You are on page 1of 6

Evaluation of Creep, Shrinkage and Modulus of elasticity models for

High Strength Concrete


N. Baidya & P. Mendis

University of Melbourne, Australia

S. Fragomeni

Victoria University, Australia

ABSTRACT: In the literature and codes of practice, various equations are recommended for calculating
typical material properties of compressive strength, elastic modulus, shrinkage strain and creep coefficient of
concrete. The various equations are reported to give significantly different estimates leading to great uncertainty, particularly when estimating axial shortening of vertical concrete elements such as columns, cores and
walls in a tall building that typically utilises high strength concrete. The main aim of this paper is to evaluate
the various equations from relevant concrete codes and literature used to estimate the material properties of
both normal strength concrete (NSC) and high strength concrete (HSC). Selected prediction equations are also compared to recent laboratory results from site delivered HSC.
1 INTRODUCTION
Axial shortening in a concrete column can be expressed as a summation of elastic strain caused by
any load application, creep strain developed by a
sustained stress over a long-term period and shrinkage strain induced by the drying of concrete. Reinforcement in the concrete member also provides
some restraining effects to axial shortening. Gilbert
(1988) gives the total strain (t) at any time (t) at a
point in a uniaxially loaded specimen at constant
temperature as

(t) = e(t) + c (t) + sh(t)

(1)

where e (t) = instantaneous strain; c (t) = creep


strain; and sh (t) = shrinkage strain.
Recent research has focused on producing elastic
modulus, shrinkage and creep models for high
strength concrete (HSC)(Ahmad & Shah 1985, Carrasquillo et al. 1981, Gilbert 2002, Huo et al. 2001,
Mendis et al. 1997, Mokhtarzadeh & French
2000a,b, Setunge 1994 and Mazloom et al. 2004). In
some cases existing material models for normal
strength concrete (NSC) have been modified for
HSC.
Until recently guidelines on HSC in international
concrete codes were limited, and it was left to the
structural designer to select equations describing the
relevant material properties. This practice is not appropriate for design of HSC structures and may be
problematic (Carrasquillo et al. 1981). Proposed
equations for prediction of material properties of
concrete are varied in complexity and reported to
give significantly different results. Some are simple
and easy to use, but others are much more compli-

cated, involving the establishment of coefficients


which account for many parameters. These parameters affect the magnitude and rate of development of
elastic modulus, shrinkage and creep. Unfortunately,
an increase in complexity does not necessarily mean
an increase in accuracy (Gilbert 1988).
The main objective of this paper is to present and
evaluate the equations used to estimate the material
models of NSC and HSC from codes and literature.
Also the performance of selected prediction models
is evaluated against HSC laboratory results from
Holcim, Australia (formerly CSR Construction Materials - Investigation report Bin 243, 2002). This
concrete was utilised in the construction of columns
and walls of the 258 m high (86 storeys) World
Tower Building, Sydney (Baidya, 2005).
2 NORMAL STRENGTH CONCRETE MODELS
2.1 ACI (American Concrete Institute)
The method for predicting compressive strength,
elastic modulus, shrinkage strain and creep coefficient given in ACI 209R-92 (1992) is based on the
work done by Branson & Christiason (1971). These
methods give reasonable results but correlations between the field and model predicted values are said
to be inconsistent (Gilbert 1988). Note the definitions of symbols in all equations are given in the notation section.
The ACI Committee 318 (2008) employs the following expression developed by Pauw (1960) for
predicting the elastic modulus of concrete (MPa) at
age t (days).

Ec(t) = 0.043 1.5

f c (t)

(2)

t
f c (28)
+t

(3)

with

f c(t) =

The shrinkage strain measured from the start of


drying (after 7 days moist curing), is given by:

sh (t) =

t
*sh
35 + t

(4)

Finally, the creep coefficient for moist cured concrete at time t is given by:

(t ,) =

(t ) 0.6
* ( )
0. 6
10 + (t - )

(5)

2.2 AS 3600 (Australian Concrete Standard)


The recommendations given in CEB-FIP (1970) for
time-dependent models, forms the basis for the relevant clauses in the previous AS 3600 (2001), valid
for compressive strengths up to 65 MPa. It is widely
published that the inaccuracy of the results obtained
from these NSC equations for calculating material
properties increases when compared with the results
from HSC equations (Mokhtarzadeh & French
2000a).
To calculate the elastic modulus, AS 3600 (2001)
recommends Equation 2, as proposed by Pauw
(1960) with the compressive strength at age t (days)
is also given by Equation 3.
The shrinkage strain at any time t after the commencement of drying is given as:

sh (t) = 0.00085 k1

(6)

The creep coefficient at time t due to a sustained


stress first applied at age is given by:

(t ,) = k 2 k 3 cc.b

(7)

These models have been significantly updated in


the new AS 3600 (2009) code to account for HSC up
to 100 MPa. These are discussed in the next section.
3 HIGH STRENGTH CONCRETE MODELS
HSC has been broadly defined as a concrete having
a 28-days characteristic compressive strength in the
range of 40 and 100 MPa (Ahmad & Shah 1985,
Carrasquillo et al. 1981 and Gilbert 2002). However,
the lower value of this strength range is 50 MPa in
some of the codes. As outlined by Ahmad and Shah
(1985), the main advantage of HSC is that it has relatively greater compressive strength to unit cost, unit
weight and unit volume ratios as compared to NSC.

HSC structures are stiffer and stronger and at the


same time comparable in cost to NSC (Ahmad &
Shah 1985).
Carrasquillo et al. (1981) conducted an experimental investigation of the properties of HSC. The
test materials used were Type I Portland cement,
gravel or crushed limestone coarse aggregate, sand
from a local deposit and for some mixes, a waterreducing retarding admixture. Uniaxial compressive
strengths ranged from about 21 to 76 MPa. The authors studied compressive strength, strain gain with
age, specimen size effect, effects of drying, stressstrain curves, static modulus of elasticity, Poissons
ratio, modulus of rupture and split cylinder strength.
They noted significant differences between the performance of HSC and NSC.
Huo et al. (2001) conducted an experimental
study of time-dependent material properties over
two years, with local materials from Nebraska, USA.
The study included compressive strength, modulus
of elasticity, shrinkage and creep of HSC mixtures.
They compared the data obtained with ACI 209R-92
and found that shrinkage strains and creep coefficients of HSC were lower than those of conventional
concrete, resulting in over-estimation of those values
of HSC by ACI 209R-92. It was also reported that
the ACI 209R-92 code produced erroneous results
for the elastic modulus. They developed prediction
equations for modulus of elasticity, shrinkage and
creep based on the test results for HSC made from
Nebraskan materials only.
Mokhtarzadeh & French (2000a, b) conducted
various tests investigating the effects of various materials on the mechanical properties of HSC, with a
compressive strength in the range of 55 to 128 MPa.
More than 6000 HSC concrete specimens from 142
mixtures for compressive strength, modulus of elasticity, shrinkage and creep were tested. The experimental results were compared to code relationships
and prediction equations proposed for HSC.
Mazloom et al. (2004) presented the development
of formulas to estimate the long-term shrinkage and
creep of HSC containing silica fume. The experimental part of the work focused on concrete mixes
having a fixed water/binder ratio of 0.35 and a constant total binder content of 500 kg/m3. The percentage of silica fume that replaced cement in this research was: 0%, 6%, 10% and 15%. Using
regression analysis, equations were proposed for
predicting time-dependent shrinkage and creep of
HSC.
The prediction of shrinkage and creep of concrete
by the CEB-FIP model code 1990 is restricted to ordinary structural concretes, having cylinder strengths
varying from 12 to 80 MPa and mean relative humidity 40-100% at mean temperature 5-30C.
The prediction of material parameters of the Bazant & Baweja model B3 (2000) is restricted to the
Portland cement concretes, having strengths varying

from 17 to 70 MPa, w/c ratio 0.30-0.85, a/c ratio


2.5-13.5 and cement content 160-720 kg/m3.
The British Standard is the official English language version of Eurocode 2 (EN 1992-1-1:2004).
In the UK, the corresponding national standards are:
BS 8110-1:1997; BS 8110-2:1985; and BS 81103:1985. Use of the Eurocodes provides more opportunity for designers to work throughout Europe.
Higher strengths of concrete are covered by Eurocode 2 (2004), up to class C90/105. However, because the characteristics of higher strength concrete
are different, some expressions in the Eurocode are
adjusted for classes above C50/60.
Recently, procedures for the estimation of the deformation characteristics of concrete were revised to
include HSC in AS 3600 (2009) with characteristic
compressive strengths up to 100 MPa. The elastic
modulus, shrinkage and creep models follow the
recommendations by Gilbert (2002).

sh = ( sh )u

E c = 3320

f c' + 6900

(8)

Alternately, Ahmad & Shah (1985) proposed the


following equation:

Ec(t) = 3.385 10 -5 2.5 ( f c' )0.325

(9)

Equation 9 is comparable to Equation 2 for low


and normal strength concrete, but is said to be more
accurate for HSC.
Shrinkage strain
Mokhtarzadeh & French (2000b) proposed Equation
10 to predict shrinkage strains (sh)t for HSC (moist
cured) at any time t.

t
sh (t) =
( sh )u
45 + t
(sh)u = 530 microstrain.

(10)

The observed smaller ultimate shrinkage of HSC


is attributed to the lower water-cement ratio (w/c)
and denser matrix of HSC.
Huo et al. (2001) suggested a modification to the
shrinkage strain given in ACI 209R-92 to cover
HSC. The prediction equation for shrinkage strain is
expressed as:

( Ks + t )

(11)

Recently, Mazloom et al. (2004) proposed Equation 12 to predict total (drying and autogenous)
shrinkage strains sh(t) for HSC at any time t.

sh (t) =

516 y ( t ) 10 -6
( 0.3 SF + 12.6) + t

y = 1.14 - 0.007 (V/S)

(12)
(13)

3.1.2 Creep coefficient


Huo et al. (2001) suggested a modification to the
creep coefficient given in ACI 209R-92 to cover
HSC given as:

vt = ( v )u

3.1 Modified equations related to ACI


3.1.1 Elastic Modulus
Equation 2 substantially overestimates the measured
modulus of elasticity for compressive strength 41
MPa or more. Therefore, for normal density concrete
with 21< fc < 83 MPa, Equation 8 is recommended
by Carrasquillo et al. (1981).

t 0.6

(14)

K c + t 0.6

Mazloom et al. (2004) defined the creep coefficient as:

(t, to ) = C(t, t o ) Ec ( to )
C(t,t o ) =

(103 - 3.65 SF) y ( t )0.6 10 -6


( 26.5 - SF) + ( t )0.6

y = 1.08 - 0.0114 t o

(15)
(16)
(17)

3.2 Modified equations related to AS 3600


3.2.1 Elastic Modulus
According to Setunge (1994), the modulus of elasticity (Ec) of HSC is dependent on parameters such
as the volume of aggregates, modulus of the aggregates and the modulus of the paste used in AS 3600.
Consequently, Mendis et al. (1997) suggested a simplified empirical Equation 18 to predict the elastic
modulus of all grades of concrete.

E c = 0.043 1.5

f c' 20 %

= 1.1 - 0.002 f c' 1.0

(18)
(19)

Gilbert (2002) states that Equation 2 works well


for NSC but it overestimates Ec when fcm exceeds 40
MPa, therefore it is recommended for use when fcm
40 MPa only. For high strength concrete, the following equation was proposed and adopted in AS 3600
(2009) for fcm > 40 MPa.

Ec = 1.5 (0.024 f cm + 0.12)

(20)

3.2.2 Shrinkage strain


According to McDonald & Roper (1993), the previous Australian Code AS 1481 (1974) formula was
found to perform better than other equations for the
prediction of shrinkage strain. This equation is essentially based on CEB-FIP (1970).
The shrinkage strain r for moist cured normal
weight concrete at time t measured from the commencement of drying concrete is obtained from the
expression:

r = b ke k h

(21)

Gilbert (2002) proposed Equation 22 for estimating shrinkage strain in HSC, which is now adopted
by AS 3600 (2009). This equation is a modified version of the shrinkage equation in AS 3600 (2001),
dividing the total shrinkage strain cs into two components: (i) endogenous shrinkage strain cse and (ii)
drying shrinkage strain csd.

cs = cse + csd

(22)

At any time t (days) after pouring, the endogenous shrinkage strain is given by

cse = cse* (1.0 e - 0.1t )

(23)

After the commencement of drying at any time t,


the drying shrinkage strain may be calculated as:

csd = k1 k 4 csd .b

(24)

3.2.3. Creep coefficient

Gilbert (2002) has suggested Equation 25 for the


creep coefficient to cover compressive strengths up
to 100 MPa. This creep equation has also been
adopted by AS 3600 (2009).
The creep coefficient at any time t is given by
Equation 25.

cc = k 2 k 3 k 4 k 5 cc,b

(25)

Equation 25, although similar to Equation 7, introduces two new k-factors (k4 and k5), revises the
factors k2 and k3, and modifies the basic creep coefficient cc.b

3.4 Eurocode 2 (2004)


The modulus of elasticity (MPa) at age t (days) for
normal weight concrete, is given as:

Ecm(t) = ( f cm(t) / f cm ) 0.3 Ecm

(28)

The total shrinkage strain cs at any time t after


the commencement of drying is as follows:

cs = cd + ca

(29)

where cd is development of drying shrinkage strain


with time and ca is autogenous shrinkage strain are
both fully defined in Eurocode 2 (2004).
The creep coefficient (t, to) is given by:

(t , t o ) = o c (t , t o )

(30)

4 LABORATORY TESTS VERSUS PREDICTIONS


This section utilizes laboratory results of site delivered HSC, Mix 97605 with compressive strength of
105 MPa, used in columns/walls of the World Tower
Building in Sydney (Investigation report Bin 243,
2002). The preparation and testing of specimens for
modulus of elasticity, shrinkage and creep followed
the procedures given in relevant parts of AS 1012.
For creep measurement, cylinder specimens
(300x150mm diameter) strains due to creep were
measured after the specimens were subjected to a
stress of 42 MPa (i.e. 40% of the 28-days compressive strength of 105 MPa). The laboratory trends
represent the best fit from a number of results.
The predicted values using selected equations of
elastic modulus, shrinkage strain and creep coefficient along with the corresponding laboratory values
are compared in Figures 1-3. Equations relating to
AS 3600 and ACI were used for the comparisons.
4.1 Elastic modulus comparison
As shown in Figure 1, the selected equations overestimated elastic modulus for concrete strengths
more than 90 MPa.

3.3 CEB-FIP (1990)


The total shrinkage or swelling strain cs (t, ts) is calculated as:

cs (t ,t s ) = cso s ( t t s )

(26)

The creep coefficient is given by:

(t ,t o ) = o c ( t t o )

(27)
Figure 1. Elastic modulus versus compressive strength for laboratory test and theoretical prediction

Reasonable prediction is given by Carrasquillo et


al. (1981) and Ahmad & Shah (1985), where as
Mendis et al. (1997) and Gilbert 2002 (AS 3600
2009) models give moderately accurate predictions.
It is worth noting however that Fragomeni et al.
(2009) found the Gilbert (2002) equation to be a
good predictor of elastic modulus for concrete
strengths greater than 65 MPa.
4.2 Shrinkage strain comparison
In Figure 2, it is noted that all equations except
McDonald and Roper (1993) grossly over-estimate
shrinkage from the beginning of shrinkage time. It
seems that Huo et al. (2001) gives reasonable estimation of shrinkage after 90 days but over estimation prior to this.

tests indicates a wide variation of results for these


properties. This suggests that numerically accurate
estimations of elastic modulus, shrinkage and creep
and their effects on structural behaviour is extremely
difficult. Therefore, it is important at the design
stage to consider a range of values for both shrinkage strain and creep coefficient. This would then
help to establish the upper and lower limits when using the models for predictions such as axial shortening.
As indicated by the trends in Figures 1-3, the behaviour of material properties of HSC follows the
general trend of material properties of NSC. It is
however obvious that much more experimental work
from a range of sources is required to develop great
confidence in prediction equations for elastic modulus, shrinkage strain and creep of HSC. Interestingly
the Huo et al. (2001) and Gilbert (2002) models performed reasonably well, with both studies were aims
at producing HSC models.
6 ACKNOWLEDGMENT
The authors acknowledge the partial funding and
support provided by Holcim Australia for this study.
NOTATIONS

Figure 2. Shrinkage strain versus time (Mix 97605, 105 MPa).

4.3 Creep coefficient comparisons


For creep prediction, it is noted in Figure 3 that the
Huo et al. (2001) model is generally closer to the lab
test results than the other equations. The second best
equation is Gilbert (2002), incorporated in AS 3600
(2009), which under-estimated results prior to 60
days and then over-estimated results afterwards.

a/c

C(t,to)
Ec, Ec(t)

=
=

Ecm

fc(t)
fc,fc(28)

=
=

fcm

fcm(t)

Kc

Ks
ke

=
=

kh

Figure 3. Creep coefficient versus time (Mix 97605, 105 MPa).

k1 =

5 CONCLUSIONS
The broad comparison of the predicted results using
different equations of elastic modulus, shrinkage
strain and creep coefficient with specific laboratory

k2, k3, k5

ratio (by weight) of aggregate to


cement
specific creep after t days loading
concrete elastic modulus at age t
(MPa)
concrete elastic modulus of 28 days
(MPa)
compressive strength at age t (MPa)
compressive strength at 28-days
(MPa)
mean compressive strength of concrete at 28 days (MPa)
mean compressive strength of concrete at an age t days (MPa)
adjustment for early age creep coefficient
adjustment for early age shrinkage
shrinkage strain coefficient depends
on theoretical thickness of member
shrinkage strain coefficient depends
on the development of creep and
shrinkage with time; coefficient depending on the notional size ho
modification factor for shrinkage
strain depends on relative humidity
and hypothetical thickness
modification factor for creep coefficient

k4

SF
t

=
=

to ,
ts

=
=

(t-)
V/S
(vu)
w/c

=
=
=
=

, =
c =
s =
b =
ca =
cd =

cd.o =
csd.b =

modification factor for shrinkage


strain and creep coefficient depends
on the environment
silica fume (in percentage)
age of concrete (days); time in days
after loading; time after the end of
the initial wet curing (days)
age of concrete (days) at loading
age of concrete (days) at the beginning of shrinkage or swelling
duration of loading (days)
volume to surface ratio of concrete
ultimate creep coefficient
ratio (by weight) of water to cementitious material
correction factor according to the
loading age to
constant used for compressive
strength
coefficient to describe the development of creep with time after loading
coefficient to describe the development of shrinkage with time
basic shrinkage strain
autogenous shrinkage strain
drying shrinkage strain
basic drying shrinkage strain
basic drying shrinkage strain

cse* = final endogenous shrinkage strain


cso = notional shrinkage coefficient

(sh)u

ultimate shrinkage strain

*sh = final shrinkage strain at time infinity


= coefficient for modulus of elasticity
= density of concrete (kg/m3)

cc.b = basic creep coefficient


o , o = notional creep coefficient
*() = final creep coefficient
REFERENCES

ACI Committee 209. 1992. Predictions of Creep, Shrinkage


and Temperature Effects in Concrete Structures (ACI
209R-92). SP-70, American Concrete Institute, Farmington
Hills, Michigan.
ACI Committee 318. 2008. Building Code Requirements for
Structural Concrete (ACI 318-08) and Commentary. American Concrete Institute, Detroit.
Ahmad, S. H., & Shah, S. P. 1985. Structural Properties of
High-Strength Concrete and its Implications for Precast
Prestressed Concrete. PCI Journal 30(6): 92-119.
Australian Standard AS 3600 2009. Concrete Structures,
Standards Australia Committee BD-002, Sydney, Australia,
p. 213.
Australian Standard AS 3600 2001. Concrete Structures,
Standards Association of Australia, Sydney, Australia,
p.176.

Baidya, N. N. 2005. Axial Shortening of Vertical Elements in


Tall Buildings Incorporating High Strength Concrete, ME
Thesis, Department of Civil and Environmental Engineering, University of Melbourne.
Bazant, Z. P. & Baweja, S. 2000. Creep and shrinkage prediction model for analysis and design of concrete structures:
Model B3, American Concrete Institute, Special Publication SP-194, A. Al-Manasseer, ed., Farmington Hills,
Michigan: 1-83.
Branson, D. E. & Christiason, M. L. 1971. Time Dependent
Concrete Properties Related to Design-Strength and Elastic
Properties, Creep and Shrinkage. Designing for Effects of
Creep, Shrinkage and Temperature in Concrete Structures,
SP-27, American Concrete Institute, Detroit: 257-278.
Carasquillo, R. L., Slate, F. O. & Nilson, A. H. 1981. Properties of High Strength Concrete Subject to Short-Term
Loads, ACI Journal 78(3): 171-178.
CEB-FIP Model Code 1990. First predraft 1988. Bulletin d Information No 190 a et b, Juillet 1988.
Comite Euro-International du Beton 1970. CEB-FIP International Recommendations for the Design and Construction
of Concrete Structures, Paris, London.
Eurocode 2 2004. Design of concrete structures. Part 1-1: General rules and rules for buildings.
Fragomeni, S., Baweja, D. & Mendis, P. 2009. Measured and
Predicted Axial Shortening in Tall Buildings: Recent Research Findings, 24th Biennial Conference of the Concrete
Institute of Australia, 1719 September, Sydney.
Gilbert, R. I. 1988. Time Effects in Concrete Structures, Elsevier Science Publishers, Amsterdam, p. 321.
Gilbert, R. I. 2002. Creep and Shrinkage Models for High
Strength Concrete - Proposals for inclusion in AS3600,
Australian Journal of Structural Engineering, Institution of
Engineers, Australia 4(2): 95-106.
Huo, X. S., Al-Omaishi, N. & Tadros, M.K. 2001. Creep,
Shrinkage, and Modulus of Elasticity of High-Performance
Concrete, ACI Materials Journal 98(6): 440-449.
Investigation Report Bin 243, 2002. Creep Characteristics of
Concrete Mixes Report I, in confidence to CSR Construction Materials, Prepared by Radhe Khatri and Ben Link,
CSIRO Building, Construction and Engineering, North
Ryde, Sydney, Australia, June, Unpublished Report.
Mazloom, M., Ramezanianpour, A.A. & Brooks, J.J. 2004. Effect of silica fume on mechanical properties of highstrength concrete, Cement & Concrete Composites 26: 347357.
McDonald, D. B. & Roper, H. 1993. Accuracy of Prediction
Models for Shrinkage of Concrete, ACI Materials Journal
90 (3): 265-271.
Mendis, P. A., Pendyala, R. S. & Setunge, S. 1997. Requirements for High-Strength Concrete in AS 3600, HighPerformance Concrete Sub-committee of the Concrete Institute of Australia (Victoria), Melbourne.
Mokhtarzadeh, A. & French, C. 2000a. Mechanical Properties
of High-Strength Concrete with Consideration for Precast
Applications, ACI Materials Journal 97(2): 136-147.
Mokhtarzadeh, A. & French, C. 2000b. Time-Dependent Properties of High-Strength Concrete with Consideration for
Precast Applications, ACI Materials Journal 97(3): 263271.
Pauw, A. 1960. Static Modulus of Elasticity of Concrete as Affected by Density, ACI Journal 32(6): 679-687.
Standards Association of Australia, AS 1481 1974, SAA Prestressed Concrete Code, Sydney.
Setunge, S. 1994. Engineering Properties of High Performance
Concrete, Paper presented at the seminar on High Performance Concrete: Technology, Design and Applications,
University of Melbourne, Australia, February.

You might also like