Construction and Building Materials 29 (2012) 659–666

Contents lists available at SciVerse ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Estimating early-age in situ strength development of concrete slabs

M. Sofi a,⇑, P.A. Mendis a, D. Baweja b
a
Department of Infrastructure Engineering, The University of Melbourne, Parkville, Victoria 3010, Australia
b
School of Civil and Environmental Engineering, University of Technology, Sydney, NSW 2006, Australia

a r t i c l e i n f o a b s t r a c t

Article history: Post-tensioned concrete members are becoming very popular around the world due to structural efficien-
Received 24 February 2011 cies. However, some important issues related to concrete properties are not resolved yet. A more accurate
Received in revised form 21 September 2011 prediction of the in situ early-age strength is needed to allow for the application of the post-tensioning
Accepted 2 October 2011
(PT) load without causing any damage to the PT anchorage zone concrete. Strength–maturity functions
Available online 24 December 2011
are conventionally used to estimate the in situ concrete strength on the basis of a datum temperature.
Temperature Match Curing (TMC) is an alternative method which provides a more representative esti-
Keywords:
mate of the in situ strength. In this paper, the prediction of in situ strength of slabs using maturity func-
Maturity
Early-age concrete
tions is compared with the results from the TMC and those cured following standard recommendations.
In situ strength development The comparisons generally reveal that the specimens cured and tested according to the standard methods
Temperature Match Curing can be used for estimating the strength development of concrete with some reservations.
Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction strength in situ and is widely used in the industry. TMC is a process
by which test specimens of concrete, from a representative sample
Concretes used in post-tensioned (PT) concrete slabs have of the concrete that is placed in an in situ or precast element, are
early-age strength requirements for initial and final stressing pro- cured by keeping them in a chamber at the same temperature as a
cesses. This is typically at 1 day age for initial stressing and three to pre-selected point in the element [7]. The TMC system can function
4 days age for final stressing. The age at which the final stressing is under isothermal, near adiabatic, pre-recorded, and even real-time
applied depends upon a confident prediction of in situ strength. profile conditions. It has been successfully used by a number of
The normal practice is to use a minimum compressive strength investigators [8–10] to predict the in situ strength. In TMC and core
of 22 MPa [1]. tests performed by Mak and Ritchie [11] on 200 mm thick slabs, a
The Australian Standard on specification and supply of concrete good match was obtained between the cores from in situ concrete
AS1379 [2] requires concrete samples to be tested in accordance slab and the TMC specimens, with less than 1% error overall. This
with AS1012 Parts 1, 8 and 9. According to AS1012 [3], the concrete finding indicates that both TMC and core cylinders subjected to a
samples are cured under standard-moist conditions where the similar temperature history have the same maturity, when consid-
samples are subjected to constant temperatures of 23 ± 2 °C and ering an allowance for converting from core to cylinder strengths.
27 ± 2 °C for a standard temperate zone and standard tropical zone, Other authors have reached similar conclusions that the TMC system
respectively. In ASTM C192/C192M [4], the curing temperature is provides the closest correlation with in situ strength of early-age
set at 23 ± 2 °C. On the other hand, the Australian Concrete Struc- concrete [12]. However, the technology remains costly and inacces-
tures Standard AS3600 [5] and ACI 318-02 [6] state that accompa- sible to small to medium manufacturers.
nying cylinder specimens of the same concrete batch should be The standard method of testing according to AS1012 [3] pre-
stored and cured under conditions similar to those of the in situ scribes a curing regime and boundary conditions that are essen-
concrete. What constitute ‘‘conditions similar’’ to the in situ con- tially different from those under which the in situ concrete is
crete are open to interpretation. cured. Slabs have a high surface to volume ratio and while the
There is a major technical problem in the construction industry in top surface is exposed to variable temperature profiles, the other
that there are no standard methods for assessing in situ strength of surfaces are sealed by formwork. It is well known that the curing
concrete used in post-tensioned concrete slab construction. Tem- temperature of concrete is an important factor governing early-
perature Match Curing (TMC) is well established for assessing age strength development, as the progress of cement hydration is
temperature sensitive. The maturity method has reportedly been
⇑ Corresponding author. Tel.: +61 3 8344 6767; fax: +61 3 8344 4616. used as an effective means of predicting the early-age of in situ
E-mail addresses: [email protected] (M. Sofi), [email protected] concrete in spite of its inherent limitations [13–19]. The maturity
(D. Baweja). method is used to account for the combined effect of temperature

0950-0618/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.conbuildmat.2011.10.019
660 M. Sofi et al. / Construction and Building Materials 29 (2012) 659–666

and time on the development of the hydration reaction and the Cylinder specimens are cured at different constant temperatures
mechanical properties of concrete. The nonlinear maturity function and their compressive strengths are determined at different age
based on the Arrhenius law, which is also known as equivalent age intervals. The compressive strength predicted by maturity method
function [Eq. (1)], is considered the most accurate, as it better rep- and standard procedures (by curing specimens under ambient and
resents the effect of temperature on strength development over a standard conditions) were compared with the in situ strength indi-
wide range of temperatures [20]: cated by the TMC results in Sections 3 and 4 respectively.

X
t
Ea 1 1
te ¼ e R ðT T r Þ Dt ð1Þ 2. Experimental procedure
0
2.1. In situ temperature measurement and data logging

where te is the equivalent age at the reference temperature, days; Tr Ten post-tensioned suspended slabs were instrumented in two construction
is the reference temperature, K; Ea is the apparent activation energy, projects, Sentinel Apartment and Austin Hospital in Melbourne during the winter
J mol1; R is the universal gas constant, 8.314 J mol1 K1; T is the time [24]. These represent typical slabs cast for multi-story buildings in Australia.
temperature of the concrete during time interval Dt, K The winter time is of particular interest as the range of average daily temperature
could be lower than that of standard 23 °C. In this case, the standard testing proce-
The reference temperature (Tr) is the standard cured tempera-
dure following the Australian Standards AS1012 [3] may not accurately represent
ture at which test specimens are cured. In many parts of the world the in situ concrete strength. In this standard procedure, the concrete samples
it is 20 °C (293 K). In Australia it is 23 °C in temperate zones and are cured under standard-moist conditions at constant temperatures of 23 ± 2 °C
27 °C in tropical zones. for a standard temperate zone.
Once the equivalent age is found, strength maturity functions Each of the concrete slab placements was approximately 570 m2 in surface area
and required 100 m3 of concrete. All slabs had a thickness of 230 mm. Thin copper-
can be used to find the strength at a particular age. Carino and constantan type double T reusable thermocouple wires with cold junction compen-
Lew [19] have reviewed the basic concepts underlying a number sation and linearisation were used to measure the temperature variation at various
of strength–maturity functions for the strength gain of concrete locations. Fig. 1 shows the arrangement of the thermocouples in the slab cross sec-
under isothermal curing. Table 1 lists these functions. An evalua- tion (100, 150, 220 mm, measured from the surface of the slab), one thermocouple
measuring the ambient temperature variation next to the slab and another one
tion of the functions and how they relate to the in situ strength
embedded in the ambient cured cylinders cast beside the slab. No insulation mate-
of concrete as measured by other means such as Temperature rials were used for the ambient cured cylinders, as recommended by AS3600 [5]
Match Curing (TMC), has received very little attention. and ASTM C31/C31M [4].
Given that an accurate prediction of in situ strength is required Temperature profiles of the slabs were obtained using five channel temperature
prior to the application of a pre-stressing load, the main objective data loggers. The data logger was programmed to log temperature and equivalent
age at 5 min intervals, starting after the concrete was poured. These data were then
of this paper is to establish whether the existing procedures can averaged to 1 h intervals and downloaded onto a host computer for the purposes of
reasonably predict the in situ strength of concrete. To that end, a strength prediction. A thermocouple was used to measure the temperature ambient
comparison of the compressive strength values obtained from temperature variation next to the anchorage zone, which is the critical region dur-
specimens cured under TMC, standard, ambient conditions, and ing the stressing process.
The temperature variation measured over 80 h (or approximately 3.5 days) is
the strength–maturity functions is carried out. In this paper, the
reported in Fig. 2. Data recorded from Channels 1–3 represent the temperature var-
TMC cured specimens will be considered to represent the in situ iation across the slab depth (see Fig. 1). Some small differences in temperature re-
strength development accurately. The assessment of the strength corded across the slab depth can clearly be seen. The thermocouple deeper inside
values is limited to 4 days age, corresponding to the age of final the slab (Channel 3) records higher temperatures. However, the in situ temperature
stressing. It is noted that the variation of temperature ranging from profile of each slab is unique depending on the ambient temperature conditions of
the day. The ambient cured cylinder temperatures were generally lower than the
10 to 35 °C used in this study represents temperature variation in
temperatures recorded in the in situ concrete slab.
temperate zones. Therefore, the results outcomes presented in this
paper are limited to those regions, such as eastern and southern 2.2. Temperature Match Curing (TMC)
part of Australia, New Zealand, southern part of Europe, California,
southern part of South America and South Africa, and other regions As mentioned earlier, TMC is a process by which test specimens of concrete,
with similar climate. from a representative sample of the concrete that is placed in an in situ or precast
element, are cured by keeping them in a chamber at the same temperature as a pre-
Two building projects in Melbourne were investigated to obtain
selected point in the element [7]. A state-of-the-art integrated TMC system was set
typical temperature profile of the slab. The temperature profile was up. The system was capable of accurately simulating the in situ curing temperature
then used as an input temperature for the TMC specimens. Details variation by controlling the temperature of a water bath for estimating the real
of the experimental procedure are presented in Section 2. The pro- strength development in concrete. The systems were successfully used on a number
cedure to determine the parameters associated with the strength– of early projects to estimate the in situ strength of structural members [8,25,26].
The main components of a TMC system are depicted in a schematic drawing (Fig. 3).
maturity functions are the same as those reported by Carino and The system has an embedded process controller to precisely regulate the curing
Lew [19], the parameters kT, t0 and S1 are obtained by fitting the temperature profile. Accurate temperature matching is enabled using proportional,
strength–age relationship to each set of strength–age data. integral and derivative control of feedback temperature data obtained from sensors

Table 1
Strength–maturity functions and parameter.

Function name Strength–maturity function Parameters Eq.

Linear hyperbolic kT ðt e t 0 Þ
S ¼ S1 1þk [15] S = strength at age te, where te is the equivalent agea; (2)
T ðt e t 0 Þ
S1 = limiting strength; kT = rate constant, 1/day;
t0 = the age at the start of strength development
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Parabolic hyperbolic pkTffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ðt e t 0 Þ As listed above (3)
S ¼ S1 [21]
1þ kT ðt e t 0 Þ
a
Exponential  ts
S ¼ S1 e ½ e  [22] te = equivalent age; s = a time constant (value (4)
1/s is the rate constant kT for this function);
a = shape parameter
Plowman S ¼ a þ b logðMÞ [23] a = strength for maturity index M = 1; (5)
b = slope of the line; M = maturity index
a
te is the equivalent age defined by Eq. (1).
 M. Sofi et al. / Construction and Building Materials 29 (2012) 659–666 661

100
Channel 1 150
Channel 2
220
Channel 3

Fig. 1. Arrangement of thermocouple wires in the slab.

35
Channel 1 Channel 2
Channel 3 Ambient
30
Temperature (oC)

Cylinder
Fig. 3. Components of TMC systems.
25

20
35
15
30

Temperature (°C)
10 25
0 10 20 30 40 50 60 70 80
20
Time (hrs)
(a) Sentinel Apartment 15

10
35
Channel 1 Channel 2
Ambient Cylinder 5
30 0
Temperature (oC)

0 40 80 120 160
25 Time (hrs)
(a) TMC C2c – Sentinel Apartment
20
35

15 30
Temperature (°C)

25
10
0 10 20 30 40 50 60 70 80 20
Time (hrs)
15
(b) Austin Hospital
10
Fig. 2. Typical temperature profiles obtained for the thermocouple arrangements.
5

0
embedded into one or two concrete specimens within the conditioning chamber. 0 40 80 120 160
The sensors are part of the eight thermocouple inputs located at the temperature Time (hrs)
controller. A thermocouple connected to one of the eight thermocouple inputs is
also permanently positioned in the chamber to provide water temperature mea- (b) TMC C3d – Austin Hospital
surements. More details about the particular system used for this project can be
found in Ref. [25]. Fig. 4. Average concrete slab temperature profiles.
Two in situ concrete slab temperature profiles were selected from Sentinel
Apartment and Austin Hospital. To best represent the temperature across the slab The mix was designed to fully comply with AS3600 [5] and AS1379 [28]. A total
section, the in situ concrete temperature profiles (Channels 1–3) were averaged of 144 concrete cylinder specimens were made. The concrete specimens cured in
to represent a single profile in each case. The average profiles for the third day were the laboratory were batched at the plant and truck-mixed in order to obtain a
duplicated to extend the profiles to a 7 day period. These averaged profiles are pre- mix that is similar to the one that was dispatched to the site. All concrete specimens
sented in Fig. 4 and referred to as TMC C2c and TMC C3d for Sentinel Apartment and were ground and conditioned before being tested at 1–28 day age. The specimens
Austin Hospital, respectively. were cured under standard and ambient conditions, isothermal conditions and
Temperature Match Curing (TMC) methods.
2.3. Testing of specimens The specimens were cured in accordance with AS1012 [3] and AS3600 [5] for
standard and ambient conditions, respectively. The average compressive strengths
The ten post tensioned slabs in the two projects (Sentinel Apartment and Austin of three cylinder specimens are presented in Table 3. As an example, only the ambi-
Hospital) have similar geometry and concrete mix. Project specifications for these ent cured results from Sentinel Apartment are presented in this paper.
post-tensioned slabs recommended a concrete mix with a characteristic compres- Some specimens were cured under isothermal conditions in order to establish
sive strength at 28 days ðfc0 Þ of 32 MPa. In this study, the concrete mix similar to that the strength–maturity functions. The test schedule and corresponding results are
used in the building projects was considered. The concrete mix details are pre- presented in Table 4. The specimens were cured in lime saturated water baths of
sented in Table 2. A general purpose (GP) Portland cement satisfying AS3972 [27] constant curing temperatures (15, 23, 35 and 50 °C). Within 1.5 h from the start
was used. of casting, the moulds were covered with steel lids and transferred to the water
662 M. Sofi et al. / Construction and Building Materials 29 (2012) 659–666

Table 2 no curing compounds or other surface treatment applied to these

Concrete mix design. slabs. With no means to prevent heat exchange between the
Mix ingredient Content hydrating concrete and the ambient, the surface area of the slab
Portland cement content (kg/m ) 3
305 remains exposed to the ambient conditions. From Fig. 2, it can
Fly ash content (kg/m3) 35 (10% fly ash) be seen that although the in situ concrete slab temperature pro-
Total cementitious (kg/m3) 340 files do not vary considerably over the slab depth, the actual tem-
W/C ratio 0.5 peratures within the first 24 h change significantly as shown in
Water reducer (ml/m3) 1360
Accelerator (ml/m3) 1360
Fig. 2a. This difference could affect the strength development at
Air entraining agent (ml/m3) 0 an early-age between the ambient cured and in situ strength
20 mm aggregate (kg/m3) 550 when a particular strength is required for post-tensioning.
14 mm aggregate (kg/m3) 550 Fig. 2b shows the in situ temperature profiles for a similar slab
Washed concrete sand (kg/m3) 770
from a different project. The sensor at the bottom of the slab
Slump (mm) 80 ± 15
(Channel 3) records a relatively higher temperature due to the
wooden formwork insulation from the bottom and the thickness
of the concrete. While the in situ concrete slab temperatures also
Table 3 tracked the ambient temperature, they lagged behind the ambient
Standard and ambient cured cylinder compressive strengths (fc0 ).
by about 4 h.
Age (days) C1b, fc0 (MPa) C2b, fc0 (MPa) Fig. 5 represents the equivalent age results of the two different
Standard cured 23 °C Ambient cured buildings. The profiles shown in Fig. 5 were obtained using the
1 11.2 13.4 Arrhenius function [Eq. (1)], the ratio of the activation energy
2 – 20.4 against the gas constant, Ea/R at 4200 with a Tr value of 23 °C
3 24.2 21.9
[29]. At the end of the test at 71 h, on average the in situ concrete
4 26.4 23.9
7 31.7 28.8 slab reached an equivalent age of 66 and 60 h for Sentinel Apart-
28 42.5 40.3 ments and Austin Hospital, respectively. The measurements repre-
sent the temperature history of the slab section at the exterior edge
of the building in each case. The temperature profiles result in a
slightly higher equivalent age for the channels positioned deeper
Table 4
Strength–maturity parameters for isothermal curing. in the slab section. The temperature profiles of Channels 1 and 2,
and the cylinder appear to be consistent.
Test C1a C1b C1c C1d
The in situ equivalent age profiles of the concrete slab at various
Temperature 15 °C 23 °C 35 °C 50 °C
Age (days) Strength Strength Strength Strength locations (i.e., next to the core of the building, mid-span and outer
(MPa) (MPa) (MPa) (MPa) area) have equally been investigated. Fig. 6 represents the equiva-
1 6.3 11.2 17.0 19.0
lent age results of the two different buildings. It was found that
2 – – 23.2 23.6 there was no significant difference in average equivalent age of
3 18.8 24.2 – – the slab profile at these points.
4 21.9 26.4 28.3 29.9
7 28.9 31.7 32.5 32.8
28 37.4 42.5 42.7 32.5
80
Channel 1
Equivalent age (hrs)

70
Channel 2
Table 5
60
Channel 3
TMC cured cylinder compressive strengths (fc0 ). 50 Cylinder
40
Age (days) TMC C2c, fc0 (MPa) TMCC3d, fc0 (MPa)
Sentinel Apartment Austin Hospital 30
1 14.0 9.6 20
2 21.0 18.0 10
3 25.4 24.4
0
4 26.7 27.4 0 20 40 60 80
7 32.3 30.4
28 41.8 39.8 Elapsed time (hrs)
(a) C2c – Sentinel Apartment
baths maintained at the intended curing temperatures. After the intended curing
period, each specimen was removed from the tank and conditioned immediately 70
before testing. The average compressive strengths obtained from testing are pre- Channel 1
Equivalent age (hrs)

sented in Table 4. 60
Channel 2
To evaluate the in situ strength development of concrete, some specimens were
50
subjected to representative temperature profiles recorded from the two building Cylinder
projects using the TMC method. The specimens were cured under average temper- 40
ature profiles shown in Fig. 4. The average compressive strengths are presented in
Table 5. The strength values are the average compressive strength results of three
30
cylinder specimens. 20
10
3. Prediction of in situ strength using maturity functions 0
0 20 40 60 80
3.1. In situ slab temperature profiles Elapsed time (hrs)
(b) C3d – Austin Hospital
As stated earlier, slabs are structural members with a very high
surface area to depth ratio compared to other members. There are Fig. 5. Equivalent age profiles for slab at different depths.
 M. Sofi et al. / Construction and Building Materials 29 (2012) 659–666 663

80 28
70
Equivalent age (hrs)
24
60

Strength (MPa)
50 20

40 16 Linear Hyperbolic

30 Interior Parabolic Hyperbolic

12 Exponential
20 Median
Plowman
10 Exterior 8
Standard cured
0 4
0 20 40 60 80 0 1 2 3 4
Elapsed time (hrs) Equivalent age (days)
(a) Sentinel Apartment
Fig. 7. Strength predictions from regression analyses.

80
70
Equivalent age (hrs)

The rate of strength development measured by the rate con-

60 stant kT is listed in Table 6. It is commonly known that kT increases
50 with the increasing curing temperature. Lower kT values for 35 °C
40 may be attributed to the discrepancies in the experimental results.
Published work reports that kT depends on the particle size distri-
30 Interior
bution of the cement and the rate constant for each particle, which
20 Median is temperature dependent [21]. General purpose Portland (GP) and
10 Exterior blended cements (GB) are the two broad categories of cements
0 used in Australia. The latter are mixtures of Portland cement and
0 20 40 60 80 other cementitious compounds. GP type cements were used in this
Elapsed time (hrs) project. Higher rate of strength gain is reported for Type GP com-
(b) Austin Hospital pared to Type GB [30]. At 4 days age, strength development of be-
tween 50% and 70% of the 28 day compressive strength is reported
Fig. 6. Equivalent age profiles for slab at different location in the buildings. for Type GP. The rate of strength development depends on cement
and concrete mix composition and therefore cannot be assumed to
be standard. The evaluation of the effects of particle size on kT is
3.2. Determination of strength development parameters beyond the scope of this study.
The strength development is assumed to commence at age t0
For determining the strength–maturity function parameters, a and varies with the strength–age function. The value of t0 in the
method similar to that presented in Refs. [19,20] have been used. strength functions is defined by the point of intersection between
Cylinder specimens were cast according to AS1012 and cured un- the strength–maturity curves and the horizontal axis. On average,
der isothermal conditions (15, 23, 35 and 50 °C). Their compressive t0 ranges from 0.2 to 0.9 days. The value of t0 generally decreases
strengths were experimentally determined at regular age intervals with increasing temperature (Table 6). However it varies depend-
(1, 2, 3, 4, 7 and 28 days). ing on the strength function used. Due to the shape of the
A linear regression analysis was used to curve fit the strength– strength–maturity curves, there is a possibility that t0 might be
maturity functions (Eqs. (2)–(5) in Table 1) to the experimental slightly larger than the actual time taken for the commencement
data. In the process, the strength function parameters S1, kT, and of strength development. Mak and Torii [8] reported that the con-
t0 values were obtained for the combination of the function-tem- servative values of the t0 assumed in practice provide a higher mar-
peratures listed. For this purpose, a Matlab program was used to gin of safety in estimating early-age strength development.
obtain the parameters using the least square curve fit function: For the Plowman function, the compressive strength values and
the average of maturity were fitted to the linear function [Eq. (5)]
1X to obtain the values of parameters a and b. Day [29] reported that
ðf ðx; xdatai Þ  ydatai Þ2 ð6Þ the Plowman function has a bilinear relationship rather than a sin-
2 i
gle linear relationship. However, it was observed from the study
where xdata and ydata are the concrete ages and the corresponding that the linear relationship has a change in gradient at about
strengths at the time of testing, respectively and f(x, xdatai) repre- 7 days. As this paper is concerned with the early-age strength
sents the strength–maturity functions as presented in Table 1. The development, the linear function is deemed valid if only the mea-
results of the regression analysis for the reference temperature of surement for concrete at very early ages (less than 7 days old) is
23 °C are presented in Fig. 7. The values of the parameters corre- being used. Although this may not be strictly correct, it was deter-
sponding to the different temperatures are listed in Table 6 and mined that this is immaterial as long as the linear relationship is
briefly discussed in this section. satisfied prior to 7 days.
The limiting strength, S1, is the asymptotic value of the strength
functions that fits the experimental data. There are significant dif- 3.3. Predictions of in situ concrete strengths using maturity functions
ferences observed in S1 values for the different strength–maturity
functions. These differences are expected as S1 is obtained by curve A number of concrete cylinder specimens were subjected to
fitting of the experimental data. The best fit value for S1 does not temperature profiles collected from the site, using the TMC setup,
represent the actual long-term strength of the concrete [19]. There- to observe their early-age strength gain. The strength–maturity
fore, the ‘‘limiting strength’’ term is rather misleading as it has no functions describing the concrete mix used in this study were
physical significance. obtained using the parameters established in Section 3.2. The
664 M. Sofi et al. / Construction and Building Materials 29 (2012) 659–666

Table 6
Strength function parameters.

Function T (°C) R2 s (day) t0 (day) kT S1 a or 1 b

Linear hyperbolic 15 0.9998 0.46 0.323 42
23 0.9998 0.18 0.429 43
35 0.9995 n/a 0.292 47
50 0.9981 0.19 1.456 34
Parabolic hyperbolic 15 0.9978 0.91 0.168 54
23 0.9988 0.78 0.298 54
35 0.9999 0.27 0.219 60
50 0.9967 0.72 3.619 37
15 0.9997 2.4 0.423 43 0.782
Exponential 23 0.9999 1.6 0.61 46 0.691
35 0.9987 4.0 0.248 75 0.299
50 0.998 0.62 1.603 34 1.023
Plowman 0.9613 21.2 23.1

parameters were obtained using specimens cured at standard A sensitivity analysis was carried out to investigate the effects of
23 °C. In order for any comparisons between the strength values variation in Ea/R values to the strength predictions. It was shown in
of cylinders cured under various conditions and the strength– Fig. 9 that the strength prediction is rather insensitive to the change
maturity functions to be made, the age at which the strength of in Ea/R values (within 10% difference) with the exception of day 1
the cylinder specimen is measured needs to be converted into measurement. For this particular mix, the insensitivity to the Ea/R
equivalent age with a reference temperature of 23 °C. The objective value suggests that the activation energy recommended by Day
of the comparison is to establish whether the strength–maturity [29] could be adopted. However, it is noted that insensitivity could
function can be used in lieu of TMC. be attributed to the variation of the in situ temperature (ranging
To obtain equivalent age [Eq. (1)], it is necessary to establish the between 10 °C and 35 °C) which could result in similar maturities
activation energy Ea. Following the method presented by Carino as the specimens cured at the reference temperature of 23 °C. Nev-
and Lew [19], the rate constant kT for each of the curing tempera- ertheless, within the temperature range, this finding could greatly
tures listed in Table 6 were used to establish the activation energy simplify the procedure by eliminating the necessity to test speci-
(which describes the effect of temperature on the rate of strength mens at different temperatures. Measurements can be performed
development). The activation energy of the mix is obtained by plot- on specimens cured at 23 °C standard curing guidelines.
ting the natural logarithms of kT versus the inverse of the curing
temperature. The negative of the slope of the straight line equals
3.4. Prediction of in situ strength using ambient and standard cured
the activation energy (Ea) divided by the gas constant (R). It was
specimens
observed that Ea/R values from the strength measurement vary sig-
nificantly between 2000 and 8000. When using a value of Ea/R of
The cylinder compressive strengths at their testing age for
4200 as suggested in Ref. [29], it is shown in Fig. 8 that the strength
ambient cured and TMC cured (Test C2c) specimens are shown in
function can reasonably predict the in situ strength of concrete
Fig. 10. There appears to be a negligible difference between ambi-
represented by the TMC results (denoted TMC C2c and C3d repre-
ent and in situ cured specimens for the first 2 days. However, the
senting results from Sentinel Apartment and Austin Hospital
difference between these becomes more significant after 2 days.
respectively). The difference between the strength predictions
This result is expected, as in the first few hours the temperature
using the maturity functions and the in situ strengths were found
profile of the cylinder raises similar to that of the in situ slab,
to be less than 10% with the exception of the 1 day age. The
due to hydration reaction. However, the slab retains the heat due
strength–maturity parameters were obtained by curve fitting to
to its larger geometry and the insulation provided by the timber
the strength measurements from 1 day age. If the average ambient
formwork. Consequently, the rate of hydration is higher in the
temperature is less than 23 °C, then the equivalent age will be low-
in situ concrete resulting in slightly higher strengths.
er than the real age. This could be an issue at winter times when
From the comparison between the in situ compressive strengths
the ambient temperature is generally well below 23 °C.
(TMC C2c and C3d) and the strength of the standard cured speci-

28
30
24
Strength (MPa)

26
Strength (MPa)

20
22
Linear Hyperbolic
16 Parabolic Hyperbolic
18 Linear Hyperbolic
Exponential Parabolic Hyperbolic
12 Plowman Exponential
Standard cured
14 Plowman
8 TMC C2c E/R 4200
10 E/R 2000
TMC C3d
E/R 6000
4 E/R 8000
0 1 2 3 4 6
0 1 2 3 4
Equivalent age (days)
Equivalent age (days)
Fig. 8. Comparison of strength functions with specimens cured under different
temperature profiles. Fig. 9. Sensitivity of strength predictions to E/R values.
 M. Sofi et al. / Construction and Building Materials 29 (2012) 659–666 665

30 30

25 25

TMC cured
Strength (MPa)

20

f'c (MPa)
20
15
15
10
10 Ambient 5

5 TMC C2c 0
0 5 10 15 20 25 30
0 Standard cured f'c (MPa)
0 1 2 3 4 5
(a) Average concrete temperature 18°C
Age (days)

Fig. 10. Comparison of strengths between ambient cured and TMC cured 30
specimens.
25

TMC cured
20

f'c (MPa)
15
30 10

25 5
Strength (MPa)

0
20 0 5 10 15 20 25 30
Standard cured f'c (MPa)
15
TMC C2c
(b) Average concrete temperature 20°C
10
TMC C3d Fig. 12. TMC cured versus standard cured strengths.
5 Standard

0
0 1 2 3 4 5
1.2
f'c TMC cured / f'c Standard

Age (days)

Fig. 11. Comparison of strengths between standard cured and TMC cured 1
specimens.
cured

0.8

mens (Fig. 11), it can be seen that the strength values are close to 0.6
those of standard cured specimens. However, the specimens were
subjected to temperature variations which fluctuate around 23 °C
and expected to have the same maturity as those cured in standard 0.4
15 20 25 30 35
condition. The average of the temperature variations presented in
Average concrete temperature (°C)
Fig. 4a and b were found to be 18 °C and 20 °C for specimens C2c
and C3d, respectively. The linear trend presented in Fig. 12 shows (a) Mix 1
that the standard cured specimens could reasonably represent the
1.2
f'c TMC cured / f'c Standard

in situ strength. However, it is noted that the standard cured spec-

imens could overestimate the in situ strength if the average in situ
temperature falls far below 23 °C. 1
To investigate the sensitivity of strength prediction to the vari-
cured

ation in in situ temperature, further analyses were undertaken

0.8
based on experimental results obtained by Sirivivatnanon et al.
[31]. The mix used in the experimental program undertaken by
Ref. [31] is similar to the one presented in Table 2 (referred to as 0.6
mix 1). Furthermore, a similar mix with 25% additional fly ash con-
tent has been tested (referred as mix 2). The details of the mix and
0.4
experimental program can be found in [31]. 15 20 25 30 35
The ratio of the in situ compressive strength (TMC cured speci- Average concrete temperature (°C)
mens) to the compressive strength of standard cured specimens
(b) Mix 2
was plotted against the average in situ temperature variations in
Fig. 13. The standard cured specimens have been found to reason- Fig. 13. Comparison of strengths between TMC cured and standard cured speci-
ably predict the in situ strength provided that the average tempera- mens for different mixes and curing temperature.
ture variation is within 15–35 °C range. This range is representative
of the temperature variation in temperate zone such as southern and
eastern part of Australia, New Zealand, southern part of Europe, other regions with similar climate. Further studies are needed to test
California, southern part of South America and South Africa, and the validity of the finding for temperatures outside the range.
666 M. Sofi et al. / Construction and Building Materials 29 (2012) 659–666

4. Conclusions [12] Price WF, Hynes JP. In-situ strength testing of high strength concrete. Mag
Concr Res 1996;48(176):189–97.
[13] Guo C. Maturity of concrete: method for predicting early-stage strength. Am
The maturity function has been shown to provide a reliable esti- Concr Inst Mater J 1989;86(4):341–53.
mate of the strength development for the early-age concrete. A fur- [14] Oluokun FA, Burdette EG, Deatherage JH. Early-age concrete strength
prediction by maturity: another look. Am Concr Inst Mater J 1990;87(6):
ther sensitivity study found that the strength predictions are not
565–72.
significantly affected by the activation energy parameter. This [15] Tank RC, Carino NJ. Rate constant functions for strength development of
could potentially simplify the procedure of prediction of the concrete. Am Concr Inst Mater J 1991;88(1):74–83.
[16] De Schutter G, Taerwe L. Degree of hydration-based description of mechanical
strength using strength–maturity functions. More significantly,
properties of early age concrete. Mater Struct 1996:335–44.
comparisons of the strength values of TMC and standards cured [17] Pane I, Hansen W. Concrete hydration and mechanical properties
specimens have not shown significant differences. The current under nonisothermal conditions. Am Concr Inst Mater J 2002;99(6):534–42.
standards method of testing can be used to provide a reasonable [18] Hattel JH, Thorborg J. A numerical model for predicting the thermomechanical
conditions during hydration of early-age concrete. Appl Math Model
estimate of the strength development of early-age concrete. The 2003;27(1):1–26.
research findings are valid for average in situ temperature varia- [19] Carino NJ, Lew HS. The maturity method: from theory to application. In: Chang
tion between 15 and 35 °C. PC, editor. Structures congress and exposition. Washington, DC: American
Society of Civil Engineers; 2001. p. 19.
[20] Carino NJ, Malhotra VM. The maturity method. In: Malhotra VM, Carino NJ,
References editors. Handbook on non-destructive testing of concrete. CRC Press; 2004.
[21] Knudsen T. Modelling hydration of Portland cement: effect of particle size
[1] CPN29. Prestressed concrete anchorage zones. Concrete Institute of Australia; distribution. In: Henniker NH, editor. Engineering foundation conference on
1996. characterization and performance prediction of cement and concrete, Neville;
[2] AS1379. Specification and supply of concrete. Committee BD-049, Concrete 1982. p. 125.
Institute of Australia; 2007. [22] Freiesleben Hansen P, Pedersen EJ. Curing of concrete structures, Comité euro-
[3] AS 1012.8.1. Method of testing concrete: method of making and curing international du béton (CEB) information bulletin; 1985.
concrete – compression and indirect tensile test specimens. Committee BD- [23] Plowman JM. Maturity and the strength of concrete. Mag Concr Res
042, Australian Standards; 1999. 1956;8(22):13.
[4] ASTM C192/C192 M-10. Making and curing concrete test specimens in the [24] Choong J, Mendis P, Mak SL, Baweja D. More accurate concrete strength
field, American society for testing and materials (ASTM). West Conshohocken, predictions for post-tensioned slab construction. In: Concrete institute of
Pennsylvania; 2007. Australia biennial conference, Melbourne; 2005.
[5] AS 3600. Concrete structures. Committee BD-002, Australian Standards; 2001. [25] Mak SL. Factors influencing the in-situ strength in high strength concrete
[6] ACI318-02. Building code requirements for structural concrete (ACI 318-02) columns. PhD Thesis. Department of Civil Engineering, Monash University;
and commentary (ACI 318R-02). ACI Committee 318; 2002. 1993.
[7] Harrison TA. Temperature matched curing. London: Civil Engineering; 1983. p. [26] Mak SL, Attard MM, Ho DWS, Darvall PL. Cross-sectional strength gradients in
26–7. high strength concrete columns. Cem Concr Res 1994;24(1):139–49.
[8] Mak SL, Torii K. Strength development of high strength concretes with and [27] AS3972. Portland and blended cements. Committee BD-010, Australian
without silica fume under the influence of high hydration temperatures. Cem Standards; 1997.
Concr Res 1995;25(8):1791–802. [28] AS1379. Specification and supply of concrete. Committee BD-049, Australian
[9] Khan AA, Cook WD, Mitchell D. Early age compressive stress–strain properties Standards; 1997.
of low-, medium, and high-strength concretes. Am Concr Inst Mater J [29] Day KW. Concrete mix design, quality control and specification. London: E. &
1995;92(6):617–24. F.N. Spon; 2006.
[10] Mani AC, Tam CT, Lee SL. Influence of high early temperatures on properties of [30] Guirguis S. Cements – properties and characteristics. Technical note. Cement
PFA concrete. Cement Concr Compos 1990;12(2):109–15. and Concrete Association of Australia; 1998.
[11] Mak SL, Ritchie D. Estimating in-situ concrete properties non-destructively [31] Sirivivatnanon V, Baweja D, Khatri R. Evaluation of in-situ concrete
using temperature match conditioning. CSIRO Sustainable Materials strengths for post-tensioning of concrete slabs. Concr Forum 2009;2(1):
Engineering, Personal Communication. 1–10.