Vous êtes sur la page 1sur 5

Construction and Building Murerials, Vol. 10, No. 5, pp.

375-319, 1996
Copyright 0 1996 CSIRO. Published by Elsevier Science Ltd
Printed in Great Britain
0950-0618/96 S15.00+0.00
09504618(95)ooo15-1
ELSEVIER

A performance specification for durable


concrete

D. W. S. Ho* and G. J. Chirgwint

* CSlRO Division of Building, Construction and Engineering, Highett, Victoria,


Australia
t Bridge Branch, Roads and Traffic Authority of NSW, Australia

Received 6 January 1995

The lack of durability of reinforced concrete structures often refers to the corrosion of the reinforc-
ing steel and the subsequent spalling of the cover concrete. With the use of water sorptivity in quan-
tifying the quality of concrete, this paper discusses the deficiencies in the common practice of
specifying concrete for reinforcement protection. The paper also demonstrates how water sorptiv-
itv can be used as a simole oerformance criterion for specifying durable concrete. Copyright Q 1996
CiIRO. Published by Elsevi’er Science Ltd.

Keywords: concrete; durability; specification

Concrete can deteriorate in many ways. Perhaps the What are the controlling parameters for durability?
greatest concern nowadays is the corrosion of the re- And how can they be specified and, more import-
inforcing steel and the subsequent spalling of the cover antly, enforced in practice?
concrete. In Australia, the cost of repair of building If water to cement ratio is important, would cement
structures other than dwellings has been estimated in of different types and source provide a similar level
1978 to be about $50 million per annum’, representing of durability for the same water-cement ratio?
10% of that expended on new buildings. This could be If blended cements are appropriate for use in
translated to repair and maintenance costs in the order aggressive environments (e.g. marine), would they
of $800 million in today’s prices for buildings alone. be equally suitable for less aggressive exposure, such
These costs could run into billions of dollars if one as inland areas with a temperate climate?
takes into account other structures such as bridges, Can the quality and therefore the durability of
sewerage pipes etc. concrete be quantified?
Many attempts have been made over the years to What is the difference in concrete quality achieved
improve the properties of concrete. Chemical admix- with different curing techniques?
tures have been developed to reduce the water demand What is the effect of prolonged or lack of curing on
the long-term performance of concrete? A typical
of a mix, its workability, the setting time, and ultimately
specification would state: ‘Concrete shall be cured
the ease of construction. Mineral admixtures such as fly
for at least seven days’. Usually, the specification or
ash, granulated blast furnace slags and silica fume are
code of practice would add: ‘Concrete shall be
being incorporated in concrete as blended cements and
covered . . . and be kept continually moist . . . consid-
there is a general recognition by concrete technologists
eration will be given to the use of curing
that these admixtures when properly used enhance the
compounds’. These specifications, whilst sounding
properties of concrete, particularly in marine environ-
good in theory, do not provide an appropriate
ments. In Australia, there is a continued request by
contractual basis for action. It is difficult, if not
some practising engineers for information relating to impossible, for the superintendent to check whether
the proper use or specification of these admixtures in or not curing has been ‘effectively’ carried out.
order to achieve the level of durability required. Of
great concern to the Roads and Transport Authority of The above list of issues can go on for several pages.
New South Wales (RTA) is premature deterioration of Some of the questions raised may seem naive to the
some of the bridges2 constructed over the past 25 years. experts, but what sounds good in theory may not neces-
The high cost of repair and relocating bridge crossing sarily be practical, particularly in the area of curing.
points especially in urban areas favours the use of The crucial issue is how to specify relevant durability
highly durable structures. Questions or issues that are requirements that can be tested for conformance.
often raised are: ‘Construction reality’ cannot be over-emphasized!

375
376 A performance specification for durable concrete: 0. W. S. Ho and G. J. Chirgwin

This paper makes no attempt to answer all the ques- and the depth is found to be related to the square root
tions raised, but points out the deficiencies in the of time, t, such as:
current practice of specification. Solutions such as
performance criteria are offered. S = d/(t)“.5

In the laboratory, the depths of water penetration were


Quality of concrete measured at l/2, 2, 4, 8, 16 and 24 h.
For concrete durability, the properties of interest
Factors affecting water sorptivity
depend on the service and exposure conditions, For
road pavements, one would be particularly interested in The influence of various factors on water sorptivity has
the abrasion resistance and shrinkage of the concrete, been reported 8- 12.In view of the limited space available
whereas for dams and high strength concrete columns in this paper, only some examples are presented here to
the heat of hydration would be a major consideration. demonstrate the importance of mix constituents and
For concrete elements exposed to the atmosphere, curing on the quality of concrete.
such as building exteriors, and the reinforcement corro- The water-cement ratio of a mix is generally consid-
sion aspect dealt with in this paper, the penetrability of ered as one of the most important parameters affecting
concrete quality. However, due to the increasing use of
substances such as carbon dioxide, chloride ions,
blended cements of various types, there has always been
oxygen and water are of prime importance3. It is gener-
a question among engineers as to whether different
ally recognized that this penetrability is controlled by
types of cements would contribute to similar durability.
the pore structure of the cover concrete.
Confusion often exists when the two ratios, water/
In this paper the quality of concrete is assessed by
Portland cement (W/C) and water/binder (W/B), are
water sorptivity, a property relating to the pore struc-
used loosely. Binder is used in this paper to represent
ture of concrete near the surface. Depending on the
the total mass of Portland cement and mineral
exposure environment, other properties such as carbon-
admixtures.
ation, chloride ingress and oxygen diffusion could be
Results on three groups of mixes with 1, 7 and 28
used additionally to assess concrete quality. The inclu- days of standard moist curing are presented in Table 1.
sion of these properties is a topic for further discussion Fly ash (FA), ground granulated blast furnace slags
and is beyond the scope of this paper. (BFS) and silica fume (SF) are incorporated in some of
The concept of water sorptivity, S, has been the group a, b and c mixes. Different sources of Type A
presented previously4. It is a property describing the Normal Portland Cementsi were used in each group of
‘rate’ of water penetration due to capillary action. It concrete. All mixes had a similar slump of about 80
was pointed out that this property related to both the mm. Plain concretes (i.e. without admixtures) are used
initiation phase of reinforcement corrosion, by affecting as controls.
the ingress of soluble aggressive materials (e.g. chlorides As expected and indicated in Table 1, the quality of
from sea spray or sulphur dioxide from polluted atmos- concrete, as indicated by water sorptivity, is improved
pheres), and the propagation phase, by providing with increased curing time. Results also indicated that
moisture and lowering the resistivity of the concrete. even with equal 28-d strength (F28) and W/B or W/C,
Accordingly, good quality concrete is represented by the quality of concrete varied depending on the source
low values of water sorptivity. Similar approaches and type of materials used.
based on water absorption have also been studied by There is a general belief that concrete incorporating
other works5-7. mineral admixtures requires longer curing compared
During testing, the depth (not volume) of water pene- with plain concretes. This is valid provided the same
tration, d, is noted visually by splitting the specimen, Portland cement is used in both mixes. However, as

Table 1 Water sorptivity of concretes with 1, 7 and 28 days of standard curing. Plain mixes with no admixtures, WR-mix with a water-reducing
agent, mixes with fly ash (FA), blast furnace slag (BFS) or silica fume (SF), mix with both water-reducing agent and fly ash (FW) or blast furnace
slag (SW).

Sorptivity (mm l?)


F28
Concrete WE W/B hIPa) Id 7d 28d

(a) Plain 0.69 _ 27.5 15 7 5


WR 0.66 26.0 8 6 3.5
FA 0.80 0.64 26.5 17 10 5.0
FW 0.80 0.64 26.0 14.5 7.5 3.0

(b) Plain 0.69 28 18.0 10 5.5


BFS 1.06 0.69 28 26.8 11 6.5
SW 1.06 0.69 28 20.5 7 4.5

(c) Plain 0.58 30 7.0 3 2


SF 0.83 0.78 30 12.0 4 2.5
A performance specification for durable concrete: D. W. S. Ho and G. J. Chirgwin 377

indicated by the SF mix of group C, it has lower sorp- moisture content of the concrete fell below 100% RH.
tivity values compared with the plain concretes of both These observations were based on data using 20 to 50
groups a and b. These results indicate that, not only do MPa graded concrete. For concrete with high strengths,
the admixtures have a large influence on concrete the trend may or may not be similar.
quality, but also the source of the Portland cement
used. In order to achieve a low sorptivity value, a Water sorptivity as a performance criterion
proper combination of materials has to be selected.
Thus, one should not make general observations on the The common practice for specifying concrete for
use of mineral admixtures or any material without also durable structures is by the 28-day compressive
taking into account the effect of other mix constituents. strength, W/B and/or binder content together with
Similar findings on mixes incorporating fly ash were curing requirements. It is obvious from the discussion
also obtained when properties such as early strength earlier that no single parameter or a combination of
developmenti and pore structurei were discussed. parameters can be used to assess the performance of
The influence of W/C, cement content and 28day concrete.
strength are further illustrated in Table 2. As indicated, One possible way to ensure durability is to specify
comparable concretes made using Portland cements directly the required quality of concrete such as its resis-
Cl and C2 had similar W/C ratios, Portland cement tance to carbonation, chloride penetration, water sorp-
contents and 28day strengths. Yet the sorptivity values tivity and oxygen diffusion. Depending on the type of
of one set of mixes were about 30% higher than those of structure and the service environment, more than one
other corresponding mixes from the other set. This was criterion may need to be specified. Somerville*9 high-
explained i6 by the low early strength development as lighted the potential of using water permeability as a
indicated by the 7- to 28-day strength ratios (F7/F28) criterion in engineering designs. Otherszo~*’ went one
for mixes incorporating cement C2. step further by proposing regular concrete permeability
testing alongside conventional strength testing. The use
of carbonation as a performance criterion has been
Effectiveness of curing techniques
suggested**.
The specification of concrete is further complicated by Water sorptivity was adopted by the RTA as a
the availability of different on-site curing techniques performance criterion in their specification of concrete.
and the general lack of understanding of their effective- This property was chosen because, besides its relevance
ness among practising engineers. to durability, it is simple to determine and without the
On-site curing techniques, for practical reasons, need for special moulds or equipment. Furthermore it
generally rely on the prevention of moisture loss from takes into account the requirements for curing.
concrete by methods such as maintaining formwork in However, one complication in the use of the perfor-
place, covering with polyethylene sheet, or the applica- mance criterion is the necessity in specifying acceptable
tion of a curing compound. In a laboratory studyi7, it quality levels for different exposure environments. In
was found that there was significant improvement in the view of this, Ho and Lewis23 recommended interim
surface quality when cured with the above techniques to limiting values of water sorptivity for building struc-
three days, and some additional improvement when tures with a service life of around 50 years.
curing was extended to seven days, but little or no
benefit arose for curing beyond this period. In other
RTA approach
words, there is no technical benefit of keeping concrete
behind formwork for more than seven days. Based on Bridges for the RTA require a service life of 100 years.
water sorptivity measurements, it was found that the In order to obtain adequate data for the determination
surface quality achieved with the above on-site curing of appropriate quality levels, the concept of water sorp-
procedures for seven days was generally equivalent to tivity as an additional durability requirement went on
about three days of standard curing. This was trial in 1990. Testing procedures for sorptivity were
explained l* by the ineffectiveness of curing when the further simplified by determining only the depth of

Table 2 Properties of plain concretes incorporating normal Portland cements Cl and C2

Cement content F28


Concrete w/c (kg m-‘) (MPa) F7lF28 Sorptivity (mm h-5

Plain (Cl) 0.42 405 45 0.95 3.5


0.56 316 38 0.85 5.5
0.69 268 27 0.73 7.0
0.77 248 22 0.78 9.5

Plain (C2) 0.44 401 47 0.79 5.0


0.57 310 38 0.70 6.5
0.69 267 29 0.63 8.5
0.77 246 23 0.68 12.5
378 A performance specification for durable concrete: D. W. S. Ho and G. J. Chirgwin

Table 3 B80 durability requirements for concrete

Exposure Minimum Minimum period of


binder content Maximum standard curing
Classification Environment (kg m-‘) W/B (days)

A (Dry climate, no industry _ 0.56 7 (7)


non-aggressive)

B1 (Industrial areas, inland) 320 0.56 7 (7)


B2 (Close to coast or 390 0.46 7 (14)
permanently in salt water)

C (Tidal/splash zone) 450 0.40 14 (21)


“Figures in brackets refer to high slag cement

water penetration after 24 h of wetting and is referred goal is a ‘user friendly’ specification that can be used to
to as the sorptivity penetration depth. ensure maintenance-free structures over their design life.
In the second edition of B80 Concrete Work for
Bridges25, the design of the curing regime has been
Conclusion
made part of the mix design and acceptance process.
The contractor shall propose only such curing regime as The quality of concrete required for reinforcement
it has confidence can and will be implemented for the protection was assessed in terms of water sorptivity.
relevant concrete in the field. The specification gives This property was found to vary depending on the
‘deemed to comply’ moist curing periods under reason- source and type of the cement used. The paper demon-
able ambient conditions for conforming mix designs strated that concrete quality cannot be predicted by mix
(Table 3). The contractor is free to choose any curing design parameters such as strength and water to binder
regime, provided that the tests demonstrate the perfor- (or Portland to cement) ratio due to the interactions of
mance of the mix when cured in accordance with the mix constituents. Specifications were further compli-
contractor’s design is satisfactory. cated by the different effectiveness of on-site curing
The acceptability of the mix and curing design is techniques.
determined by the resulting sorptivity and other Specification based on performance criterion was
strength requirements. The contractor’s proposed mix suggested. This concept was adopted by the Roads and
and curing design must be no weaker or more perme- Traffic Authority of New South Wales and has been on
able than those obtained from the ‘deemed to comply’ trial since 1990. With data generated on water sorptiv-
mixes shown in Table 3. ity over the past years, maximum sorptivity limits have
Over the last few years, many data on sorptivity been set and formed part of the durability requirements
penetration depth were generated from mixes that satis- in the latest edition of their B80 document - Concrete
fied Table 3 requirements. From these data, maximum for Bridge Works. Furthermore, the design of the
sorptivity levels for various exposure classifications curing regime has been made part of the mix design and
have been determined and became an additional acceptance process.
durability requirement in the third edition (1994) of the
B80 document. The maximum sorptivity depths using
proposed curing regime are:
References
1 Beresford, F.D. and Ho, D.W.S. Proc. Biennial Conjtirence of
Concrete Institute of Australia, Cunberra 1919
Exposure classification A Sorptivity depth 45 mm 2 Chirgwin, G.J. Seminar on How to Specijj and Produce Concrete
Bl 35 mm for Durable Structures, Concrete Institute of Australia,
B2 17 mm Melbourne, 1994
Tuutti, K. Swedish Cement and Concrete Research Institute,
C 11 mm
Stockholm, No. 4, 1982
Ho, D.W.S. and Lewis, R.K. Civil Eng. Trans. 1984 CE26(4), 306
With the additional durability requirement on water Senbetta, E. and Scholer, C.F. Proc. Second Int. Conf&ence on
Durability of Building MuteriaO and Components, Maryland,
sorptivity, the incidence of damage due to reinforce- 1981, p. 153 e
ment corrosion is expected to be greatly reduced. Hall, C. and Yau, M.H.R. Build. Environ. 1987, 22 (l), 77
Structures with this additional requirement are being Aldred, J.M. Fourth Internutional Conjtirence on Durubiliry oj
Building Muferials and Components, Singapore, 1987, p. 152
monitored for their long-term performance. The RTA Ho, D.W.S. and Lewis, R.K. Dumb. Build. Mater. 1987, No. 4,
recognizes that the use of sorptivity alone as a perfor- 241
mance test may not be adequate to ensure durability Ho. D.W.S. and Lewis, R.K. Durub. Build. Mater. 1987, No. 5, 1
Ho, D.W.S. and Lewis, R.K. Fourth Int. Conjerence on
under all situations, but considers it a step in the right
Durability of Building Materia1.r and Components. Singapore,
direction. Other criteria on chloride diffusion are being 1987, p. 508
assessed and are under consideration for use in marine 11 Ho, D.W.S., Hinczak, I., Conroy, J.J. and Lewis, R.K. American
structures. The B80 is an on-going document which will Concrete Institute SP 91, 1986, 1463
12 Ho, D.W.S. and Lewis, R.K. Proc. CANMET/ACI International
be modified from time to time as more research data Workshop on Silica Fume in Concrele, Wushington, D.C., 1991,
and field information become available. The ultimate p. 145
A performance specification for durable concrete: D. W. S. Ho and G. J. Chirgwin 379

13 AS 1315 SAA Portland Cement, Australian Standard, 1982 19 Somerville, G. Struct. Eng. 1986, 654 (2), 60
14 Ho, D.W.S. and Lewis, R.K. Proc. Concrefe 85 Conference, 20 Pritchard, B. Constr. Repairs Mainten. 1986, 2 (6), 6
Brisbane, Australia, 1985, p. 95 21 O’Brien, T., Cather, R. and Figg. J. American Concrete Institute
15 Sydney, 1990 and Cao, H.T. Proc. Onoda Pacific Conference,
Ho, D.W.S. SPIOO, 1987, 257
16 Ho, D.W.S. and Lewis, R.K. J. Cement Concr. Res. 1988, 18 (4) 22 Lewis, R.K. and Ho, D.W.S. CIA News, 1989, 15 (4), 8
584 23 Ho, D.W.S. and Lewis, R.K. Proc. Concrete jbr the Nineties,
17 Ho, D.W.S. and Lewis, R.K. Concr. Ausfralia, 1992, 18 (2) 3 Leura, Australia, 1990, 10
18 Ho, D.W.S., Cui, Q.Y. and Ritchie, D.J. Cement Concr. Res. 24 Part B80 - Concrete for Bridgeworks, Ed. 1 Revision 0, Roads
1989, 19 (3) 457 and Traffic Authority of New South Wales, 1990, p. 49

Vous aimerez peut-être aussi