Creep Behavior and Its Prediction for Normal Strength Concrete Made

Creep Behavior and Its Prediction for Normal Strength Concrete Made from Crushed Clay Bricks as Coarse Aggregate Syed Ishtiaq Ahmad, A.M.ASCE1 and Sushanta Roy2 Abstract: To study the effect of crushed clay bricks as coarse aggregate on creep behavior of concrete, a comprehensive testing program was conducted. Concrete cylinder specimens having characteristic or specified compressive strength of 17.2, 24.0, and 27.5 MPa were prepared from both natural stone and crushed clay brick aggregate. Mix design ratios were evaluated in a way so that volumetric content of coarse aggregate, both brick and stone, remained same for all concrete samples. Specimens were subjected to creep testing at the 7th and 28th day after casting and creep strain data were recorded up to 300 days. Results show that although strength and other environmental parameters remain the same, concrete made from crushed clay brick as coarse aggregate have a higher creep strain than that of concrete made from natural stone aggregate. This increase in creep strain ranges from 30% to as high as 45% for the 300-day loading history considered. Additionally, to select an appropriate model to predict creep in brick aggregate concrete, the effectiveness of five widely used prediction models were examined. Predicted creep strain from ACI 209R, CEB-FIP, B3, GL2000, and Eurocode 2 models were compared with experimental results. By using statistical analysis, the authors established that prediction of creep by GL2000 model is closest to the experimental results. Finally, a modification factor has been proposed that may be incorporated so that prediction of creep strain by the GL2000 model in brick aggregate concrete becomes more realistic. DOI: 10.1061/(ASCE)MT.1943-5533.0000391. © 2012 American Society of Civil Engineers. CE Database subject headings: Concrete; Clays; Bricks; Aggregates; Creep; Numerical models; Time dependence. Author keywords: Concrete; Clays; Bricks; Aggregate; Creep; Numerical models; Time dependence.

Introduction In countries such as Bangladesh and parts of India, where natural stone is scarce and hence expensive, crushed burnt clay bricks are used extensively as an economic alternative coarse aggregate in preparation of concrete. Here, concrete prepared from brick aggregate is commonly used for construction of up to 6-story buildings, rigid pavements, and small-and medium-span bridges and culverts (Mansur et al. 1999). Additionally, in regions where natural stones are abundant, a survey has shown that approximately 5 to 10% of bricks manufactured in modern automated factories are rejected because of nonconformity with relevant specifications (Mansur et al. 1996). Utilizing these bricks as coarse aggregate will provide a good use of otherwise waste materials. Further, aggregate production from such bricks may be an effective way to make the most of energy and efforts spent on their production. These economic and environmental issues have led to increasing attention and research in properties of brick aggregate and concrete made from it. Recent successful studies on the use of crushed bricks as aggregate in concrete have been reported from several parts of the world. Akhtaruzzaman and Hasnat (1983) carried out research by using well-burned brick as coarse aggregate in concrete in which 1

Associate Professor, Dept. of Civil Engineering, BUET, Dhaka, Bangladesh (corresponding author). E-mail: [email protected] 2 Graduate Student, Dept. of Civil Engineering, BUET, Dhaka, Bangladesh. Note. This manuscript was submitted on February 19, 2011; approved on September 14, 2011; published online on September 16, 2011. Discussion period open until August 1, 2012; separate discussions must be submitted for individual papers. This paper is part of the Journal of Materials in Civil Engineering, Vol. 24, No. 3, March 1, 2012. ©ASCE, ISSN 08991561/2012/3-308–314/$25.00.

they found that it was possible to achieve concrete of high strength using crushed brick as the coarse aggregate. Khalaf (2006) determined several physical and mechanical properties of fresh and hardened concrete produced from crushed clay brick aggregate and compared those with concrete produced by using granite aggregate. Cachim (2009) found that brick aggregate could be used as partial replacement of natural aggregates in concrete without reduction of concrete properties for up to 15% replacement, and with reductions up to 20 to 30% replacement. Debieb and Kenai (2008) showed that it is possible to manufacture concrete containing crushed bricks (coarse and fine) with characteristics similar to those of natural-aggregate concrete provided that the percentage of brick aggregates is limited to 25 and 50% for the coarse and fine aggregate, respectively. Most of the studies noted previously considered strength, workability, and modulus of elasticity as the main parameters for comparison except Debieb and Kenai (2008), who included permeability and shrinkage properties of concrete produced from partially replaced brick aggregate. In a recent paper, Domingo et al. (2010) studied long-term deformation by creep and shrinkage for concrete in which natural aggregate was partly substituted by recycled aggregate from waste concrete. Nevertheless, very few works so far have comprehensively dealt with creep behavior of concrete produced from newly crushed clay brick as coarse aggregate. There is no extensive database nor any available model to predict the extent and nature of creep for concrete produced from brick aggregate. With these as background, the present paper examines the creep behavior for normal strength concrete produced from brick aggregate. For this purpose, a comprehensive testing program was undertaken at Bangladesh University of Engineering and Technology, Dhaka, Bangladesh, in which three different types of normal strength concrete were prepared from both brick and

308 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / MARCH 2012

Downloaded 04 Apr 2012 to 123.49.42.194. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org

Cumulative Percent (%) Retained

natural stone aggregate. These concrete samples were then tested for creep for up to 300 days. The difference in creep behavior between concrete produced from brick and stone aggregate was first observed. Then a thorough study was conducted to examine the effectiveness of five widely used creep prediction models, namely ACI 209R (ACI 1992), CEB-FIP90 (CEP-FIP 1994), B3 (RILEM 1995a, b), GL2000 (Gardner and Lockman 2001a, b, c), and Eurocode 2 (ECS 2002) in predicting creep for brick aggregate concrete. From these studies, the model that best fits the creep behavior of brick-aggregate-made concrete was identified.

100.0 90.0 80.0 70.0

Stone Aggregate

60.0 50.0

Brick Aggregate

40.0

ASTM C 33 (U B)

30.0 ASTM C 33 (L B)

20.0 10.0 0.0

Materials Used ASTM Standard Sieve

Cement

Fig. 1. Grading of natural stone and brick aggregate

Ordinary Portland cement (Type 1) having 28 days compressive strength of 46 MPa (ASTM 1994) was used for preparation of all concrete samples. By using one type of cement, the effect of varying the types of coarse aggregate in concrete was investigated.

28 days, depending on the age at which they were subjected to different tests.

Fine Aggregate

Concrete Mix Design

One type of natural coarse sand was used throughout the experimental work so as to keep the fine aggregate parameter constant. Sieve analysis was carried out in accordance with ASTM C136-06 (ASTM 2006). The results of this analysis showed that the sand used fitted within the limits set out in ASTM C33-03 (ASTM 2003).

The procedure for design of concrete mixes with normal aggregate can be used to design mixes using crushed brick aggregate (Khalaf 2006). In this work, mix design ratios for both stone and brick aggregates with target strength of 17.2, 24.0, and 27.5 MPa were evaluated from the ACI method (ACI 2002). Design ratios were evaluated by keeping a constant volume for both stone and brick aggregate so that the variation in creep behavior of concrete attributable to parameters other than the properties of coarse aggregates remains minimal. Mix ratios thus evaluated were converted to equivalent weight and are shown in Table 1. Table 2 shows strength test results on concrete samples prepared by using mix ratios of Table 1. From these results, it is seen that except for one instance (30.2 MPa in case of 27.5), the strength

Clay Bricks Bangladesh standard BDS 208:2002 (BSTI 2007) classifies bricks into three categories depending on their use. Of these three, type “S” is normally used for aggregate production and the same has been employed in this work. Before the new bricks were crushed down into coarse aggregate, their compressive strength was measured as per ASTM C67-03a (ASTM 2009) and was found to be 31.9 MPa which is above 27.5 MPa, the minimum limit set in BDS 208-2002 (BSTI 2007) for Type S bricks. Moisture content of the bricks was found to be 13.2%.

Table 1. Concrete Mix Design (weight basis)

Clay Brick and Stone Aggregate Brick aggregate was produced by using a hammer to break down whole new bricks on a solid concrete surface. Natural sandstone boulders crushed by a stone crusher were used as stone aggregate. For comparison purposes, bricks and stone boulders were crushed in a way so that they possessed similar gradation and approximately the same fineness modulus (FM) to negate the effect of size and shape on creep behavior of concrete. Additionally, it was also ensured that grading limits set out in ASTM C33-03 (ASTM 2003) were strictly maintained. Size distribution and gradation of both types of aggregates are shown in Fig. 1, from which the FM of stone and brick aggregate were found to be 8.3 and 7.9, respectively.

Brick aggr.

Stone aggr.

w=c

OPC

Coarse aggr.

Fine aggr.

Water

17.2 24.0 27.5 17.2 24.0 27.5

0.55 0.42 0.37 0.58 0.40 0.38

26.4 30.0 34.3 22.8 30.0 34.3

87.1 87.1 87.1 81.3 81.3 81.3

67.2 62.9 59.5 68.4 64.9 60.5

14.6 12.7 12.7 13.2 12.0 12.5

Note: Quantity for single batch: 12 cylinders per batch.

Table 2. Concrete Compressive Strength Test

Experimental Program The experimental work consisted of a concrete compressive strength test [ASTM C39/C39-M05 (ASTM 2005)], modulus of elasticity evaluation [ASTM C469-02 (ASTM 2002a)], and creep testing [ASTM C512-02 (ASTM 2002b)]. Concrete having three characteristics or a specified strength of 17.2, 24.0 and 27.5 MPa was considered in this study. For every set of concrete with a particular characteristic strength, a total of 24 specimens consisting of 300 mm × 150 mm cylinders were cast in two batches. All the testing specimens were moist cured for a duration of 7 or

Weight (kg)

Target strength (MPa)

Brick aggr.

Stone aggr.

Compressive strength (MPa)

Modulus of elasticity (GPa)

Target strength (MPa)

7 days

28 days

28 days

17.2 24.0 27.5 17.2 24.0 27.5

12.4 17.6 19.4 12.3 17.7 21.2

17.8 25.1 27.7 17.7 25.2 30.2

21 25 26 21 27 29

JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / MARCH 2012 / 309


Relative Humidity (%)

60 55 50 45 40 35 Time (Days)

Fig. 2. Relative humidity data at creep testing machine

Creep Testing For measurement of creep, specimens were loaded at 35% of their stress/strength ratios inside the testing frame at the age of 7 and 28 days after casting. Additionally, three concrete cylinders from every representative strength category were left just beside the creep testing machine as control specimen. After loading, both control and loaded concrete cylinders were kept under room temperature at 28–4°C. Relative humidity in the laboratory was also recorded for the entire loading period and is shown in Fig. 2. The strain values of loaded specimen and those of control specimens were measured by using a 200-mm strain gauge. Strain in the control specimen is, in effect, because of shrinkage, temperature, and other secondary causes except creep. For this work, strain readings of up to 300 days have been considered. The magnitude of creep was calculated by subtracting the strain of the control specimens and elastic deformation from the total deformation of loaded specimens.

relations have a similar pattern for both types of concrete. There is an initial steep increase in creep strain, after which, the rate of increase remains approximately constant throughout the 300-day time period. Nevertheless, in brick aggregate concrete, creep strain is always higher than that in stone aggregate concrete of same strength. Initially, creep strain in brick aggregate concrete is approximately 45% higher approximately that in stone aggregate concrete. As the loading age progresses, this difference narrows and at the time of loading at 300th day, becomes approximately

2000 1800 1600

Creep Strain (micron)

achieved is fairly close and lies within 5% of the target strength. Therefore, it may be concluded that the applied mix ratios were appropriate and justified. Results in Table 2 also indicate that strength rather than types of aggregate used is the main determining factor for modulus of elasticity of concrete. This is in accordance with findings of other researchers e.g., Cachim (2009).

1400 1200 1000 800 Brick Aggregate

600

Stone Aggregate

400 200 0 0

Results and Discussion

50

100

150

200

250

300

Time (Days)

Figs. 3–8 show creep strain for both brick and stone aggregate concrete of three different strengths and loaded at 7th and 28th day after casting. From these graphs, it is apparent that creep strain-time

Fig. 4. Comparison of creep strain for 24.0 MPa concrete loaded on 7th days

2000 2000 1800 1800 1600



1600 1400 1200 1000 800 600

Brick Aggregate Stone Aggregate

400

1400 1200 1000 800


600 400 200

200

0

0 0

50

100

150

200

250

0

300


50

100

150

200

250

300

Time (Days)

Time (Days)




2000

32%. Furthermore, creep strain decreases with the increasing strength of concrete. Interestingly, the ratio of this decrease is found to be approximately the same for both types of concrete. Age at which specimen is loaded affects creep strain behavior during the initial days of loading only. However, at the 300th day loading, specimens of identical strength show almost the same creep strain irrespective of the loading age.

1800


1600 1400 1200 1000

Comparison between Experimental and Predicted Creep Strains

800 Brick Aggregate

600

Stone Aggregate

400 200 0 0

50

100

150

200

250

300

Time (Days)


2000 1800


1600 1400 1200

Widely used models to predict creep such as ACI 209R, CEBFIP90, B3, GL2000, and Eurocode 2 do not include parameters that can clearly distinguish properties of coarse aggregates. However, from the preceding discussions, it is apparent that the type of aggregate influences creep behavior of concrete markedly. Therefore, it is of interest from designer’s perspective to see which model can eventually predict creep strain for concrete made from brick aggregates. For this purpose, experimental results have been compared with predictions using different models. The experimental and predicted results obtained from the ACI 209R, CEB-FIP90, B3, GL2000, and Eurocode 2 models for brick aggregate concrete of strength 17.2, 24.0, and 27.5 MPa loaded at the 7th and 28th days are shown in from Fig. 9–14. For comparison purposes, creep coefficient has been evaluated from experimental results and from all the creep prediction models under consideration. Creep coefficient from experimental data was calculated using the following formula: ϕ ¼ Ecσεc c, in which ϕ ¼ c Creep coefficient Ec ¼ Modulus of elasticity of concrete

1000 800

εc ¼ Creep strain


600

σ ¼ Applied stress

400 200 0 0

50

100

150

200

250

300

Time (Days)


As these figures show, creep coefficients evaluated from the five models examined in this work differ from experimental creep coefficient considerably. The creep coefficient predicted by the ACI 209R model gives a good approximation of experimental behavior up to 40 days; after that it always overpredicted the creep strain. The Eurocode 2 and B3 models always underpredicted the creep deformation and in the case of Eurocode 2 model, greater

12.00

2000

Experimental Value EC2 Model GL 2000

1800

CEB-FIP 90 ACI 209 B3 Model

10.00 1400

Creep Coefficient


1600

1200 1000 800


600 400

8.00

6.00

4.00

2.00

200 0

0.00 0

50

100

150

200

250

300

0

50

100

150

200

250

300

Time (days)

Time (Days)


Fig. 9. Comparison of creep coefficient for 17.2 MPa concrete loaded on 7th day



12.00

12.00 Experimental Value EC2 Model GL 2000


8.00

8.00

Creep Coefficient

10.00

Creep Coefficient

10.00

6.00

4.00



6.00

4.00

2.00 2.00 0.00 0

0.00 0

50

100

150

200

250

50

100

300

150

200

250

300

Time (days)

Time (days)


12.00



12.00


10.00

Creep Coefficient

10.00

Creep Coefficient


8.00

6.00

4.00


8.00

6.00

4.00

2.00 2.00

0.00 0

0.00 0

50

100

150

200

250

300

50

100

150

200

250

300

Time (days)

Time (days)


deviation from the experimental value is observed. For relatively low strength concrete (17.2 MPa), the CEB-FIP90 model initially underestimates creep strain, after which it overestimates. In the case of higher strength concrete (27.5 MPa), it underestimates creep deformation for the entire time period. Initially, the GL2000 model shows good agreement with the experimental results for all categories of concrete. After approximately 20 days, it starts to underpredict creep strain. However, the amount of underprediction by the GL2000 model has been found to be far less than those of Eurocode2 or B3 model.

12.00 Experimental Value EC2 Model GL 2000

Creep Coefficient

10.00


8.00

6.00

4.00

2.00

0.00 0

50

100

150

200


250

300

Time (days)


Ranking of Prediction Models Among the numerous methods of analyzing experimental data to determine the best prediction models, the residuals squared and error percentage methods have been used in this study (Kendall et al. 1973). “Residual” is the difference between the model and experimental value. Residual is a useful tool to identify whether a prediction model is overpredicting or underpredicting at a given time. Residual squared method is the summation of residuals squared for all data points. The model with the smallest value



Table 3. Statistical Performance of Prediction Models Age at loading (days)

CEB-FIP

EC2 model

ACI 209R

GL 2000

B3 model

Str. (MPa)

Err. (%)

R sqr.

Err. (%)

R sqr.

Err. (%)

R sqr.

Err. (%)

R sqr.

Err. (%)

R sqr.

17.2 24 27.5 17.2 24 27.5

30.7 42.0 45.1 33.3 43.5 48.1

42.9 91.8 104.1 27.1 66.5 94.3

64.7 73.6 75.0 59.7 74.3 76.4

163.7 274.7 289.5 75.8 203.0 237.3

55.1 45.9 45.4 74.5 45.7 54.5

172.7 131.4 128.3 198.5 102.4 164.0

20.4 29.4 29.7 14.1 32.8 36.1

29.4 62.0 63.8 9.3 58.7 68.6

38.4 55.3 57.9 19.5 49.2 54.3

75.0 178.6 196.3 14.09 110.8 138.5

7

28

Table 4. Overall Ranking of Prediction Models Age at loading (days)

CEB-FIP

EC2 model

ACI 209R

GL 2000

B3 model

Str.(MPa)

Err. (%)

R sqr.

Err. (%)

R sqr.

Err. (%)

R sqr.

Err. (%)

R sqr.

Err. (%)

R sqr.

17.2 24 27.5 17.2 24 27.5 Sum

2 2 2 3 2 2

2 2 2 3 2 2

5 5 5 4 5 5

4 5 5 4 5 5

4 3 3 5 3 4

5 3 3 5 3 4

1 1 1 1 1 1

1 1 1 1 1 1

3 4 4 2 4 3

3 4 4 2 4 3

7

28

Overall ranking

26 2

57 5

indicates the best prediction model. In contrast, error percentage is calculated as: Error Percentage ¼ Residuals 100=Experimental Value × ðat a given timeÞ The average error percentage for all the date points is evaluated and a smaller error percentage indicates a better model. Table 3 shows the residual squared and error percentage of creep coefficient for concrete of 17.2, 24.0, and 27.5 MPa strength loaded at the 7th and 28th days. The ranking of prediction models on the basis of residual squared and error percentage analysis is shown in Table 4. It shows that two different analyses provide almost the same results, i.e., deviation from the observed and predicted strain is approximately the same from the two methods used. Furthermore, as was evident from graphical representation, none of the models give an accurate approximation of the observed creep behavior with deviation varying from 15% (GL2000) to as high as 75% (ACI 209R and Eurocode 2). For ranking of models, the following procedure was used: The model that gives the closest result to the observed creep strain is given the first ranking. Subsequent close approximations are ranked as 2, 3, and so on. Ranking is done for all three different strengths of concrete loaded at the 7th and 28th days after casting. Finally, all these rankings have been summed and the best model is indicated by the lowest value of these summations. From Table 4, note that GL2000 is the best

Table 5. Multiplication Factor for Creep (α) Age of loading

Concrete strength (N=mm2 )

7 days

28 days

17.2 24 27.5

1.2903 1.4535 1.4615

1.1822 1.5433 1.6126

45 4

12 1

40 3

model to predict creep coefficient for normal strength concrete made from brick aggregate, whereas CEB-FIP is the second-best predictor followed by B3, ACI-209R, and Eurocode 2. Proposed Modification Factor for GL2000 Model From the preceding analysis, it is apparent that none of the models considered in this work accurately predicts the observed creep behavior. Analysis results also show that prediction of the GL2000 model is closest among the five models considered. However, to estimate creep deformation more closely by using the GL2000 model, a modification is required to the existing equation. For this, a multiplication factor “α” has been derived by using the correlation method (Kendall et al. 1973), according to which, the slope of the best fit straight line between experimental and model values is termed as “α”. Suggested multiplying factors for three different strengths of concrete are shown in Table 5. From the results presented, it may be seen that the multiplying factor “α” for the existing equation of the GL2000 model varies from 1.18 to 1.62, in which the higher value (1.62) is associated with higher strength concrete (27.5 MPa) and vice versa. However, to be “α” conservative “α”, a multiplying factor of 1.7 may be considered as appropriate for normal strength concrete made from crushed clay bricks as coarse aggregate.

Conclusion The experimental results clearly show that although strength and other environmental parameters are identical, the concrete specimen made from crushed clay bricks as coarse aggregate have 30 to 45% higher creep strain compared with those made from natural stone aggregate. Results also shows that creep strain/time behavior have patterns that are identical in both types of concrete. Additionally, observed creep behavior in brick aggregate concrete is compared with five creep prediction models, namely ACI 209R, CEB-FIP90, B3, GL2000, and Eurocode 2, that are widely used.



From the analysis, it is shown that none of the models predicts the experimental creep strain of brick-aggregate-made concrete accurately. However, comparative study using statistical parameters such as residual squared and error percentage methods show that GL2000 is the best model that predicts the observed creep strain with some degree of accuracy. Again, to estimate creep strain more accurately, a multiplying factor “α” has been proposed for GL2000 equation, the best predictor model of observed creep behavior. Statistical tools such as the “correlation method” show that the value of “α” may be taken as 1.7 to the be on conservative side.

References Akhtaruzzaman, A. A., and Hasnat, A. (1983). “Properties of concrete using crushed brick as aggregate.” ACI Concr. Int., 5(2), 58–63. American Concrete Institute (ACI). (1992). “Prediction of creep, shrinkage and temperature effects in concrete structures.” ACI 209R, Famington Hills, MI. American Concrete Institute (ACI). (2002). “Standard practice for selecting proportions for normal, heavyweight and mass concrete.” ACI 211.1-91, Famington Hills, MI. ASTM. (1994). “Specification for Portland cement.” ASTM C150-94, West Conshohocken, PA. ASTM. (2002a). “Standard test method for static modulus of elasticity and poisson’s ratio concrete in compression.” ASTM C469-02, West Conshohocken, PA. ASTM. (2002b). “Standard test method for creep of concrete in compression.” ASTM C512-02, West Conshohocken, PA. ASTM. (2003). “Standard specifications for concrete aggregate.” ASTM C33-03, West Conshohocken, PA. ASTM. (2005). “Standard test method for compressive strength of cylindrical test specimen.” ASTM C39M-05, West Conshohocken, PA. ASTM. (2006). “Standard test method for sieve analysis of fine and coarse aggregates.” ASTM C136-06, West Conshohocken, PA. ASTM. (2009). “Standard test method for sampling and testing brick and structural clay tiles.” ASTM C67-03a, West Conshohocken, PA.

Bangladesh Standards and Testing Institution (BSTI) (2007). “Specification for common building clay bricks.” BDS 208:2002, Dhaka, Bangladesh. Cachim, P. B. (2009). “Mechanical properties of brick aggregate concrete.” Constr. Build. Mater., 23(3), 1292–1297. CEB-FIP. (1994). “CEB-FIP model code 1990: Design code 1994.” Thomas Telford, London. Debieb, F., and Kenai, S. (2008). “The use of coarse and fine crushed bricks as aggregate in concrete.” Constr. Build. Mater., 22(5), 886–893. Domingo, A., Lázaro, C., Gayarre, F. L., Serrano, M. A., and LopezColina, C. (2010). “Long term deformations by creep and shrinkage in recycled aggregate concrete.” Mater. Struct., 43(8), 1147–1160. European Committee for Standardization. (2002). “Eurocode 2: Design of concrete structures—Part 1: General rules and rules for buildings.” prEN 1992-1-1, Brussels, Belgium. Gardner, N. J., and Lockman, M. J. (2001a). “Design provision for drying shrinkage and creep for normal-strength concrete.” ACI Mater. J., 98(2), 159–167. Gardner, N. J., and Lockman, M. J. (2001b). “Discussion on ‘Design provision for drying shrinkage and creep for normal-strength concrete.’” ACI Mater. J., 99(1), 111. Gardner, N. J., and Lockman, M. J. (2001c). “Errata: Discussion on ‘Design provision for drying shrinkage and creep for normal-strength concrete.’” ACI Mater. J., 99(1), 112. Kendall, Y., Maurice, R., and Stuart, A. (1973). The advanced theory of statistics. Vol 1 & 2, Griffin Publication, London. Khalaf, F. M. (2006). “Using crushed clay brick as coarse aggregate in concrete.” J. Mater. Civ. Eng., 18(4), 518–526. Mansur, M. A., Wee, T. H., and Lee, S. C. (1996). “Crushed bricks as coarse aggregate for concrete.” Concrete in the service of mankind, Chapman and Hall, London, 505–514. Mansur, M. A., Wee, T. H., and Lee, S. C. (1999). “Crushed bricks as coarse aggregate for concrete.” ACI Mater. J., 96(4), 478–484. RILEM TC-107-GCS. (1995a). “Creep and shrinkage prediction model for analysis and design of concrete structures—Model B3.” Mater. Struct., 28(6), 357–365. RILEM TC-107-GCS. (1995b). “Errata: Creep and shrinkage prediction models—Model B3.” Mater. Struct., 29(2), 126.

