Professional Documents
Culture Documents
net/publication/282450118
CITATIONS READS
13 8,616
2 authors, including:
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Performance of Hot Mix Asphalt Mixtures Made of Recycled Aggregates View project
All content following this page was uploaded by Muhammad Murtaza Nasir on 20 April 2016.
DOI: 10.1520/JTE20140139
Reference
Rafi, Muhammad Masood and Nasir, Muhammad Murtaza, “Models for Prediction of 28-Day Concrete
Compressive Strength,” Journal of Testing and Evaluation, Vol. 44, No. 3, 2016, pp. 1–12, doi:10.1520/
JTE20140139. ISSN 0090-3973
ABSTRACT
Manuscript received March 28, 2014; The design codes usually specify 28-day concrete compressive strength as design strength
accepted for publication November 13,
for reinforced concrete (RC) structures. Concrete specimens are cast and tested at 28 days
2014; published online February 2, 2015.
to ensure compliance of concrete strength with the design requirements. Prediction of
1
Professor, Department of Earthquake
concrete strength can help in reducing waiting time and can result in speeding up
Engineering, NED Univ. of Engineering
and Technology, Karachi-75270, construction activities. This paper presents prediction models for concrete compressive
Pakistan, e-mail: rafi-m@neduet.edu.pk strength up to 28 days. The data of experimentally tested concrete cylinders were employed
2
Lecturer, Department of Civil in the development of these models. The effects of cement chemical composition and
Engineering, Sir Syed Univ. of fineness were included by defining two parameters in the models. The predictions are based
Engineering and Technology,
Karachi, Pakistan, on 7-day concrete strength. The proposed models provided good correlation with the
e-mail: mmuhammad5790@hotmail.com observed concrete strength data. The models were also validated using the strength results
of concrete mixes in the available literature. Generalized forms of the models have been
suggested for cement brands available in Pakistan.
Keywords
concrete, strength, cement, cylinders, cubes, prediction model
Introduction
Concrete is one of the popular construction materials around the world and is employed in both
structural and non-structural applications. It is a strong material in compression and is, therefore,
used to resist compressive stresses. The properties of concrete are generally co-related with its
compressive strength; better compressive strength implies enhanced concrete properties. The
theory of reinforced concrete (RC) design is based on the 28-day concrete compressive strength,
Copyright V
C 2015 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959. 1
2 Journal of Testing and Evaluation
which is determined by testing the specimens in the laboratory. (2 brands), (2) Power cement, (3) Falcon cement, (4) Thatta
ASTM C192/C192M-02 [1] recommends the use of cylindrical cement, (5) State cement, and (6) Pakland cement. All these
specimens for this purpose. are Type-I ordinary Portland cements (OPC) [32]. These
In order to ensure that the concrete strength satisfies the brands were designated by letters from A to G, which were
design requirements, several trial mixes are made in the labora- randomly assigned to them to keep the data of testing anony-
tory and are tested at 28 days. This is a time-taking process, mous. Tests for determining chemical composition and fine-
which requires long waiting time and may cause delays in the ness of grinding of cement were carried out in the laboratory
project execution. In addition, estimate of concrete strength at using the relevant ASTM standards. The results indicated that
different time intervals is required to plan the removal of form- cement brand A has the highest value of fineness compared to
work and subsequent construction activities. Several attempts the other brands, followed by cement brand B. On the other
have been made by the researchers to develop strength predic- hand, cement brand G indicated the lowest value of fineness. It
tion models for concrete to keep a check on its quality and its was noted that this brand also contains the least amount of sil-
conformity with the design specifications. The suggested models icates (tri-calcium silicate (C3S) and di-calcium silicate (C2S))
were based on different approaches, which employ cement (67 %). The silicates were found to be the highest in cement
chemical and compound composition [2–6], specified age of brand B (75 %). All the aforementioned cement brands (A–G)
concrete [7–9], empirical approach [10–15], or water-cement were also employed for casting and testing of concrete cylin-
(w/c) ratio [16–19]. Furthermore, studies related to prediction ders and mortar cubes for the studies presented. The results of
models for concrete strength have also been based on computa- chemical analysis and fineness were used to understand the
tional modeling, statistical techniques and artificial neural net- differences in the mechanical properties of the employed
works [20–31]. However, most of these models have limited cements. Note that the limitation on the identity of cement
application for the construction industry and are heavily brands does not affect the significance of presented studies, as
research oriented; these models employ several parameters and discussed in the forthcoming sections.
coefficients that require separate efforts of determination. In
addition, the models are intended to determine the long-term
strength over several years and cannot be used directly to esti- Experimental Program
mate 28-day strength. The experimental program consisted of casting and testing of
This paper presents the studies that were carried out to concrete cylinders and mortar cubes. The cylinders were tested
develop analytical models for the prediction of 28-day concrete both in compression and tension, whereas mortar cubes were
compressive strength. Cylindrical concrete specimens were tested in compression. The details of these have been presented
experimentally tested in the laboratory, both in compression in the following sections.
and tension. Mortar cubes were also tested in compression to
supplement (qualitatively) results of concrete cylindrical
Materials
strength. The data of concrete compressive strength were
employed in the development of analytical models; these were AGGREGATE
validated using the strength results from the presented study Normal weight crushed stone coarse aggregates (CA) (19 mm
and those available in the literature. Attempts were made to (0.75 in.) size) and fine aggregates (FA) (sand) were used in the
keep the models simple so that these could be conveniently concrete mix. The properties of aggregates were determined
used by the professionals and practitioners. The use of models using the relevant ASTM standards and are given in Table 1.
avoids waiting for 28 days to determine the concrete strength. Sieve analysis was carried out for both aggregate types, which
As a result, project managers and project engineers may become complied with the requirements of ASTM C33-03 [37].
able to better plan construction activities. Although the pre-
sented studies are in the context of cement brands available in CEMENT
Pakistan, the proposed models are generic in nature and can be As mentioned earlier, seven locally available cement brands
applied in any region of the world. were used in this study for preparing mortar and concrete
mixes. All these cements were equivalent ASTM C150 Type I
OPC [32]. Of the employed cement brands, Lucky cement
Review of Previous Work samples were provided by the manufacturer, whereas the rest
of the cement bags were purchased from the local market. The
by Authors cement bags were carefully stored and were covered by the
The authors carried out studies to investigate the chemical and plastic sheets in the laboratory. The cement was inspected reg-
physical properties of cements locally available in Pakistan. ularly to ensure that its quality is not compromised due to
The employed cement brands include: (1) Lucky cement lump formation.
RAFI AND NASIR ON MODELS FOR CONCRETE STRENGTH 3
Mix Type Cement (Kg/m3) Sand (Kg/m3) Coarse Aggregate (Kg/m3) w/c Ratio Admixture (Liters/m3)
M21 315 804 1135 0.53 —
M34 450 689 1098 0.40 2.25
M48 570 479 1098 0.40 2.17
4 Journal of Testing and Evaluation
Tensile strength of concrete was also determined by carry- On a qualitative scale, the trend of 28-day strength of con-
ing out splitting tests on concrete cylinders. Four cylinders were crete cylinders corresponds to the mortar cube strength for all
tested at 28 days and an average value was used as concrete the cement brands employed.
tensile strength.
TENSILE STRENGTH
Direct tension tests are not reliable for predicting the tensile
Discussion on Test Results strength of concrete, due to stress concentrations in the grip-
MORTAR CUBES
ping devices. Therefore, an indirect tension test, usually known
Figure 1 illustrates the results of compressive strength of mortar as splitting test, is performed. The splitting strength of concrete
cubes. The cubes were tested at 7 and 28 days. It is seen in Fig. 1 (ft) can be related to its compressive strength; this is estimated
that the 28-day strength of mortar made with cement brand B as 10 % of compressive strength up to 41 MPa and 9 % for
is larger than the other cement brands. On the other hand, the higher compressive strength of concrete [39].
7-day strength of cement brand A is larger compared to the rest Splitting cylinder tests were performed on the concrete
of the cement brands. The 7- and 28-day strengths of cement cylinders at an age of 28 days to determine concrete tensile
brands C–F are close to each other. Cement brand G yielded the strength. Four cylinders were used for each of the concrete mix
lowest strength at both the aforementioned test ages. employed in this study and an average strength was calculated.
The results of average tensile strength are plotted in Fig. 3 for
the three employed strength classes.
Concrete Cylinders It is seen in Fig. 3 that, similar to compressive strength, the
tensile strength of concrete made with cement brand B is higher
COMPRESSIVE STRENGTH than the concrete made with other cements, for each concrete
Figure 2illustrates the results of average compressive strength of mix type. Furthermore, cement brand G has the lowest tensile
concrete made with each cement brand. A similar trend is strength. As mentioned earlier, this cement also yielded the low-
observed in Figs. 2(a)–2(c). The 7-day strength of concrete est concrete compressive strength. This trend is the same for
made with cement brand A is higher. Beyond this age, concrete M21, M34, and M48 concrete types. The percentage of tensile
made with cement brand B gains strength at a faster rate com- strength in relation to compressive strength is similar for all the
pared to the other cement brands (similar to mortar cubes). concrete mix types and is close to 10 % of compressive strength
Note that this trend is similar for all the three concrete mix for all the cement brands.
types (M21, M34, and M48). Although the maximum strength It is seen in Figs. 1 and 3 that the results of cube compres-
at 14 and 28 days was provided by cement brand B, the strength sive strength and concrete tensile strength support the trends of
of concrete made with cement brand A match closely at both concrete compressive strength for the cement brands employed
these ages. The curves of cement brands A–C are distinctly (Fig. 2).
separated from the rest of the cement brands. Cement brands
D–G provided similar results. The least strength at any age was
provided by cement brand G. Analytical Modeling
Nasir [40] found that the locally available cements vary in
chemical and compound composition and fineness. Of the four
compounds of cement, C3S and C2S are important for cement
FIG. 1 Average compressive strength of cement-sand mortar cubes. and concrete strength. C3S largely contributes to the early age
strength gain of cement and C2S contributes to the long-term
(3 months and onwards) strength of cement. Therefore, silicate
contents are the logical variables for the concrete strength pre-
diction model; other parameters may include cement content,
w/c ratio, CA, and FA. However, an approach based on age of
concrete has been employed for the prediction model in the
studies presented. The effects of chemical and compound
composition, and fineness of cement have been incorporated
implicitly in the model. Furthermore, in order to avoid use of
multiple variables and to keep the model simple only 7-day
compressive strength has been employed as a single variable to
predict concrete strength up to 28 days. This is owing to the
fact that the concrete strength is influenced by compound
RAFI AND NASIR ON MODELS FOR CONCRETE STRENGTH 5
FIG. 2
Data of observed average concrete compressive
strength.
composition (in particular C3S) and fineness of cement largely The suggested model for prediction of concrete strength
up to 7 days and their influence becomes small at later ages. As beyond 7-day strength is based on Eq 1
a result, 7-day strength includes the effects of both the chemical
composition of cement and mix proportion of concrete. (1) fc ¼ a fc;7 tnb ; f7 < tn 28g
6 Journal of Testing and Evaluation
FIG. 3
Tensile strength of concrete at 28 days.
a b
Cement Brand Eq 1 Eq 8 Eq 1 Eq 8
FIG. 4
Observed versus predicted strength of
concrete: (a) cement brand A, (b) cement
brand C, (c) cement brand E, and (d) cement
brand G.
RAFI AND NASIR ON MODELS FOR CONCRETE STRENGTH 7
where:
fc,7 ¼ 7-day compressive strength of concrete, FIG. 5 Comparison of observed and predicted concrete strength using Eq 6.
(2) a ¼ f ðFineness; C3 SÞ
(3) b ¼ f ðC3 S þ C2 SÞ
(4) y ¼ f ðx; cÞ
fc,28 (MPa)
Cement Brand Mixes Cement (kg/m3) FA (kg/m3) CA (kg/m3) W/C fc,7 (MPa) Reported (A) Predicted (B) A/B
Commercial OPC [30] 1 400 600 1200 0.45 24.50 34.01 34.00 1.00
2 390 588 1170 0.5 22.50 32.50 31.23 1.04
3 380 570 1140 0.55 21.60 32.50 29.98 1.08
4 370 555 1110 0.6 21.50 32.30 29.84 1.08
5 360 540 1080 0.65 21.10 30.50 29.28 1.04
6 350 525 1050 0.7 20.40 30.30 28.31 1.07
7 325 650 1300 0.45 20.30 29.21 28.17 1.04
8 320 640 1280 0.5 20.00 28.90 27.76 1.04
9 315 630 1260 0.55 18.50 27.70 25.68 1.08
10 310 620 1240 0.6 17.60 25.90 24.43 1.06
11 305 610 1220 0.65 17.30 24.50 24.01 1.02
12 300 600 1200 0.7 14.60 23.8 20.26 1.17
13 517 517 1034 0.45 31.01 44.01 43.03 1.02
14 504 504 1008 0.5 29.90 39.41 41.50 0.95
15 491 491 982 0.55 28.30 37.50 39.28 0.95
16 479 479 958 0.6 26.70 36.10 37.06 0.97
17 468 468 936 0.65 25.80 35.21 35.81 0.98
18 457 547 914 0.7 25.70 34.61 35.67 0.97
Ambuja cement [14] 1 370 781 1055 0.5 13.90 21.83 21.82 1.00
4 385 767 1056 0.48 14.44 23.07 22.67 1.02
5 356 797 1057 0.52 15.08 24.85 23.67 1.05
6 396 744 1056 0.48 15.72 25.58 24.66 1.04
8 396 744 1057 0.48 16.12 26.85 25.29 1.06
10 365 775 1056 0.52 16.94 25.98 26.59 0.98
12 356 825 1021 0.52 17.13 28.01 26.89 1.04
13 396 776 985 0.48 17.66 28.58 27.72 1.03
15 402 780 1023 0.46 19.23 28.56 30.17 0.95
16 365 808 1023 0.52 19.57 29.78 30.72 0.97
17 402 780 985 0.46 19.86 27.69 31.18 0.89
18 440 740 1021 0.42 19.86 27.69 31.18 0.89
19 396 776 1021 0.48 20.41 32.56 32.02 1.02
20 380 790 1021 0.5 20.41 32.56 32.02 1.02
OPC N [42] N7 270 900 900 0.7 24.70 31.60 30.54 1.03
N6 320 860 900 0.6 33.00 41.40 40.80 1.01
N5 380 810 900 0.5 41.90 50.50 51.80 0.97
OPC C [42] C7 270 900 900 0.7 20.20 27.80 27.15 1.02
C6 320 860 900 0.6 25.20 35.10 33.88 1.04
C5 380 810 900 0.5 34.50 46.90 46.38 1.01
OPC Type-I [43] C1 1:2:0 — — 0.86 15.52 21.10 23.19 0.91
C2 1:2:3 — — 0.86 9.10 13.66 13.61 1.00
C3 1:2:3 — — 0.86 11.38 16.55 17.01 0.97
C4 1:3:0 — — 0.86 12.28 18.48 18.34 1.01
C5 1:3:4.5 — — 0.86 8.76 14.21 13.09 1.09
Type-I Portland cement [44] Mix 1 310 800 1100 0.55 19.59 27.59 24.99 1.09
Mix 1 310 800 1100 0.55 22.00 25.86 28.08 0.98
Mix 1 310 800 1100 0.55 22.28 26.21 28.43 0.91
Mix 1 310 800 1100 0.55 20.83 26.34 26.58 0.99
Mix 2 310 800 1100 0.55 19.31 25.45 24.64 1.07
Mix 2 310 800 1100 0.55 18.83 23.52 24.03 1.06
Mix 3 310 800 1100 0.55 19.24 27.95 24.55 0.96
— — — — — 20.76 27.59 26.49 1.06
— — — — — 22.00 26.25 28.08 0.98
— — — — — 20.85 25.87 26.61 0.99
— — — — — 22.25 27.59 28.39 0.91
RAFI AND NASIR ON MODELS FOR CONCRETE STRENGTH 9
TABLE 5 a and b for cement brands reported in literature. where a and b are constants that range from 0.05 to 9.25 and
Cement Brand a b 0.67 to 0.98, respectively
Equation 7 was modified to predict 28-day strength of con-
Commercial OPC [30] 0.63 0.24
crete using fc,7. The modified expression is given as Eq 8.
Ambuja cement [14] 0.53 0.33
OPC N [42] 0.76 0.15 (8) fc;t ¼ fc;7 t=ða þ bt Þ
OPC C [42] 0.69 0.20
OPC Type-I [43] 0.55 0.30 The definitions of a and b in Eq 8 are the same as given by
Type I Portland cement [44] 0.59 0.23 Eqs 2 and 3, respectively. The values of these parameters in Eq
8 were determined by using the procedure mentioned earlier for
the cement brands employed in the presented studies and are
Figure 4 compares the observed and predicted strength
given in Table 3. It is seen in Table 3 that only the values of a are
plots for some of the cement brands. The predicted strength within the aforementioned suggested range; the values of b
was calculated with the help of Eq 1 and values of a and b taken are less than those suggested by ACI code. This could be attrib-
from Table 3 were employed. It is seen in Fig. 4 that a good uted to the use of fc,7 in Eq 8 instead of fc,28. Based on the val-
correlation exists between the two data. ues of a and b as given in Table 3, Eq 8 was employed to predict
It is noted in Table 3 that the values of a and b are close for the concrete strength for the three strength classes included in
the employed cement brands and may provide constant values the presented studies. The results were demonstrated in Fig. 4.
of these parameters for all the cement brands with a negligible It is seen in Fig. 4 that the observed and predicted strengths
comprise on accuracy of results from Eq 1. This avoids the need match closely and for the majority of data the difference is less
of identifying the cement brand for selecting a and b. The aver- than 10 %.
age values of a and b from Table 3 come out to be 0.56 and 0.29, Similar to Eq 1, it is noted in Table 3 that the values of a
respectively. On substitution of these average values, Eq 1 can and b are close for all the cement brands employed and may
be written as Eq 6. Figure 5 illustrates the results of observed lead to their constant values in Eq 8. The average values of a
and predicted strengths using Eq 6. It is seen in Fig. 5 that the and b for the cement brands in Table 3 come out to be 3.2 and
two data match closely for most of the concrete ages; the maxi- 0.58, respectively. Equation 8 can, therefore, be written as Eq 9
mum error for the majority of data is less than 10 %. As a result, using the average values of a and b. The results of predicted
Eq 6 can be employed as a general expression to predict the strengths using Eq 9 are given in Table 6. A good correlation
strength of concrete made with the cement brands manufac- between the observed and predicted strengths exists in Table 6.
tured in Pakistan.
(9) fc;t ¼ fc;7 tn =ð3:2 þ 0:58t Þ; f7 < tn 28g
(6) fc ¼ ð0:56Þ fc;7 tn0:29 ; f7 < tn 28g
In order to validate the model further, help was taken from the
Limitations of Models
data in the available literature. Table 4 shows the details of
mixes along with their 7-day and 28-day (fc,28) strengths and Although a reasonably large data set of concrete mixes was
cement brands employed for the data selected from literature. employed for the studies presented, there are certain limitations
The parameters a and b were determined for all the mixes in that apply on the use of Eqs 1, 6, 8, and 9. The employed mixes
Table 4 using the aforementioned procedure and are summar- were made with OPC and normal weight aggregates; the con-
ized in Table 5. Table 4 provides the ratio of reported and pre- crete strength varies between 14–48 MPa. The concrete samples
dicted 28-day strengths. It is noted in Table 4 that the difference were wet cured by immersing them in a water tank. Further-
between the reported and predicted strengths is within 10 % more, only the workability increasing admixture was used in the
and is insignificant. Note that in the absence of information, it concrete. It is the responsibility of the reader to verify the results
has been assumed that the data of observed strength in Ref. [30] of Eqs 1 and 8 (international readers), and Eqs 6 and 9 (Paki-
is based on cylindrical strength. If this is not the case, then a stani readers) for conditions differing than those mentioned
and b values given in Table 5 will not apply. above.
Cement Brand Mix Type Observed (A) Predicted (B) A/B Observed (C) Predicted (D) C/D
A M21 18.00 18.58 0.97 20.17 21.63 0.93
M34 32.74 36.06 0.91 35.68 42.00 0.85
M48 41.34 41.54 1.00 48.10 48.38 0.99
B M21 18.53 16.96 1.09 21.32 19.76 1.08
M34 34.18 34.80 0.98 36.28 40.53 0.90
M48 44.67 39.52 1.13 48.82 46.03 1.06
C M21 16.23 14.61 1.11 18.95 17.02 1.11
M34 28.06 28.60 0.98 32.50 33.31 0.98
M48 36.61 37.30 0.98 45.89 43.44 1.06
D M21 14.58 14.67 0.99 18.91 17.09 1.11
M34 24.27 26.14 0.93 31.49 30.45 1.03
M48 34.46 36.32 0.95 44.86 42.30 1.06
E M21 14.15 14.52 0.97 18.50 16.91 1.09
M34 23.82 25.53 0.93 31.10 29.73 1.05
M48 34.33 37.33 0.92 43.74 43.48 1.01
F M21 14.47 14.31 1.01 18.60 16.67 1.12
M34 23.94 25.60 0.94 31.48 29.81 1.06
M48 34.46 36.66 0.94 44.57 42.69 1.04
G M21 13.88 14.40 0.96 17.85 16.77 1.06
M34 23.17 24.44 0.95 30.47 28.46 1.07
M48 33.31 34.75 0.96 43.43 40.47 1.07
cylinders were also carried out. The conclusions drawn from the also based on the determination of parameters a and b
presented studies are listed as under. similar to Eq 1. The generalized form of Eqs 8 and 9 was
found to provide the strength results with sufficient accu-
1. The compressive strength of concrete qualitatively corre-
racy for the cement brands available in Pakistan.
sponded to the compressive strength of cement–sand
mortar for each cement brand. Therefore, the same
cement brand provided the highest strength in both the ACKNOWLEDGMENTS
concrete cylinders and mortar cube tests. This applied at The writers wish to acknowledge the support provided for
7 and 28 day ages of concrete and mortar, and all the this research by the laboratory technical staff members.
three concrete strength classes. Financial assistance provided by Lucky Cement is gratefully
2. The splitting tensile strength of concrete cylinders fol- acknowledged.
lowed the trends observed for 28-day concrete compres-
sive strength. The cement with the highest compressive
strength yielded the highest tensile strength. The tensile References
strength of concrete for all the cement brands employed [1] ASTM C192/C192M-02: Standard Practice for Making and
was found to be nearly 10 % of its compressive strength. Curing Concrete Test Specimens in the Laboratory, ASTM
3. An analytical model to predict 28-day strength of con- International, West Conshohocken, PA, 2002, www.astm.org.
crete using its 7-day strength is suggested (Eq 1). The sug- [2] Blaine, R. L., Arni, H. T., and Defore, M. R., “Compressive
gested model is generalized to predict strength of concrete Strength of Test Mortars. Interrelations Between Cement
made from the cements manufactured in Pakistan (Eq 6). and Concrete Properties,” Part 3, Building Science Series 8,
4. The suggested model (Eq 1) also provided satisfactory National Bureau of Standards, Washington, D.C., 1968,
pp. 1–65.
results of strength for data of concrete mixes available in
[3] Popovics, S., “Generalization of the Abrams Kaw-
the technical literature. It can, therefore, be employed in
predication of Strength Development of Concrete from
any part of the world and shall require one-time determi- Cement Properties,” ACI J. Proc., Vol. 78, No. 2, 1981,
nation of parameters a and b for a particular cement pp. 123–129.
brand. [4] Popovics, S., “Extended Model For Estimating the Strength
5. A modification of ACI code equation is suggested to pre- Developing Capacity of Portland Cement,” Mag. Concr.
dict 28-day strength of concrete (Eq 8). This equation is Res., Vol. 33, No. 116, 1981, pp. 147–153.
RAFI AND NASIR ON MODELS FOR CONCRETE STRENGTH 11
[5] Popovics, S. and Popovics, J. S., “Ultrasonic Testing to Compressive Strength at the Ages of 7 and 28 Days Within
Determine Water–Cement Ratio For Freshly Mixed 24 Hour,” Mater. Struct., Vol. 36, No. 264, 2003, pp.
Concrete,” Cem. Concr. Aggr., Vol. 20, No. 20, 1998, 693–701.
pp. 262–268. [22] Sang, H. H., Jin, K. K., and Yon, D. P., “Prediction of Com-
[6] Alexander, K. M., “The Relationship Between Strength and pressive Strength of Fly Ash Concrete by New Apparent
the Composition and Fineness of Cement,” Cem. Conc. Activation Energy Function,” Cem. Concr. Res., Vol. 33,
Res., Vol. 2, No. 6, 1972, pp. 663–680. No. 7, 2003, pp. 965–971.
[7] Popovics, S., “Analysis of the Concrete Strength Versus [23] Snell, L. M., Roekel, J. V., and Wallace, N. D., “Predicting
Water–Cement Ratio Relationship,” ACI Mater. J., Vol. 87, Early Concrete Strength,” Concr. Int., Vol. 11, No. 12,
No. 5, 1990, pp. 517–529. 1989, pp. 43–47.
[8] Mehta, P. K. and Monteiro, P. J. M., Concrete–Structure, [24] Zelic, J., Rusic, D., and Krstulovic, R., “A Mathematical
Properties and Materials, Prentice Hall, Englewood Cliffs, Model for Prediction of Compressive Strength in
NJ, 1993. Cement–Silica Fume Blends,” Cem. Concr. Res., Vol. 34,
[9] Creskoff, J. J., “Estimating 28-Day Strength of Concrete No. 12, 2004, pp. 2319–2328.
From Earlier Strengths—Including the Probable Error of [25] Akkurt, S., Tayfur, G., and Can, S., “Fuzzy Logic Model
the Estimate,” ACI J. Proc., Vol. 41, No. 4, 1945, For Prediction of Cement Compressive Strength,” Cem.
pp. 493–512. Concr. Res., Vol. 34, No. 8, 2004, pp. 1429–1433.
[10] Neville, A. M., “A General Relation for Strengths of [26] Hwang, K., Noguchi, T., and Tomosawa, F., “Prediction
Concrete Specimens of Different Shapes and Sizes.” ACI J. Model of Compressive Strength Development of Fly–Ash
Proc., Vol. 63, No. 10, 1966, pp. 1095–1109. Concrete,” Cem. Concr. Res., Vol. 34, No. 12, 2004, pp.
[11] Popovics, S., “Discussion of ‘General Relation for Strengths 2269–2276.
of Concrete Specimens of Different Shapes and Sizes’ by [27] Mehta, P. K. and Monteiro P. J. M., Concrete, Microstruc-
A. M. Neville,” ACI J. Proc., Vol. 64, No. 6, 1967, pp. ture, Properties and Materials, McGraw-Hill, New York,
1566–1568. 2006.
[12] Voellmy, A., “Strength Control of Concrete Pavements,” [28] Hamid-Zadeh, N., Jamil, A., Nariman-Zadeh, N., and
Betonstrasse–Jahrbuch, Vol. 3, Fachverband Zement, Koln Akberzadeh, H., “A Polynomial Model for Concrete Com-
S., 1958, pp. 179–204. pressive Strength Prediction Using GMDH-Type Neural
[13] Zain, M. F. M. and Abd, S. M., “Multiple Regression Model Networks and Genetic Algorithm,” presented at the 5th
for Compressive Strength Prediction of High-Performance WSEAS International Conference on Systems Science and
Concrete,” J. Appl. Sci., Vol. 9, No. 1, 2009, pp. 155–160. Simulation in Engineering, Canary-Island, Spain, Dec
[14] Hasan, M. M. and Kabir, A., “Prediction of Compressive 16–18, 2006, ASTM International, West Conshohocken,
Strength of Concrete from Early Age Test Result,” PA (to be published).
presented at the 4th Annual Paper Meeting and 1st Civil [29] Akbarzadeh, H., Hamidzadeh, N., and Ranjbar, N. M.,
Engineering Congress, Dhaka, Bangladesh, Dec 22–24, “Study on the Relations Between Compressive Strength of
2011, ASTM International, West Conshohocken, PA (to be Concrete at Different Ages Using 165 Case Studies,” pre-
published). sented at the 7th International Congress on Civil Engineer-
[15] Onwuka, D. O., Okere, C. E., Arimanwa, J. I., and Onwuka, ing, Tehran, Iran, May 8–10, 2006, ASTM International,
S. U., “Prediction of Concrete Mix Ratio Using Modified West Conshohocken, PA (to be published).
Regression Theory,” Comput. Methods Civ. Eng., Vol. 2, [30] Zain, M. F. M., Suhad, M., Abd-Hamid, R., and Jamil, M.,
No. 1, 2011, pp. 95–107. “Potential for Utilizing Concrete Mix Properties to Predict
[16] Hobbs, D. W., “The Compressive Strength of Concrete: A Strength at Different Ages,” J. Appl. Sci., Vol. 10, No. 22,
Statistical Approach to Failure,” Magn. Concr. Res., 2010, pp. 2831–2838.
Vol. 24, No. 80, 1972, pp. 127–138. [31] Jee, N., Sangchun, Y., and Hongbum, C., “Prediction of
[17] Zielinszki, S., “The Development of the Setting of Roman Compressive Strength of In-Situ Concrete Based on Mix-
and Portland Cements in Pastes, in Mortars, and in ture Proportions,” J. Asian Arch. Build. Eng., Vol. 3, No. 1,
Concrete,” Proceedings of the International Association for 2004, pp. 9–16.
Testing Materials, Copenhagen, Denmark, ASTM Interna- [32] ASTM C150-04: Standard Specification for Portland
tional, West Conshohocken, PA, 1990, pp. 1–55. Cement, ASTM International, West Conshohocken, PA,
[18] Zielinszki, S., and Zhuk, J., “Comparative Investigation of 2004, www.astm.org.
Roman Cements,” presented at the Third International [33] ASTM C127-12: Standard Test Method for Density, Rela-
Congress of the Association for Testing Materials, Budapest, tive Density (Specific Gravity), and Absorption of Coarse
Hungary, 1901, ASTM International, West Conshohocken, Aggregates, ASTM International, West Conshohocken,
PA (to be published). PA, 2012, www.astm.org.
[19] Popovics, S., “Factors Affecting the Relationship Between [34] ASTM C29/C29M-09: Standard Test Method for Bulk
Strength and Water–Cement Ratio,” Mater. Res. Standard, Density (“Unit Weight”) and Voids in Aggregates,
MTRSA, Vol. 7, No. 12, 1967, pp. 527–534. ASTM International, West Conshohocken, PA, 2009,
[20] Tsivilis, S. and Parissakis, G., “A Mathematical Model for www.astm.org.
the Prediction of Cement Strength,” Cem. Concr. Res., [35] ASTM C128-12: Standard Test Method for Density, Rela-
Vol. 25, No. 1, 1995, pp. 9–14. tive Density (Specific Gravity), and Absorption of Fine
[21] Kheder, G. F., Al-Gabban, A. M., and Suhad, M. A., “A Aggregates, ASTM International, West Conshohocken,
Mathematical Model for the Prediction of Cement PA, 2012, www.astm.org.
12 Journal of Testing and Evaluation
[36] ASTM C136-06: Standard Test Method for Sieve Analysis [41] Montgomery, D. C. and Runger, G. C., Applied Statistics
of Fine and Coarse Aggregates, ASTM International, West and Probability for Engineers, John Wiley and Sons,
Conshohocken, PA, 2006, www.astm.org. New York, 2003.
[37] ASTM C33-03: Standard Specification for Concrete Aggre- [42] Ben-Bassat, M., Nixon, P. J., and Hardcastle, J., “The Effect
gates, ASTM International, West Conshohocken, PA, of Differences in the Composition of Portland Cement on
2003, www.astm.org. the Properties of Hardened Concrete,” Magn. Concr. Res.,
[38] ASTM C109/C109M-02: Standard Test Method for Com- Vol. 42, No. 151, 1990, pp. 59–66.
pressive Strength of Hydraulic Cement Mortars (Using [43] Gilkay, H. J., “Water Cement Ratio Versus Strength—
2-in. or [50-mm] Cube Specimens), ASTM International, Another Look,” J. Am. Conc. Inst., Vol. 2, No. 58, 1961, pp.
West Conshohocken, PA, 2002, www.astm.org. 1851–1878.
[39] Hassoun, M. N. and Al-Manaseer, A., Structural [44] Naseer, K. W. and Al-Manaseer, A. A., “It’s Time for a
Concrete, Theory and Design, John Wiley & Sons, Change From 6 x 12 to 3 x 9-in. Cylinders,” ACI Mater. J.,
New York, 2008. Vol. 84, No. 3, 1987, pp. 213–216.
[40] Nasir M. M., 2013, “Investigation of Properties of [45] American Concrete Institute, “Prediction of Creep, Shrink-
Cements Manufactured in Pakistan,” M. Eng. thesis, NED age and Temperature Effects in Concrete Structures,” ACI
University of Engineering and Technology, Karachi, 209R-92 (Reapproved 1997), ACI Committee 2011 Detroit,
Pakistan. MI, 1997.