Discussion on Bond Strength 1) The vibration rate of 0.3 cm/sec increased the strength to maximum 22.21% gradually over the vibration period. 2) The vibration rate between 0.45 to 0.6 cm/sec increased the strength for the initial 3 hours, decreased it for the next 9 hours (i.e. 3 to 12 hours), and finally recovered approx. 7% of the strength for the next 12 hours (i.e. 12 to 24 hours). 3) The vibration rate of 1.0 cm/sec decreased the strength for the intial 12 hours and increased it by maximum 10% for the next 12 hours (i.e. 12 to 24 hours).

7. CONCLUSION

The objective of this study was to evaluate the effects of external vibration on the strength when concrete in initial curing is exposed to the environment where vibration cannot be avoided due to the external restrictions. It is recommended, in result, to maintain the vibration rate at 0.3 cm/sec or below when the vibration period is within 6 hours. The vibration rate below 0.6 cm/sec is recommended, on the other hand, for vibration periods from 12 to 24 hours in that the bond strength decreases by 7.89% at 12 hours and increases by 8.46% at 24 hours.

8. REFERENCES

YoungWoong, Kwon (1988). “How Vibration Affects Early Concrete Strength.” Korea National Housing Corporation, Journal of Housing, V. 49, pp. 95~100.

SoonGill, Hong and Dongil, Chang (1994). “Study on the Flexural Behavior of the RC Slab in the Width-Expansion Bridge.” Journal of Korea Concrete Institute, V. 6, No. 3, pp. 152~161.

ByungHwan, Oh, HyeKum, Song and JaeYoul, Cho (1998). “How Vibration Affects Concrete in Curing.” Journal of Korea Concrete Institute, V. 10, No. 5, pp. 531~537.

JoongKi, Han (2000). “Experimental Study on the Horizontal Continuous Vibration Affects Early Age Concrete in Curing.” Master of Science Thesis, University of Pukyong, pp. 40~97.

JungSoo, Park (2003). “Mechanical Characteristics of Vibrated Concrete in Curing.” Master of Science Thesis, University of Sungkyunkwan, pp. 38~89.

Bastian, C. E. (1970). “The Effects of Vibrations on Freshly Poured Concrete.” Concrete International Design and Construction, V. 1, No. 12, pp. 31~34.

Howes, E. V. (1979). “Effects of Blasting Vibrations on Curing Concrete.” Proceedings, 20th, U.S. Symposium on Rock Mechanics, Austin, Texas, pp. 455~460.

30 CMT - 255


CMT -

Krell, W. C. (1979). “Effect of Coal Mill Vibration on Fresh Concrete.” Concrete International Design and Construction, V. 1, No. 12, pp. 31~34.

Hulshizer, A. J. (1996). “Acceptable Shock and Vibration Limits for Fleshly Placed and Maturing Concrete.” ACI Materials Journal, V. 93, No. 6, pp. 524.

30 CMT - 256


CMT -

THE ESTIMATION ON COMPRESSIVE STRENGTH OF CONCRETE BY NEW ‘SAND-CAPPING METHOD’

Park, Young-Shik1 and Suh, Jin-Kook2

ABSTRACT : The most typical capping method is a sulfur-mortar compound capping, provided

as a standard criterion by ASTM C 617, but this conventional bonded-type method has many

problems. These are underestimated strengths in high-strength concrete due to the differences of

elasticity and strength between the cylinder and the cap, and have bad serviceability such as a

dangerous work or a waste of working time. To prevent from occurring these problems in recent

years, unbonded-type capping methods have been taken the place of the conventional methods.

One of the popular methods is the use of synthetic rubber like a neoprene pad. Serious problems

still remain in this method, which are the selection, the safekeeping and the economy of pads by

the chemical characteristics. Moreover, the synthetic rubber pads cannot be used in the concrete

cylinder with the higher strength than 80MPa according to ASTM C 1231-00. New 'sand-capping

method' presented in this study, can be applicable to the compressive strength evaluation of the

high strength concrete with the range of 70~100MPa. This new method has better simplicity and

reliability than those of the existing 'sand-box', because general material such as standard sand and

simply-devised apparatus are used for the capping system. In the statistical analysis of test results,

new sand-capping method with the smallest deviation and dispersion shows much better reliability

than any other methods under ASTM C 1231/1231M.

KEYWORDS : capping, sand-cap, neoprene pad, compressive strength, high strength concrete,

reliability

1. INTRODUCTION

Various kinds of errors tend to be made due to the test method and condition in measuring compressive strength of concrete. It should be based on the specific standards for capping of the end of concrete cylinder (ASTM C 617, 2000), in order to obtain accurate results for the compressive strength of concrete. Especially in the case of high strength concrete cylinder, the reliable data of compressive strength can be measured only by special end-capping method. The method of cutting and grinding an end of cylinder is used for the compression test of high strength concrete, because the compressive strength by conventional bonded-type methods using a cap with sulfur-mortar, high-strength mortar or high-strength gypsum, has been underestimated. Unbonded capping methods such as grinding or capping with pads, is not also alternatively efficient in economy and difficulty of the test.

1 Professor, Ph.D., Dept. of Civil Engineering, Kundong University, Republic of Korea 2 Professor, Ph.D., Dept. of Civil Engineering, Kyungdong College, Republic of Korea

31 CMT - 257


CMT -

New ‘sand-capping method’ which is introduced in this study, has the simpler device as against the

other complicated ‘sand-box’ system by Purington and McCormick(1926) or Boulay and Larrard (1993). It will be ascertained that this new capping system is not only more economic in the test time and process but more reliable in the test results than any other unbonded method. These results were verified by the average compressive strength of concrete as compared with sand-capping, neoprene pad capping and sulfur-mortar capping methods according to ASTM C 1231/1231M(2000).

2. EXPERIMENTAL PROGRAMS

In this study, only the effect of capping method for concrete cylinder-end is considered in evaluation of the compressive strength. The other factors such as specimen size, loading method, testing-machine type etc. depend on ASTM C 39(2000). Three capping methods are sulfur-mortar compound capping, neoprene pad capping and sand capping. Each design strength of concrete cylinder groups is made for 21MPa(Group A), 34MPa(Group B) and 70MPa(Group C) by the following mix proportions.

Table 1. Mix Proportions of Concrete Cylinders

Batch Weight Specimen Group

Group A Group B Group C

W/b 0.53 0.37 0.28

Cement (kg/m3) 321 490 520

Water (kg/m3) 171 181.3 147

Fine Aggregate (kg/m3) 905 683 634.6

Coarse Aggregate (kg/m3) 945 982 1128.4

Superplasticizer (%) 0.5 0.8 2.0

Silica-fume (%) 0 0 10

Slump (cm) 12 8 6

Compressive Strength at 28 days (MPa) 21 34 70

2.1 Test Specimens Three mixes as shown in Table 1 were prepared to investigate the effect of capping on the compressive

strength, and cylindrical specimens of ∅100mm×H200mm were cast by moulds, and they were composed of 20 specimens for each mix type and each capping method. Ordinary Portland cement (ASTM type І) was used for all the specimens. Coarse aggregate was crushed stone with 19mm of the maximum size and river sand was used as fine aggregate. Powdered silica-fume was used as admixture for high strength concrete in this test. All the materials were mixed in 80L concrete pan mixer. The compression test of concrete cylinder was performed after the age of 90days passed by, considering

31 CMT - 258


CMT -

the differences of making and curing days. The curing of all specimens was performed in water until the age of 28 days.

2.2 Capping Methods Three kinds of capping methods were used for comparing the compressive strength each other in this study. Molten sulfur-mortar compound capping was formed by a vertical capping apparatus specified in ASTM C 617(2000), and neoprene pad was provided for an unbonded capping system in testing hardened concrete cylinders in accordance with the capping standards described in ASTM C 1231/ C 1231M (2000). The elastomeric pad like neoprene pad deforms at initial loading to conform to the contour of the cylinder-end, and it is restrained from excessive lateral spreading by plate and metal ring to provide a uniform distribution of load from the bearing blocks of testing machine to the end of concrete cylinder. In the case of sand capping suggested in this study, the cylinder is placed only on the top of sand in a simple capping device, without any other sealing process or sub-device of the existing sand-box systems. Dry and fine sand which 100% of it was passed through No.20 sieve was used as sand capping material.

Figure 1. Device for Sand-Capping System and Neoprene pads

31 CMT - 259


CMT -

Figure 2. Post-Testing Status of Neoprene Pad and Sand Cap (Hardening of Neoprene Pad)

The average thickness of caps was calculated in Table 2 for a selection of objects for statistical analysis. The compression test of concrete cylinders was performed by 3 cylinders per capping method in sequence, in order to study the influence of eccentricity error for testing machine. The weight of sand in a device was measured by 0.1g-precision, because of the difficulty from taking accurate measurement of cap-thickness.

Table 2. Thickness of Cap

Types of Cap Average

Thickness Retainer Ring Mixture of Cap Materials

Unbonded Neoprene & Natural

Pads 12.8mm 102mm*254mm

Neoprene & Natural Rubbers/ Steel &

Aluminum Alloys

Sulfur-Mortar Compound Cap 2.4mm - Sulfur Compound

Sand Cap 38gr 101.1mm*10mm Fine Sand/ Steel & Aluminum Alloys


3.1 Verification of Test Results Sampling of test results was accomplished to investigate the compressive strength from test cylinders except underrated data, because of excessive cap-thickness, imperfect capping, or failure by eccentric loading. Test results must be demonstrated at 95% confidence level (α=0.05) so that the results are to be acceptable. The average compressive strength obtained by unbonded capping shouldn’t be less than 98% of that by bonded capping or grinding in accordance with ASTM C 1231(2000).

The calculation process for verifying the unbonded capping system was performed by the followings. For every strength Group, the difference in strengths for each pair of cylinders (‘sulfur-mortar capped cylinders vs. neoprene pad capped cylinders’ or ‘sulfur-mortar capped cylinders vs. sand capped cylinders’) was computed. Next, it was verified that the average strengths for two kinds of unbonded

31 CMT - 260


CMT -

capping cylinders were over 98% for those of sulfur-mortar capping cylinders.

di = xpi - xsi

xs = ( xs1 + xs2 + xs3 · · · xsn )/n

xp = ( xp1 + xp2 + xp3 · · · xpn )/n

where,

di : difference in strength of a pair of cylinders according to ASTM C 617(2000)

xpi : cylinder strength using unbonded capping

xsi : cylinder strength using Practice for ASTM C 617

n : number of pairs of cylinders tested for strength Groups

xs : average strength of cylinders for each Group according to Practice for ASTM C 617

xp : average strength of unbonded capping cylinders for each Group

average difference, d = ( d1 + d2 + d3 · · · dn )/n

standard deviation, sd = [∑(di – d)2/(n-1)]1/2

In order to comply with the Practice, the following relationship must be satisfied,

xp ≥ 0.98xs + (t⋅ sd)/(n)1/2

where, t is the value of ‘student’s t-test ’ for (n-1) pairs at α =0.05 from the following table:

(n-1) t(α =0.05)A

9 1.833

14 1.761

19 1.729

100 1.662

* Use linear interpolation for other values of (n-1) or refer to appropriate statistical tables. The calculation process and results are shown in Table 3 for the verification of the test. All the compressive strengths of cylinders by unbonded capping methods were presented more than 98% of the reference values for Group A, B and C, respectively.

3.2 Analysis of Test Results In the statistical analysis of test results, the compressive strengths of cylinders by capping using neoprene pads have the smallest values of deviation and dispersion for normal strength concrete, but those by sand-capping method get the smallest deviation and dispersion for high strength concrete. The average strengths of cylinders by sand-capping method are shown as the highest values of all Groups in Table 4.

31 CMT - 261


CMT -

Table 3. Verification Process of Test Results

Capping Types Calculations (MPa) Group A Group B Group C

① xp 21.75 35.42 79.02

xs 21.6 35.12 76.91

d 0.15 0.305 2.115

sd 0.6977 1.6599 3.6881

②0.98xs + (t⋅sd)/(n)1/2 21.44 35.07 76.83

Neoprene Pad vs.

Sulfur-Mortar Capping

System Qualifies ① ≥ � O.K O.K O.K

① xp 22.28 36.8 81.09

xs 21.6 35.12 76.91

d 0.68 1.685 4.185

sd 0.7516 1.6769 2.3640

②0.98xs + (t⋅sd)/(n)1/2 21.47 35.08 76.30

Sand Capping vs.

Sulfur-Mortar Capping

System Qualifies ① ≥ � O.K O.K O.K

Common Details n=20, t=1.729

Table 4. Statistical Analysis of Test Results

Capping Types Analysis of Results Group A Group B Group C

Average Strength (Mpa) 21.6 35.1 76.9

Standard Deviation 0.86 1.37 2.27

Dispersion 0.75 1.87 5.14 Sulfur-Mortar Capping

No. of Specimens 20 20 20



Dispersion 0.4 1.23 4.23 Unbonded Neoprene Pad Capping




Dispersion 0.56 0.53 1.49 Sand Capping


In the cases of sulfur-mortar capping, neoprene pad capping and sand capping, the compressive strengths for Group A, B and C in the range of 20MPa~80MPa are shown in (a), (b), (c) of Figure 3. One point on a graph indicates the compressive strength for one cylinder. The variation of average compressive strengths for each capping method is shown in Figure 4(a).

31 CMT - 262


CMT -

0

20

40

60

80

100

0 20 40 60

Test num ber

Comp

ress

ive

stre

ngth

(MPa

(a) Sulfur Mortar Capping

0

20

40

60

80

100

0 20 40 60

Test num ber

Compressive streng

th(M

Pa

(b) Neoprene Pad Capping

0

20

40

60

80

100

0 20 40 60

Test num ber

Comp

ressive strength(MPa

(c) Sand Capping

Figure 3. Compressive Strengths of Cylinders by Three Capping Methods

31 CMT - 263


CMT -

A verage C om pressive strength

0

50

100

G roup A G roup B G roup C

Strengrh level

Comp

ress

ive

stre

ngth

(MPa

)

N eoprene pad

sulfate cap

sand cap

0.7%0.9%

2.8%

3.1%

4.8%

5.4%

-2.0%

0.0%

2.0%

4.0%

6.0%


Strength level

Comp

ress

ive

stre

ngth

rat

i

N eoprene pad

sulfate cap

sand cap

(a) Average Compressive Strength of Concrete (b) Ratio of Compressive Strength

0

1

2

3


S trength level

Devi

atio

n

N eo prene pad

sulfate cap

sand cap

0

2

4

6


S trength level

Disp

ersi

on

N eoprene pad

sulfate cap

sand cap

(c) Standard Deviation (d) Dispersion

Figure 4. Statistical Analysis of Test Analysis

In Figure 4(a), The compressive strengths of cylinders make a little difference between those by neoprene pad capping and sulfur mortar capping in the lower strength level, Group A and B, but the difference of strengths is occurred by 28% in Group C. As the strength is higher, the more difference of strength is occurred. The compressive strengths of concrete by sand capping method are 3~5.4% higher than those by sulfur mortar capping, and the difference of strengths in the higher strength level is more increased. When the compressive strength of cylinder by sulfur mortar capping is 1 as a reference value, the relative values of strength by unbonded capping methods are shown by ratio in Figure 4(b). The standard deviations and the dispersions of compressive strengths in Figure 4(c) and 4(d) are increased for the higher strength level, Group C. The smallest value of deviation and dispersion is obtained in the case of sand capping method. For the cylinder test of compression strength, the higher relative strength cannot be close to proper strength value of the concrete cylinder, and it would rather be overestimated. Therefore, in the case of the smaller deviation and dispersion for estimating the compressive strength, it can get near to proper value of strength for concrete cylinder only by these reliable test results.

31 CMT - 264


CMT -

4. CONCLUSION It was found out in this study that the strength of concrete cylinder is obviously affected by capping methods and the effect increased significantly as the strength is higher. Strength variation due to various capping methods is following tendency in normal and high strength concrete cylinders. The strength of sand-capped cylinders is greater than that of sulfur-mortar capped cylinders by about 4.4%. Sand capping is a little more effective than unbonded neoprene pad capping in the strength, but sand-capping method is the most reliable in the dispersion and the strength variation when every capping method is analyzed statistically. Therefore, the conventional bonded-capping methods might be improper for test of high strength concrete. Also, unbonded neoprene pad capping is not adequate to estimating for high strength concrete, because it is not prior to sand capping system in the economy and the simplicity. Consequently, it can be made generally use of sand capping system proposed in this study to estimate the strength for normal and high strength concrete cylinder, more reliably and exactly.

5. REFERENCES

ASTM (2000). “Standard Practice for Capping Cylindrical Concrete Specimens (ASTM C 617).” Annual Book of ASTM Standards, V.04.02, ASTM, Philadelphia, pp.380-384.

ASTM (2000). “Standard Test Method for Compressive Strength of Cylindrical Concrete Specimens (ASTM C 39).” Annual Book of ASTM Standards, V.04.02, ASTM, Philadelphia, pp.24-29.

ASTM (2000). “Standard Method of Making and Curing Concrete Test Specimens in the Field (ASTM C 31).” Annual Book of ASTM Standards, V.04.02, ASTM, Philadelphia, pp.5-10.6.

ASTM (2000). “Standard Practice for Use of Unbonded Caps in Determination of Compressive Strength of Hardened Concrete Cylinders (ASTM C 1231/ C1231M-00).” Annual Book of ASTM Standards.

Boulay, C. and de Larrard, F. (1993). “The Sand-Box.” Concrete International, Vol.15, No.4, April, pp.63-66.

Purington, W. F. and McCormick, J. (1926). “A Simple Device to Obviate Capping of Concrete Specimens.” ASTM Proceedings, Vol.26, Part �, pp. 488-492.

31 CMT - 265


CMT -

DEVELOPMENT OF MATERIAL FLOW SIMULATION SYSTEM BY USING MULTI-AGENT MODEL TO EVALUATE RECYCLING THECHNOLOGIES

FOR CONCRETE RELATED MATERIALS

Manabu Kanematsu1, Satoshi Fujimoto2, Ryoma Kitagaki3 and Takafumi Noguchi4

ABSTRACT : Striking an optimal balance between the reduction of waste and the reduction of CO2 emission is a controversial issue during production of cement and concrete, and construction of concrete structures to achieve the sustainable development of cement and concrete industries in Japan. They, however, usually have a trade-off relationship. In this paper, comparative case studies are made in Tokyo area to induce the best available recycling technologies, government policies and supply chain, with the scenarios on lifetime extension of buildings, continuous investment of road construction, corporation strategies as well as the supply chains of high quality recycle aggregate, recycle roadbed. Multi-agent system is used to simulate and evaluate the environmental impacts, which are caused during production and construction activities, in relation with geographical information and governmental corporation strategies. This paper consequently shows some feasible scenarios for zero-emission with compensation of high CO2 emission..

KEYWORDS: Multi-agent, material flow, recycle, sustainability

1. INTRODUCTION

Total amount of industrial waste in Japan remains at about 400 million tons after 1990s. Waste from construction industries account for 18% of the above and is the leading waste generating industry category along with agricultural and energy industries. According to the survey by Ministry of Land, Infrastructure and Transportation Ministry Japan, demolished concrete accounts for 30-40%, and used asphalt concrete accounts for 25-35% of the total construction waste. On the above background, green procurement policy is taken to use the demolished concrete as recycled roadbed, and recycle rate of demolished concrete have a increased up to 98%. However, high road ratio and recent recession indicate that investment to road construction will rapidly fall in the very near future. On the other hand, many researchers have qualitatively pointed out that the RC buildings built in high growth period of 1970s will be demolished in the near future, which naturally result in the increase of the amount of demolished concrete. Therefore demolished concrete that are beyond the capacity of demand of recycled roadbed is considered to be land-filled, and depletion of landfill sites would be a serious problem. In this paper, in order to evaluate both in technical and social aspects of the resource-flow of concrete related materials, we will firstly introduce a resource flow simulation system named “ecoMA”. The system uses the concept of Multi-Agent-System and is designed to focus on the decision-making dynamics between each company and government within the city-scale so that social constraint of resource flow can be simulated properly. The system also uses the concept of graph theory to model the supply-chain and time. Secondly, comparative case studies are made, using the above-mentioned system, to evaluate the effect of enlargement of building lifetime as a mean to reduce the demolished concrete. The studies especially focus on the quantitative balance between the building demand and waste emission. Geographical distribution of building demand and difference between decision makings of each companies are also taken into consideration. As a sample of the most serious case, 100km grid area of Tokyo is chosen as the evaluation target. The simulations target is set from 2005 to 2055 which is intended to include long enough era for the evaluation of enlargement of building lifetime. 1 Assistant Professor, Dept. of Architecture, Tokyo University of Science, Japan. 2 JSPS Research Fellow, Graduate School of Eng., Dept. of Architecture, The Univ. of Tokyo, Japan. 3 Researcher, Center for Sustainable Urban Regeneration, School of Engineering, The Univ. of Tokyo, Japan. 4 Associate Professor, Graduate School of Eng., Dept. of Architecture, The Univ. of Tokyo, Japan.

32 CMT - 266


CMT -

2. OVERVIEW OF THE SYSTEM “ecoMA”

2.1 Design and characteristics

The main characteristic of the ecoMA is in the modeling of social, time and geographic factors of the resource flow. In comparison with the "Framework of life-cycle-assessment"(ISO 14040:1997), the system design target of ecoMA corresponds to the "Goal and Scope Definition"(ISO 14041) and "Inventory Analysis"(ISO 14042). The system focuses on the production and decision making process of each plant and company, which indicates that the system is most well-suited for the evaluation of city-scale resource flow.

2.2 Modeling of social factor

Production policies of each material companies, governmental policies and economic circumstances significantly affect the resource flow as well as the environmental impacts from the whole resource flow system. Therefore, the decision making process at the branch of the flow should be taken into consideration. In ecoMA, Multi-Agent-System (Abbrev. MAS) is used as a modeling method for the social factors, such as decision making and economic circumstances. MAS, in general sense, is a

Figure 1. Schematic diagram of MP agent

Table 1. Common Characteristics of Agents in ecoMA( Bradshaw, 1997)

Characteristics Explanation Examples of corresponding behavior in the real world of construction industries

Adaptivity Ability to learn and improve with experience.

price change of ready mixed concrete in accordance with the cost of aggregate procurement

Autonomy Goal-directedness, proactive and self-starting behavior.

zero-emission and eco-conscious procurement in building construction management

Collaborative behavior

Ability to work with other agents to achieve a common goal

price cartel in ready-mixed concrete, bid-rigging on public construction projects.

Inferential capability Ability to act on abstract task specifications.

capital investment on high quality recycled aggregate and eco-cement

communication ability:

Ability to communicate with other agents price negotiation in the procurement process of building construction

Personality Ability to manifest attributes of a ``believable'' human character.

Each companies of building materials have different corporation strategy.

Reactivity Ability to selectively sense and act. Answers differently to the price query in respect to the attributes of the agent who asks.

Temporal continuity Persistence of identity and state over long periods of time

Each material plant has consistency over time for its stock, pricing strategy etc.

32 CMT - 267


CMT -

system generated from a set of decision-makers called agent. In ecoMA, agent is categorized into 3 types, MP(Material Plant) agent, TC(Transportation Companies) agent, and GO (Governmental Organization) agent. For example, each MA agent behaves individually to meet the social demand according to the own characteristic information about specific material input/out for unit product, specific environmental load , cost information, site location , manufacturing capability, corporate strategy, and so on. In addition, agent in ecoMA have some characteristics in common as shown in Table 1.

2.3 Information and actions of agents

MP agent is a model for the organization with action/re-action of material input and output. It purchase the stock, produces the products, dispose of wastes and receive waste for recycling etc, based on its information and strategy. For example, general contractor, material production companies, building scrapping companies, landfill sites are categorized as MP agent. And the ready-mixed concrete plant, cement plant, aggregate plant, electricity generating plant, steel plant and etc. were built in ecoMA as material production companies agent. Influence of the process iP of agent iA is shown as

)),(( jinjji

out QMCombiPM = (1)

where outM :output material from the agent, inM : input to the agent iA , jQ : amount of material, and

jCombi represents combination. MP agent have 3 different interaction with each other; "query and answer about product", "query and answer about waste acceptance", and "choosing the raw material and plant". MP agent also autonomously "update self information" based on its rational strategies and self-collected information. The strategy is, of course, adaptively changeable from time to time. TC agent is a model for the organization that geographically moves the materials. For example, railway companies, truck companies, shipping agents are categorized as TC agent. Transportation process is represented as (2), note that influence of the TC agent ‘s process to the resource flow also satisfy equation (1).

)( 1out

kinn MPM =∃ (2)

where process iP and jP of agent iA and jA satisfy,

)),((1 minmmi

out QMCombiPM = (3)

)),((2 ninnnj

out QMCombiPM = (4)

GO agent is a model for the organization of which action only indirectly affects the resource flow. They collect the information of other MP/TC agents and rationally change and notify their policies.

2.4 Design of agent’s strategy

Strategy of each agent is represented as a set of preference relation, which satisfies reflectiveness, completeness, indifference and transitivity. strategy changes from time to time, thus represented as

}...3,2,1|{ niPrefS i == (5)

where i :agent choice, iPref : set of preference relation of the agent.

32 CMT - 268


CMT -

2.5 Modeling of time factor

Concept of time is introduced by the event model. Event, in ecoMA, is defined as a social activity that evokes the reaction of agents. For example, demand of construction, repair demand of road, rise of tax and time ticking etc are defined as events. The demand of construction influences directly on material flow of concrete, but also the demand of road influences indirectly because the of concrete wastes are used as a roadbed material. This modeling is on the assumption that production and demolition of any material never occurs without their needs. Events are queued as shown in Figure 2 based on the survey and estimation and evoked in order. Supply chain is defined as a tree structure, as in graph theory, having the event as a base node. Social constraints, such as recycling without the usage demand, is well introduced by this modeling. Numerical efficiency is also gained.

Figure 2. Event –Driven Model Diagram

2.6 Modeling of lifetime

Assuming that amount of past production of material iM is known, and probability distribution of lifetime is represented as )( tutwprob = , disposal time disposeut of material iM is represented as follows;

)(1 randWutut tdispose−+= (6)

where cumulative distribution of lifetime: )( tutW , its inverse function: )(1 probW − , and random number of (0,1): )1,0[=rand : ,Generally speaking, )( tutw is whether Weibull, log-normal or normal distribution.

2.7 Modeling of geographic factor

Construction is relatively heavy in its density and usage mass in comparison with materials in other industries. Consequently, fluctuation of energy consumption induced from the difference in geographic distribution of building demand or the transport decision making are negligible even within the city scale of the target system. In ecoMA, we have proposed the geographic model that enables to link transportation and plant location with the statistical data of the past building history.

2.8 Modeling of geography

Target area is approximately 100km square partitioned by 1km by 1km grid using Grid Square Code. Note that plant outside the target area is also taken into consideration.

2.9 Modeling of building demand distribution

Statistics of building history is developed in every local government in Japan. Based on the research by Nachi, population weighted building history is used in ecoMA.

32 CMT - 269


CMT -

cipr

cipri MeshPopulation

PopulationStatArea 1

= (7)

where gross floor area of mesh i : iArea , prefecture and city of mesh i : pr , ci as index, and statistic value of gross floor in pr : prStat , total mesh number of ci : ciMesh

2.10 Modeling of distance

Transport distance is calculated from correlation between linear distance and road distance. Based on the studies by Davidson, Koshiduka (1999) and Nakamura(2003), road distance in ecoMA is defined as follows; note that 25.1=Rd considering that most of the construction related vehicle are heavy.

linearroad DRdD ⋅= (8)

Table 1. Input conditions for time and geography

Conditions Target area 100km square in Tokyo

Target period 2005-2055

Figure 3. Overhead view of the target area

Table 2. Input Conditions of Material Input/Output

Material Raw Materials Production Process Building

(construction) reference consider only pumping process (JSCE, 2002)

Building (Demolition) - BCS(BCS,2002)

Road (construction) not considered

Road (Repair)

based on interview and reference and statistics (RS,2004) Amano based on interview (Nakamura, 2003) Concrete Reference (Shima) reference

based on interview (Nakamura, 2003) Cement particle as clinker (Shima) based on inteview (Nakamura, 2003)

Aggregate based on interview (Nakamura, 2003) based on inteview (Nakamura, 2003) Asphalt

Concrete Amano (2000) Amano (2000)

Recycled Aggregate based on interview (Nakamura, 2003) based on interview (Nakamura, 2003)

32 CMT - 270


CMT -

3. SIMULATION CASE STUDY

As discussed in intruduction, rapid increase of demolished concrete from the 1970s building is qualitatively indicated. The case study is designed to evaluate the effect of enlargement of building lifetime as a mean to reduce the amount of demolished concrete to be land-filled.

3.1 Input conditions

Basic input/output unit of plants is based on the interview survey by Nakamura and authors, JEMAI-LCA is referred for the basic unit of greenhouse-gas. (see Table 3) Target area is set to almost 100km grid, (northwest-end 5239-31 to southeast-end 5340-22 of the Grid Square Code). Target time era is set from 2005 to 2055. prices of materials are based on the survey by ERA (2003), with constant discount based on our interview survey. Mining and transportation of stone and other natural resources is regarded as the production process of the plants. Maximum production amount of each plant is set based on the production history survey by the corresponding industry organization. Although there have been many qualitative comments on the decrease of road construction demand, almost none has estimated quantitatively. We have estimated the demand according to the statistical survey on road area, road ratio(RS, 2003) and the empirical mixture ratio by interview. The estimation is verified by the production history survey by the industry organization, and on the assumption that, until now, emitted demolished concrete from the site is 98% recycled as roadbed. Demand event of road is set “constant” or “constant decrease” during the simulation. To simplify the situation, amount of buildings’ construction is also set constant after 2003.

3.2 Comparative conditions

Weibull regression based on the survey in Tokyo area by Oikawa et. Al (2002). is used. Enlargement of building lifetime is simulated changing the characteristic lifetime parameter η of Weibull Equation (9).

m

ttw⎭⎬⎫

⎩⎨⎧ −

=ηδ)( ( 9)

Table 3. Comparative condition of building lifetime

Lifetime Equation

Characteristic lifetime

Weibull shape

parameter

Road demand

current 20 60 70 80 90

Weibull regression by Oikawa(2002) and statistics

200

2.974 1%

decrease per year

3.3 Simulation result

Figure 4 shows the estimated amount of land-filled concrete. The estimation exclude backfilled and illegal dumping for comparison with real world statistics. The result have shown that emission of demolished concrete will drastically increase in next 30 years even the full promoted usage of recycled roadbed. Figure 5 shows the estimated result of the amount of land-filled concrete including the backfilled and illegal dumping with a series of building lifetime. On the assumption that building lifetime satisfies Oikawa model, enlargement of 50 year in building lifetime will halve the emission. This scenario is not realistic since the quality of buildings in 1970s are not enough, but at least indicates the possible solution to the run-out of landfill sites.

32 CMT - 271


CMT -

4. CONCLUSIONS

LCA simulation system named “ecoMA” based on multi-agent-system and event-based directed tree model is introduced. The system is shown to be effective in the timeline, geographical and socio-economical evaluation of environmental impacts. The case study has shown that enlargement of building lifetime will make a contribution to the reduction of land-filled concrete. It is also shown that, in the non-realistically ideal situation, recyclable non-backfilled demolished concrete could be zero-emitted. However, the required enlargement of lifetime for effective reduction of demolished concrete is too high for the current maintenance technologies and social circumstances.

5. ACKNOWLEDGEMENT

This research was partially supported by the Ministry of Environment, Grant-in-Aid for Scientific Research K1701, 2005-2007, "Development of Resource-low Simulation System and Optimization Supporting System for the Evaluation of Environmental Impact of Concrete Related Materials", Representative Researcher Takafumi Noguchi , The University of Tokyo.

6. REFERENCES

K. Amano, K. Makita (2000). Journal of Materials, Concrete Structures and Pavement (In Japanese), VII-16 Vol. 657, pp81-89.

J.M Bradshaw (1997), Software Agents, AAAI Press. Building Contractor Society (2002). “Report on the composition and basic units of construction mixed

waste.”(In Japanese). Economic Research Association (2003). “Price Data for Construction Cost Estimating, Apr. 2003.” (In

Japanese) Japan Society of Civil Engineering (2002). “Environmental Evaluation of Concrete.” (In Japanese),

Concrete Engineering Series 62. T. Koshiduka, M. Iri (1999). “Numerical Geometry and Geographical Information.” (In Japanese). Y. Nakamura (2003). “The Research on Concrete Inventory Analysis and its Optimization.” Master

Thesis of Graduate School The Univ. of Tokyo (In Japanese). T. Oikawa, T. Urabe (2002). “Annual Report of Tokyo Metropolitan Research Institute for

Am ount of dem olished concrete

0

500

1000

1500

2000

2500

2000 2010 2020 2030 2040 2050

tim e(year)

Amount(t)

Landfilled

recycled roadbed

Figure 4. Amount of land-filled concrete

( excluding backfilled etc.) Figure 5. Amount of Land-filled and backfilled

concrete

32 CMT - 272


CMT -

Environmental Protection.” pp.182-190 (In Japanese). Ministry of Land Infrastructure and Transport (2004). “Road Statistics.” (In Japanese). H. Shima, H. Tateyashiki, K. Hashimoto, Y.Nishimura, (In Japanese).

32 CMT - 273


CMT -

PBD FOR CONCRETE MATERIAL DESIGN USING BAYESIAN METHOD

Jang-Ho Jay Kim1, Jing Li2, Jongsung Sim3, Sung-Gul Hong4 and Ha-Sun Jeong5

ABSTRACT: Recently, performance-based design (PBD) method has been widely studied as a next generation structural design method. The performance-based design must satisfy the required performance of structures being designed, constructed, and maintained during its life cycle. Performance of structures should be classified by the degree of performance level such as excellent, good, moderate and poor. A suitable target corresponding to each degree can be selected to determine whether safe/fail probability is achieved. One method of deciding the required performance satisfaction is Bayesian method, which has been commonly used in seismic analysis. Generally, it’s presented as a conditional probability of meeting or exceeding some limit state (i.e., collapse) for a given ground motion intensity (i.e., peak ground acceleration (PGA)). But in PBD of material design, the same methodology can be applied to assess concrete material target performance based on some conditional parameters (i.e. load, material, and environmental parameters). Since Bayesian method is a statistical way to estimate the conditional probability, a great number of data points are needed to obtain proper satisfaction curve or envelope. Using these data points and performance criterion, the success/failure status for each data point can be determined. Finally, the satisfaction curve based on degree of quality can be used for a material design of concrete. Since the parameters used in developing satisfaction curve is a case specific parameter and the structure’s degree of quality has to be considered, this method is a performance based material design method. In this paper, a detailed explanation of the procedure of drawing satisfaction curve based on various material parameters is shown. The results show that the required performance can be evaluated by using Bayesian method. KEYWORDS: PBD (Performance Based Design), Bayesian Method, Satisfaction Curve

1. INTRODUCTION

The development of concrete structure design method has experienced several stages. A new design method called PBD is studied as a next generation structural design method. PBD is Performance Based Design, in which structures can be designed for the required performance. In PBD, structures should be classified into the degree of seriousness, such as critical, major, normal and minor. And based on the degree, a suitable target can be selected to determine whether safe/fail probability is achieved. One method of deciding the required performance satisfaction is Bayesian Method, which has been commonly used in seismic analysis. Therefore, in this paper, a detailed explanation is presented to show the procedures of applying Bayesian Method to PBD.

2. PERFORMANCE BASED DESIGN

Concrete structure design standard was first developed in UK in 1902, and gradually evolved into the current design method. Depending on the design methodology, concrete design methods can be divided into 4 classifications as follow (Figure 1): 1. WSD (Working Stress Design Method), which was applied from 1900 to1960, and is in the level of satisfaction of structure serviceability. 2. USD (Ultimate Strength Design Method), which is applied from 1940 to now, and is in the level of satisfaction of structure safety. 3. LSD (Limit State Design Method), which is applied from 1980 to now, and is in the level of satisfaction of structure serviceability, safety, and restorability. 4. PBD (Performance Based Design Method), which is being studied currently, and is being developed to integrate and optimize WSD, USD and LSD, considering all of serviceability, restorability and safety.

1 Associate Professor, Department of Civil and Environmental Engineering, Sejong University, Korea. 2 Master Student, Department of Civil and Environmental Engineering, Sejong University, Korea. 3 Professor, Division of Construction and Transportation Engineering, Ansan Campus, Hanyang University, Korea. 4 Associate Professor, Department of Architecture, Seoul National University, Korea. 5 Director, Korea Concrete Institute Research Center, Korea

33 CMT - 274


CMT -

The performance-based design concept is a concept that tries to satisfy required performance of a structure being designed, constructed, and maintained during its life time. So in PBD, a structure should be designed, constructed, and maintained according to the importance and the required performance.

Since Performance Based Design Method must satisfy the required performance of structures, structures need to be classified into the degree of quality, such as excellent, good, moderate and poor. Then, depending on the determined degree, a suitable target probability of achievement can be selected for design parameters. To determine the design parameter values, Bayesian method can be used to evaluate the probability of satisfaction based on different design parameter values and develop a satisfaction curve or envelope for design usage. This means that the relationship between the design parameter value and structure probability of achievement can be established by Bayesian method in a satisfaction curve. More specifically, by using this satisfaction curve, design parameter value can easily be determined for the target probability (Figure 2). This is the general procedure of PBD using Bayesian Method. This method shows that, in PBD, determination of design parameter values are totally based on the structure seriousness or performance requirement.

3. BAYESIAN METHOD

3.1. Theory background

Generally in Bayesian Method, a satisfaction curve is used as a probabilistic way of assessing the vulnerability of a bridge or any structure under a seismic event, and it is presented as a conditional probability of meeting or exceeding some limit state (i.e., collapse) for a given ground motion intensity (i.e., peak ground acceleration (PGA)), which is developed by Shinozuka. It assumes that the curves can be expressed in the form of two-parameter lognormal distribution functions, and the estimation of the two parameters (median and log-standard deviation) is performed with the aid of the maximum likelihood method. For this purpose, the peak ground acceleration (PGA) is used to represent the intensity of the seismic ground motion. The likelihood function for the present purpose is expressed as follows:

Figure 1．Classification of Design Methods

33 CMT - 275


CMT -

Degree of Quality

(Excellent / Good / Moderate / Poor)

Material and Construction Parameter checking

Completion of Satisfaction Curve by Bayesian Method

Target Value Selected - 10% failure (satisfy) probability - 50% failure (satisfy) probability

Parameter Study

Summating by combining the effect from each parameter

∏=

−−=N

i

xi

xi

ii aFaFL1

1)](1[)]([ (1)

where, F(.) represents the satisfaction curve for a specific state of damage; ai is PGA value to which bridge i is subjected; xi is 1 or 0 depending on whether or not the bridge sustains the state of damage under PGA equal to ai; and N is total number of bridges inspected after the earthquake. Under the current log-normal assumption, F(a) takes the following analytical form:

⎥⎥⎥⎥

⎦

⎤

⎢⎢⎢⎢

⎣

⎡

Φ=ζ

)ln()( c

a

aF (2)

in which, a represents PGA; and Φ[.] is standardized normal distribution function. The two parameters c and ζ in Eq. (2) are computed as ce and ζe satisfying the following equations to maximize ln L:

0lnln==

ζdLd

dcLd (3)

This computation is performed by implementing a straightforward optimization algorithm. In this study, the same methodology has been used for estimation of satisfaction probability of a structure based on material or construction parameters instead of PGA. And, in the process, depending on design seriousness, some threshold parameter value is given as criterion by which the structure is determined whether damage has occurred under certain parameter value, just like the xi, which is mentioned previously. Then a satisfaction curve can be developed based on selected parameters. For PBD, any probability value which is greater then required probability value can be selected as design target probability value. And from the satisfaction curve, the corresponding material or construction parameter value can be determined for a structure design. Especially, required probability value can be

Figure 2. Main procedures of PBD Method

33 CMT - 276


CMT -

based on indirect consideration of social, economy, technology, regional, environment, etc of the structure being designed.

3.2. Procedure of satisfaction curve drawing and examples

Generally, the procedure includes three main parts: ① generation of virtual data points from limited number of actual data points by bell curve implementation, ② determination of success/failure status of each data point by assigned criterion, and ③ completion of final Bayesian curve. A practical example is used to verify the procedure of developing material parameter based satisfaction curve. A satisfaction curve for the workability of concrete mixture based on "unit water" has been performed. A detailed example explaining the procedure of drawing a satisfaction curve based on the parameter "unit water" is discussed. Two types of original datum have been used in the example: (1) water-cement ratio "w/c" and (2) unit water “w”. In this example, "w/c" and "w" are the parameters influencing the workability of concrete mixture content as "slump" measurement. A criterion of slump is used to determine the success/failure status of the workability of the concrete mixture content.

3.2.1. Generation of virtual data points

Bayesian curve is a statistical method of deciding the success/failure of a chosen parameter. Therefore, abundant experimental data points are required for a proper evaluation. However, the experimental data are generally insufficient for rigorous calculation of satisfaction curve. Therefore, in order to overcome this limitation, virtual data are generated from limited number of actual data points by using bell curve implementation. To draw a bell curve, two parameters are needed, mean value and standard deviation. The original data used in this example are shown in Table 1.

w/c (%)

Unit Water(kg/m3)

155 165 175

40

185 155 165 175

50

185 155 165 175

60

185 Using the original data, the mean value and the standard deviation are calculated. The mean value and standard deviation calculation of "unit water" are shown in Eqs. (4) and (5), respectively. 170

12)185175165155(321 =

+++×=

+++=

nxxx

x nL (4)

Table 1. Original data of "w/c" and "unit water"

33 CMT - 277


CMT -

0

1

2

3

4

5

6

7

8

9

130 140 150 160 170 180 190 200 210Unit Water

Num

ber

[ ] 18.1112

)170185()170175()170165()170155(3

)()()(

2222

222

21

=−+−+−+−×

=

−++−+−=

nxxxxxx nL

δ (5)

The data points for the bell curve are generated using mean value and standard deviation of 170 and 11.18, respectively. Table 2 shows the virtual data obtained from bell curve (Figure 3). In Figure 3, x-axis represents "unit water" (w) and y-axis represents the number of virtual data points generated for each "unit water" value.

The same procedure of generating virtual datum of "unit water" has been used to generate "w/c" datum. However, since the bell curve generated data points are based on mathematical equation, the generated data points may lie outside of the physically possible range. Since the data points existing outside of the physically possible range are unusable, they are eliminated in the next analysis stage of developing satisfaction curve. Table 3 shows the generated virtual data points for "w/c". The next step is to combine these two groups of data points as combinations. Here, for combining "w" data points with "w/c" data points, the total number of "w" is set equal to the number of each "w/c" value. For example, when "w/c" is equal to 40.5, the number of data points is 248. So the bell curve for "w" is redrawn with the total number of "w" data points as 248. This procedure can be described as calibrating the curve data points using the Eq. (6).

valueeach w/c ofnumber wofnumber total wofnumber original wofnumber erecalculat ×= (6)

When w is equal to 132.5, a calibrated number of "w" is 0248

1122111

≈×++++++ L

. By calibrating

data points, all of "w/c" data points have corresponding "w" combinations. Table 4 shows the completed (w/c, w) combinations.

w number 131.5 0 132.5 1 133.5 1 134.5 2

… … 205.5 2 206.5 1 207.5 1 208.5 0 total 9958

Table 2. “w” and number of data points

Figure 3. Bell curve of “unit water”

33 CMT - 278


CMT -

3.2.2 Determination of success/failure status of each data point by assigned criterion

The next step in the procedure is to check the status of data points by applying a predetermined threshold criterion to check success or failure: In this case, "w/c" and "w" should be checked with slump criterion to see whether (w/c,w) combination satisfies the success/failure status of slump criterion. In this example, the experimentally obtained curves of slump based on w/c and w (Figure 4) and the threshold value of 15cm slump are used as the slump criterion.

In this criterion, (w/c,w) data points above the 15cm line are success, and the points below the line are failure. For example, when "w/c" is equal to 40.5 and "w" is equal to 169.5, the slump is approximately 13cm, which lies under the line. Therefore, for this (w/c,w) data point, it is determined as failure. In determination of success/failure, "0" stands for success, and "1" stands for failure. This process of checking success/failure is performed for all of data points. Table 5 shows the completed result of success/failure determination for "w".

w number w/c number 132.5 1 40.5 248 133.5 1 41.5 284 134.5 2 42.5 320

… … … … 205.5 2 57.5 320 206.5 1 58.5 284 207.5 1 59.5 248

w/c w number 40.5 132.5 0 40.5 133.5 0 40.5 134.5 0 … … …

41.5 157.5 5 41.5 158.5 6 41.5 159.5 7 … … …

59.5 205.5 0 59.5 206.5 0 59.5 207.5 0

Table 3. Selected data of “w”, “w/c” Table 4. Combined data

Figure 4．“w” and “w/c” – slump curve

33 CMT - 279


CMT -

3.2.3. Completion of final satisfaction curve

To calculate the probability, the cumulative log normal distribution method is used. "LOGNORMDIST" command in "MS Excel" is used to calculate the value of cumulative lognormal distribution. The function requires three parameters: parameter value, mean value, and standard deviation. In our case, x is the parameter "w", "w" data point number, and the success/failure status ("0" or "1") is inputted to obtain the mean value and standard deviation using a computer program that calculates log normal distribution. Figure 5 is the satisfaction curve drawn using obtained mean value and standard deviation. X-axis is "w" and y-axis is the "probability" of success/failure of the task.

Satisfaction curves for “w/c” and “w” by different criterion (15cm, 16cm, 17cm slump) are also shown in Figures 6 and 7, respectively.

w number of data points s or f

139.5 1 1 140.5 1 1 140.5 1 1 140.5 1 1 … … …

198.5 1 0 198.5 1 0 198.5 1 0

w probability 140.5 0.994142228 140.5 0.994142228 … …

162.5 0.469276278 162.5 0.469276278 … …

198.5 0.000130864 198.5 0.000130864

Table 5. Success/failure status of "w"

Table 6. Probability of "w"

Figure 5. Satisfaction curve of “w”

33 CMT - 280


CMT -

Actually, the probability value in satisfaction curve should be a continuous curve form “1” to “0” or “0” to “1” as the parameter increases. However, in some cases, the unrealistic data points are eliminated and the curve becomes a truncated curve between “1” to “0. This is the case for the analysis of “w/c” where the shape of satisfaction curve for “w/c” is a truncated type.

4．SATISFACTION CURVES FOR OTHER PARAMETERS

The method has been used to analyze other concrete material parameters to develop satisfaction curves for PBD.

4.1 Satisfaction curve for strength based on “max aggregate size” and “cement content”

The experimental data of maximum size of aggregate and cement content on the 28-day compressive strength of concretes are analyzed. For satisfaction criterion, 25 MPa of compressive strength is selected as suggested by ACI Standard. The satisfaction curve of 28 day concrete strength for “max aggregate size” and “cement content” are shown in Figures 8 and 9, respectively.

4.2 Satisfaction curve for carbonation depth based on “strength” and “relative humidity”

The experimental data of carbonation depth based on compressive strength of concrete after 2 years of air exposure at relative humidity of 65% and 100% are analyzed. For satisfaction criterion, 5 mm of

Figure 7．Satisfaction curve for “unit water” Figure 6．Satisfaction curve for “w/c”

Figure 8. Strength satisfaction curve of “max aggregate”

Figure 9. Strength satisfaction curve of “cement content”

33 CMT - 281


CMT -

carbonation depth is selected as suggested by ACI Standard. The satisfaction curve of carbonation depth for “strength” and “relative humidity” are shown in Figures 10 and 11, respectively.

4.3 Satisfaction curve of shrinkage for “water content” and “cement content” The experimental data of shrinkage strain as a function of cement content, water content, and water/cement ratio are analyzed. For satisfaction criterion, 100×10-6 of shrinkage strain is selected as suggested by ACI Standard. The satisfaction curve of shrinkage strain for “water content” and “cement content” are shown in Figures 12 and 13, respectively.

5．CONCLUSION

PBD is being developed as a more reasonable design method for a next generation of design method. To implement PBD, Bayesian method is applied. It estimates the material performance based on various material parameters. Using Bayesian method, simple and practical performance estimation of concrete material can be achieved. The method allows the development of satisfaction curves, which can be used for a PBD. From the examples, practical concrete material performance estimations for various parameters are possible and valid.

Figure 11. Depth of carbonation satisfaction curve of “relative humidity”

Figure 10. Depth of carbonation satisfaction curve of “strength”

Figure 13. Shrinkage strain satisfaction curve of “cement content”

Figure 12. Shrinkage strain satisfaction curve of “water content”

33 CMT - 282


CMT -

6. ACKNOWLEDGEMENT

This research was supported by subsidiary research grant given to Korea Concrete Institute from 2005 Construction Core Technology Program D11, “Center for Concrete Corea” from the Ministry of Construction and Transportation. The first author wishes to thank partial financial support from “Bio-Housing research institute” from the Ministry of Science and Technology in Korea. These financial supports are gratefully acknowledged.

7. REFERENCES

M. Shinozuka, M. Q. Feng, H. Kim, T. Uzawa, and T. Ueda (2001). “Statistical Analysis of Fragility Curves.” Department of Civil and Environmental Engineering University of Southern California, Los Angeles, California, 90089-2531

Masanobu Shinozuka, M. Q. Feng, Jongheon Lee, Toshihiko Naganuma (2000). “Statistical Analysis of Fragility Curves.” J. Engrg. Mech., Volume 126, Issue 12, pp. 1224-1231

Jernigan, J. B., and Hwang, H. M. (1997). “Inventory and fragility analysis of Memphis bridges.” Tech. Rep., Ctr. for Earthquake Res. and Information, University of Memphis, Memphis.

33 CMT - 283


CMT -

ANOTHER APPROACH FOR THE PREDICTION OF THE COMPRESSIVE STRENGTH AND ELASTIC MODULUS OF RECYCLED AGGREGATE

CONCRETE

Jongsung Sim1, Cheolwoo Park2, Hongseob Oh3 and Sungjae Park4

ABSTRACT : The reserves of natural aggregate for construction become depleted rapidly. Social and environmental pressures, therefore, on the construction wastes drive greater significance on the recycling of the waste. Particularly, waste concrete is crucial among the construction wastes in terms of conservation of natural construction resources as well as disposal crisis. Since 1970s, studies on the applicability of recycled concrete aggregate and concrete using recycled concrete aggregate have been conducted in many countries, and to some extent practical applications to the field have already initiated. The application of recycled concrete aggregate, however, has been sometimes limited in the construction sites and remained in the low-valued purposes only. The latest progresses on the recycled concrete can be enabled to produce recycled concrete aggregate of which quality is closer to that of natural one. The mechanical properties and durability of recycled aggregate concrete is still lower than that of natural aggregate concrete. Specially, decrease of compressive strength and elastic modulus of recycled aggregate concrete is known as that is due to increment of replacement ratios of coarse and fine recycled concrete aggregate. This study was addressed on the mechanical properties such as compressive strength and elastic modulus of concrete used recycled coarse and fine aggregate. The fundamental characteristics of recycled aggregate were preliminarily analyzed and the adopted recycled aggregate satisfied the standards for quality certification of recycled concrete aggregate in Korea. As the replacement ratio increased, the compressive strength and elastic modulus of recycled aggregate concrete decreased. When the coarse and fine aggregates were completely replaced with the recycled, the compressive strength and elastic modulus were decreased by 13% and 31%, respectively. Based on the test results, the authors suggest the equations for prediction of the compressive strength and elastic modulus of the recycled aggregate concrete considered the replacement ratio. The values from the equations were in good agreement with the test data from this study and others.

KEYWORDS: Coarse and Fine Recycled Concrete Aggregate, Recycled Aggregate Concrete, Compressive Strength, Elastic Modulus

1. INTRODUCTION

The aggregate, component occupying 70% of concrete volume, is known as an important factor deciding the quality of concrete and plays an important role in maintaining the stability of structure. The collectable volume of aggregate is known as 4.4 billion m3 and aggregate of 1.4 billion m3 has been permitted for collection and 30% of this volume has been already collected and used in Korea. According to data provided by the Korea Ready Mixed Concrete Industry Association in 2004, the supply of aggregate is expected to be critically limited in 30 or 40 years in consideration of current 1 Professor, Hanyang University, Korea. 2 Professor, Kangwon National University, Korea. 3 Professor, Jinju National University, Korea. 4 Ph. D. Candidate, Hanyang University, Korea.

34 284


CMT -

annual aggregate demand of about 0.2 billion m3, so the supply and demand of aggregate will be emerged as a critical issue. In this regard, the recycled aggregate acquired from the waste concrete may be considered as the most eco-friendly alternative to solve the supply and demand issue of aggregate in consideration of current annual supply and demand of aggregate. The Korean government has actively and positively developed the technologies required for the application of recycled aggregate by instituting the applicable laws and recycled aggregate quality standard as a part of political effort (Ministry of Construction & Transportation, 2005). However, the recycled aggregate concrete using the recycled aggregate is inferior to the general concrete using the general aggregate in terms of physical, mechanical and endurable properties such as compressive strength and modulus of elastic (J. W. Shim, 2004). Also, this phenomenon is diversely shown according to substitution rate of recycled coarse aggregate and recycled fine aggregate. In this regard, we are to suggest the compressive strength presumption equation and elastic modulus equation according to substitution rate shown when applying the recycled aggregate concrete to the construction site in order to review the change of compressive strength and modulus of elastic known as the most fundamental physical properties in design and analysis of structure when applying the recycled aggregate to the structural concrete in this study.

2. EXPERIMENTAL

2.1 Test Overview In this study, we calculate the variables according to substitution rate of recycled coarse aggregate and recycled fine aggregate as shown in the table 1 in order to presume the compressive strength and elastic modulus of recycled aggregate concrete. The crushed coarse aggregate meeting the requirements for the KS F 2527 (crushed concrete aggregate) and aggregate made by mixing the sea sand with crushed sand are used for the natural coarse aggregate and fine aggregate respectively. The recycled coarse aggregate and recycled fine aggregate meeting the requirements for the recycled aggregate quality standard are used for the recycled aggregate. Currently, the specific policy is not instituted in relation with the mix design of concrete using the recycled aggregate, so no change is considered for the mix design in this study as shown in the table 2. The recycled aggregate is substituted based on the volume and target design strength is set to 45MPa in consideration of application of structural member.

Table 1 Test Variables Used Aggregates Specimens

Coarse aggregate Fine aggregate NN Natural Natural RN Recycled Natural

RR30 Recycled 30%, substituted by Recycled RR60 Recycled 60%, substituted by Recycled

RR Recycled Recycled

34 285


CMT -

2.2 Test Method For the test, we fabricate the cylindrical container to contain the mixed concrete at the size of φ100×200mm and cure the mixed concrete at the temperature of 20±2℃ for 28 days in the water. The test is performed under the condition of compressive strength meeting the requirements for the KS F 2405 (concrete compressive strength test method) and elastic modulus is calculated by measuring the strain rate resulted from the load by means of mounting the strain gages on both sides of cylindrical container.

3. TEST RESULTS AND CONSIDERATION

3.1 Compressive Strength While the strength is reduced about 2MPa when substituting the coarse aggregate by the recycled aggregate, strength is reduced about 6MPa when substituting the natural aggregate by the recycled aggregate. Thus, the strength is reduced when the substitution rate of recycled aggregate is increased. According to results of study, reduction of strength is increased when the substitution rate of recycled aggregate is above 60%, so we may see that the compressive strength of recycled aggregate is considerably changed when the substitution rate of fine aggregate is 30~60%.

35

40

45

50

NN R N R R 30 R R 60 R R

S pecim ens

Compressive strength,M

Pa

Figure 1 Compressive Strength According to Substitution Rate of Recycled Concrete Aggregate

Table 2 Mix Table

Unit volume weight (kg/m3) Mix agent Fine aggregate (kg) Coarse aggregate (kg) Specimens W

(kg) C

(kg) Natural Recycled Natural Recycled HRWRA

(kg) AEA (kg)

NN 189.10 580.49 661.74 0 851.37 0 4.4 0.022 RN 189.10 580.49 661.74 0 0 841.70 4.4 0.022

RR30 189.10 580.49 463.22 184.83 0 841.70 4.4 0.022 RR60 189.10 580.49 264.70 369.66 0 841.70 4.4 0.022

RR 189.10 580.49 0 616.10 0 841.70 4.4 0.022

34 286


CMT -

We are to calculate or presume the reduced compressive strength of recycled aggregate concrete resulted from the substitution rate based on these test results. We suggest the equation [1] as the presumption equation. According to scientific inference method, big volume of test results including, the test results of this study are required for the calculation of presumption equation. However, the properties of recycled aggregate concrete are considerably changed according to quality and mix strength of recycled aggregate and test results required for the suggestion of presumption equation are not sufficient, so we suggest this equation as the basic data for the future studies under the condition of limiting the application of presumption equation to the recycled aggregate concrete showing the mix strength of 30-50MPa and using the high quality recycled aggregate.

ckckr fGSf +−−= 0166.004.0 [1]

Here, ckrf , ckf , G and S indicate the presumed compressive strength (MPa) of recycled aggregate

concrete, compressive strength of normal aggregate concrete (MPa), substitution rate (%) of recycled coarse aggregate and substitution rate of recycled fine aggregate respectively.

3.2 Elastic modulus The figure 2 shows the elastic modulus acquired through the test, and also, this figure shows that the elastic modulus is reduced when the substitution rate of recycled aggregate is increased like the compressive strength. The elastic modulus presumption equation of general concrete suggested by the current concrete structure design standard is as shown in the equation [2], and figure 2 shows the value of elastic modulus calculated according to change of compressive strength.

700,703.0 5.1 += ckc fE ω ( ,30MPafck ≥ 3/500,2~450,1 mkg=ω ) [2]

In this test, the difference of about 5GPa is measured between measured value and presumed value under the condition of using the general aggregate concrete. We suggest the presumed value under the intention of allowing the elastic modulus presumption equation to presume the value close to the actual value using the regression analysis, so the presumed value differs from the results of this study. However, the elastic modulus of recycled aggregate concrete is highly decreased as compared with the one of general aggregate concrete when the substitution rate is increased. In addition, the presumption equation based on the current concrete structure design standard does not reflect these phenomenons. The figure 2 shows that the inclination of elastic modulus presumption equation based on the current concrete structure design standard differs from the one acquired when using the recycled aggregate. Therefore, it is inappropriate to apply the elastic modulus presumption equation of general aggregate concrete to the recycled aggregate concrete. In this regard, we suggest the changed elastic modulus presumption equation applicable to the recycled aggregate concrete as follows:

34 287


CMT -

NN RN RR30 RR60 RR

10

15

20

25

30

35

353841444750

Compressive strength (MPa)

Ela

stic m

odulu

s (

GPa)

Test result

Presumed (KCI)Presumed (This study)Trend line (test result)

Figure 2 Comparison of Elastic Modulus

)9.102.14700,7(03.0 5.1 SGfE ckrcr −−+= ω [3]

( ,30MPafck ≥ 3/500,2~450,1 mkg=ω )

Here, crE ,ω , ckrf , G and S indicate the presumed elastic modulus of concrete using the recycled

aggregate (MPa), unit weight of concrete (kg/m3), presumed compressive strength of recycled aggregate concrete (MPa), substitution rate of recycled coarse aggregate (%) and substitution rate of recycled fine aggregate (%) respectively. In above equation, we consider the elastic modulus reduced according to recycled aggregate substitution rate of recycled aggregate concrete by adding the coefficient on the substitution rate of aggregate to the constant stated in the elastic modulus presumption equation of general concrete. As a result, we may see that the elastic modulus reduction of recycled aggregate concrete is reflected as shown in the figure 2. The table 3 shows the compared values of elastic modulus presumed according to compressive strength of concrete in the current concrete structure standard (when the concrete strength is above 30MPa) like the measured elastic modulus regarding the test variables and equation [2].

Table 3 Comparison of Measured and Presumed Elastic Modulus

Specimens Compressive strength (MPa)

Test elastic modulus (MPa)

Elastic modulus presumed based on the concrete structure design standard(KCI)

(MPa)

Changed presumed elastic modulus

(MPa) NN 45.0 24,185 29,836 29,897 RN 43.3 20,440 29,417 28,058 R30 42.8 17,981 29,292 26,902 R60 41.7 20,000 28,870 25,485 RR 39.1 16,516 28,325 23,881

34 288


CMT -

4. CONCLUSION

This study has been conducted for the purpose of studying the compressive strength and modulus of elastic known as the most fundamental physical properties in design and analysis of structure when applying the recycled aggregate to the structural concrete. 1. The recycled aggregate concrete shows the bigger slump than the general concrete does under the

same mix and pouring conditions. 2. The test results show that the strengths are reduced about 4% and 13% respectively when

substituting the recycled coarse aggregate only and substituting all coarse aggregates and fine aggregates. We suggest the presumption equation enabling to presume the substitution rate of recycled aggregate based on these characteristics.

3. The measuring results of elastic modulus show that the elastic modulus is reduced about 15% and 31% respectively when substituting the recycled coarse aggregate only and substituting all coarse aggregates and fine aggregates. We suggest the presumption equation enabling to presume the elastic modulus based on these measuring results. We may see that the measured value is relatively close to the presumed value by comparing with the existing data.

The presumption equations suggested in this study are expected to be useful for the design and analysis of concrete structural member using the recycled aggregate. Also, if more test data is accumulated, these presumption equations are expected to be more reliable in terms of engineering value.

5. ACKNOWLEDGEMENTS

This research is funded by Korea Institute of Environmental Science & Technology. Authors are thank you for the support of Korea Institute of Environmental Science & Technology. Also, we are very thanks to corps, INSUN ENT for the support of recycled aggregate.

6. REFERENCES

Korea Ready Mixed Concrete Industry Association (2004). “Project intended to solve the unstable aggregate supply and demand issue.” pp. 5~6.

Ministry of Construction and Transportation (2005). “Quality standard for recycled concrete aggregate.” pp. 2~27.

J. W. Shim, M. H. Lee and S. H. Lee (2004). “Study and comparison of different properties of recycled coarse aggregates resulted from the different recycled aggregate production methods.” Journal of Korean Concrete Institute, pp. 195~198.

H. Ch. Lee (2004). “Performance evaluation of recycled aggregate concrete.” Thesis of M. S., Hanyang University, pp. 35~54.

34 289

Documents

2nd Asian Concrete Federation Conference_Bali_Nov 20-21 2006