You are on page 1of 7

Construction and Building Materials 36 (2012) 585591

Contents lists available at SciVerse ScienceDirect

Construction and Building Materials


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

A method for the use of accelerated carbonation tests in durability design


R. Neves b, F.A. Branco a,c, J. de Brito a,c,
a

DECivil-IST, Technical University of Lisbon, Av. Rovisco Pais, 1049-001 Lisbon, Portugal
ESTBarreiro, Polytechnic Institute of Setbal, R. Amrico da Silva Marinho, 2839-001 Lavradio, Portugal
c
Department of Civil Engineering and Architecture, Section of Construction, IST, Technical University of Lisbon, Av. Rovisco Pais, 1049-001 Lisbon, Portugal
b

h i g h l i g h t s
" Carbonation model calibrated with long term results.
" Deterioration model considering only initiation stage.
" Safety factors determined using Monte Carlo simulations.
" Comparative study with model code for service life design semi-probabilistic approach.

a r t i c l e

i n f o

Article history:
Received 20 March 2012
Received in revised form 22 May 2012
Accepted 4 June 2012
Available online 10 July 2012
Keywords:
Concrete
Carbonation
Corrosion
Service life
Performance-based design

a b s t r a c t
This work aims to provide a simple semi-probabilistic approach to the service life design of reinforced
concrete structures with respect to reinforcement corrosion induced by concrete carbonation.
This paper presents an analytical model for the initiation period, calibrated with long-term carbonation
results, which uses the accelerated carbonation resistance and the environmental class as input parameters. The issue of dening a limit for deterioration is discussed and a maximum accepted level of deterioration and reliability indexes are dened. The corresponding partial safety factors are derived from a
full probabilistic approach. The performance of the proposed method is compared with that proposed on
the b Model Code for Service Life Design.
2012 Elsevier Ltd. All rights reserved.

1. Introduction
The service life design of reinforced concrete structures is an
important subject that has led to extensive research worldwide.
That research has given us the performance-based design methods
in use today [16,26,14]. These methods are founded on prediction
models and reliability concepts and, unlike the deemed-to-satisfy
approach, allow lifecycle cost analyses [31,37]. To encourage the
widespread use of the performance-based design its methods must
be as simple (user-friendly) and accurate as possible.
Reinforcing steel corrosion is the most common deterioration
mechanism in reinforced concrete structures [13,19,29]. Steel corrodes when in contact with humidity and oxygen but when it is inside concrete an oxide lm builds up that makes steel passive and
protects it against corrosion. This lm is stable in an alkaline envi Corresponding author at: Department of Civil Engineering and Architecture,
Section of Construction, IST, Technical University of Lisbon, Av. Rovisco Pais, 1049001 Lisbon, Portugal. Tel.: +351 21 8419709; fax: +351 21 8497650.
E-mail addresses: rui.neves@estbarreiro.ips.pt (R. Neves), fbranco@civil.ist.utl.pt
(F.A. Branco), jb@civil.ist.utl.pt (J. de Brito).
0950-0618/$ - see front matter 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.conbuildmat.2012.06.028

ronment such as young concrete. But it can be destroyed if the pH


falls or if the chloride content reaches a critical threshold. The pH
can decrease because of concrete carbonation, and so CO2 and chloride action are the most relevant processes in this context.
Although chloride-induced corrosion is considered to be worse
than carbonation-induced corrosion, Parrott [35] apud Jones et al.
[22] states that 2/3 of all structural concrete is exposed to environmental conditions that favour carbonation-induced corrosion. In
Taipei, signicant carbonation-induced corrosion of existing concrete bridges has been observed within 1015 years of construction [24].
The service life of a structural component is the period after
construction when all the properties exceed the minimum acceptable values when routinely maintained [1]. Predicting service life
requires a model that simulates the behaviour of the structure under environmental actions. The model must take into account the
relevant physical and chemical mechanisms and its performance
when simulating real behaviour must be assessed [14].
Tuutti [40] presented a model where the service life for concrete structures with regard to reinforcement corrosion is broken
down into an initial and a propagation stage. This division is

586

R. Neves et al. / Construction and Building Materials 36 (2012) 585591

needed because the primary ruling factors in these sub-processes


differ.
For the specic deterioration mechanism of carbonation-induced corrosion, the initial stage is the period from the beginning
of a structures life until the carbonation depth is the same as that
of the reinforcement cover, when corrosion is considered possible.
The propagation stage is the period from the end of the initial
stage until the reinforcement reaches an unacceptable level of corrosion degradation.
2. Limit states and reliability

Deterioration

Dening a level of unacceptable corrosion degradation, i.e. a


limit state, is not a simple matter. If the maximum acceptable deterioration level is associated with a service requirement, a serviceability limit state (SLS) is dened, while if the maximum
acceptable deterioration level is associated with collapse or similar
forms of structural failure, an ultimate limit state (ULS) is dened.
Criteria such as corrosion onset, cracking and spalling dene SLSs
[18,25], while spalling with apparent risk from falling pieces, loss of
bond between concrete and reinforcement and loss of cross-section
leading to fracture of structural component dene ULSs [23].
The criterion for service life design, especially for prestressed
structures or those affected by chloride ion penetration, is often
corrosion onset [2,5,15,36], which can be designated as a depassivation limit state. The advantages of this option are: ease of prediction
and less expensive pro-active protection measures. In addition, as
the rate of steel corrosion in chloride-induced corrosion is fast,
the impact on the initial costs of disregarding the propagation stage
will be small. But if the propagation stage of carbonation-induced
corrosion in dry exposure conditions is disregarded, this will
signicantly raise production costs.
In a scenario of cracking limit state, for carbonation-induced
corrosion, the ratio of initial and propagation stages in service life
is variable since, contrary to what happens with chloride-induced
corrosion, the conditions most favourable to depassivation are generally not ideal for the development of corrosion [34,38]. Therefore,
it is possible that identical elements, when exposed to different
environments, have the same lifetime but with different contributions from the initial and propagation stages, as illustrated in Fig. 1.
Thus, for a certain required service life, adopting a depassivation limit state instead of a cracking limit state implies extending
the initial stage. The relevance of this extension may range from
negligible (RHc > 85%) to overriding (45% < RHc < 65%). The most
direct ways to achieve such an extension are to increase rebar cover thickness and to improve concretes resistance to carbonation.
Any of these actions will raise costs, either directly by increasing

the cost of concrete, or indirectly, by reducing the reinforcements


exural strength efciency.
The verication of a certain limit state is performed using a
model which numerically expresses that limit state through a
function known as the reliability function, commonly expressed as:

Z RS

where R is a function that quanties a response; S is a function that


quanties the effect of an action.
According to Eq. (1), Z 6 0 represents failure [21]. Functions R
and S have input parameters that are stochastic variables. Moreover, the functions (models) themselves exhibit uncertainty.
Hence, a probabilistic approach to service life design is recommended [8,25,23].
In probability-based design, the probability of reaching a given
limit state is computed. That probability is designated as probability of failure, pf, and may be expressed as the probability of response falling below the action:

pf pR 6 S

Based on probability of failure, a reliability index, b, may be


dened:

b U1 pf

where U1(.) is the inverse standardised normal distribution


function.
In general, the R and S functions, and therefore also the Z function, are time-dependant. Their expected values, and even scatter,
may vary with time. Therefore, the probability of failure should not
be determined as an absolute value, but relative to a certain period
of time, usually the service life. Two approaches are possible: limit
state format and service life format [27]. Although they have different formats, they use the same basic theoretical concepts and deliver exactly the same results [8].
In the limit state format, the probability of failure is determined
by considering the relevant time interval: intended service life (Eq.
(4)).

pf;T pZ 6 0

where pf,T is the probability of failure during the period 0  T; T is


the target service life.
In the service life format, a lifetime function must be dened
whose result is a given service life, to be compared with the target
service life:

pf;T pL 6 T

where L is a lifetime function.

Cracking

45%<RHc<65%
65%<RHc<85%
RHc>85%

Time
Service life

Fig. 1. Concrete relative humiditys (RHc) inuence on carbonation-induced corrosion.

587

R. Neves et al. / Construction and Building Materials 36 (2012) 585591

Fig. 2 shows the differences between these approaches for a


reinforcement depassivation process, and highlights the use of
the same information. In Fig. 2, R stands for reinforcement cover,
S stands for chloride threshold or carbonation depth, fL(t) stands
for probability density function of service life and T stands for target service life.
According to Maes et al. [27], the service life format approach is
of particular interest when the majority of uncertainties lie on the
deterioration model, when the deterioration model has several
stages, e.g. initiation and propagation stages, or when the element
is subjected to more than one deterioration mechanism.
The numerical quantication of reliability is important as it enables target values to be dened according to the causes and processes evolved in the attainment of a limit state, the corresponding
consequences, and the cost of the construction works [10]. Furthermore, the possibilities of detecting damage and taking corrective
measures in the event of failure must be also considered when
dening target reliabilities [23].
The consequences associated with the attainment of a limit
state depend on the maximum level of deterioration accepted.
Nevertheless, attaining a limit state always entails economic costs,
which is why setting a reliability index must be supported by an
economic analysis aimed at cost optimisation. In this analysis,
the production cost needed to guarantee extra protection and the
cost of the consequences must be balanced to pursue the optimum
reliability index from an economic point of view.
This approach is of particular interest for limit states related to
reinforcement corrosion. Actually, the cost of repairs (consequences)
increases as higher levels of deterioration, i.e. depassivation, cracking
and spalling, are allowed. Higher reliability indexes will therefore be
required. Thus, if a limit state of spalling is considered, which increases the service life when compared to a depassivation limit state,
the required reliability index will also be higher than for depassivation limit state. Consequently, the requirements for reinforcement
cover and concrete durability-related properties will be similar if
either a depassivation or spalling limit state is considered.
3. Carbonation rate model

p
xK t

where x is the carbonation depth (mm); K the carbonation coefcient (mm/year1/2); and t is the exposure time to CO2 (year).
In Eq. (6), K is a parameter that takes into account all factors
affecting carbonation. Separating those factors into environmental
and concrete intrinsic factors, K may be dened as:

ka
ke

where ka is the parameter related to concrete intrinsic factors; and


ke is parameter related to environmental factors.
Under the same environmental conditions, including CO2 concentration, the coefcient K is a relative indicator of concrete carbonation resistance and the most straightforward and reliable
parameter for characterising the inuence of concrete intrinsic factors on carbonation. To evaluate carbonation resistance with
timely delivery of results, accelerated tests have to be used in
which concrete is exposed to mediums with relatively high CO2
concentration.
Neves et al. [33] investigated the relationship between carbonation resistance in accelerated and natural conditions, using cores
drilled from 96 spots in real structures, using different structural
elements, exposed to environmental class XC3 or XC4 [12] whose
age ranged from 4 to 32 years. Based on their conclusions and using
Eq. (6), which has exhibited an ability to simulate carbonation
depth evolution over time in natural conditions [32,7,28,17,30,
39], the following model is proposed:

R, S

Adequate modelling of carbonation is a delicate issue since several factors must be taken into account and users expect research
to deliver simple and user-friendly models. Factors affecting car-

bonation can be organised into two major groups: environmental


factors and factors intrinsic to concrete. The environmental factors
include: the occurrence of drywet cycles [4], relative humidity
[41] and CO2 concentration [41]. Intrinsic factors are mainly related to concrete porous structure and the amount of matter that
can be transformed into carbonates.
The main transport mechanism in concrete carbonation is CO2
diffusion through its porous system, which can be described using
Ficks rst law. Assuming that the CO2 concentration, the amount
of CO2 required to carbonate a unit volume of concrete and the diffusion coefcient for CO2 through carbonated concrete are all constant, and also that a complete reaction takes place before
carbonation proceeds, and integrating the diffusion equation, the
following formula is obtained:

Limit state format


R

S(t)

pf,T

Time

fL(t)

Mean service life

Service life format


Fig. 2. Reliability analyses by two different approaches.

588

R. Neves et al. / Construction and Building Materials 36 (2012) 585591

4. Design criteria

Fig. 3. Measured vs. predicted carbonation depths for environmental class XC3.

ka p
t
ke

The maximum accepted level of deterioration is corrosion onset,


thus it corresponds to an SLS: depassivation limit state. Corrosion
models are not as developed and reliable as initiation stage models; hence it is considered preferable to limit the deterioration level
to the end of the initiation stage. However, as stated, disregarding
the propagation stage in certain conditions of carbonation-induced
corrosion may lead to a signicant increase of the initial cost. But
in these conditions the propagation stage can be seen as an additional safety margin, which can be balanced against allowing a larger failure probability [23].
A semi-probabilistic approach (safety factors) is commonly used
in structural design. The semi-probabilistic approach provides a
relatively simple reliability analysis of sufcient accuracy. Unlike
the full-probabilistic approach, the intended values are achieved directly, i.e. without iterations.
According to EN 1990 [10], a reliability index b = 1.50 must be
considered for SLSs. This reliability index is related to a failure
probability pf = 0.066. The failure probability is the probability that
resistance R may fall below the action S. In a semi-probabilistic
procedure the restriction of failure probability can be expressed as:

Rc
where x is the carbonation depth (mm); ka the resistance to accelerated carbonation (mm/year1/2); ke the environmental parameter,
being 9.9 for environmental class XC3 and 15.0 for environmental
class XC4; and t is the exposure time to CO2 (year).
Resistance to carbonation in accelerated conditions (ka) is quantied by means of K in Eq. (1), i.e. an accelerated carbonation coefcient, which has proven to be suitable for simulating carbonation
in accelerated conditions [20,32,3,6], using the following test
conditions:
Preconditioning: 3 weeks at 20 3 C and 65 5% RH.
Testing age: from 5 weeks.
Accelerated carbonation conditions: 20 3 C, 65 5% RH and
5 0.1% CO2 concentration.
The performance of the model is illustrated in Figs. 3 and 4, for
environmental classes XC3 and XC4, respectively, using the results
from Neves et al. [33]. According to these authors, when using
different CO2 concentrations in accelerated testing for carbonation
resistance, a correction factor, ka, given by the square root of the
accelerated testing concentrations ratio may be used if CO2 concentrations are still near 5%.

Fig. 4. Measured vs. predicted carbonation depths for environmental class XC4.

cR

 cS Sc > 0

where Rc is the resistance based on characteristic values; cR, cS the


safety factors; and Sc is the action based on characteristic values.
Within the aim of a carbonation-induced corrosions reliability
analysis, R is the thickness of reinforcement cover and S is carbonation depth. The safety factor related to reinforcement cover is often replaced by a safety margin. The reinforcement cover design
value is thus given by:

cd cnom  Dc

10

where cd is the reinforcement cover design value; cnom the nominal


(specied) reinforcement cover; and Dc is the reinforcement cover
safety margin, usually 10 mm [16,9].
To quantify the safety factors related to carbonation depth, lifetime simulations using the Monte Carlo technique were performed.
In the process, the reinforcement cover was considered a deterministic variable (cd) and parameters ka and ke, from the carbonation model (Eq. (8)) were assumed to be stochastic variables.
For environmental class XC4, ke was considered as a normal distributed variable with a mean value of 15 and a standard deviation
of 0.88 [33]. The resistance to accelerated carbonation (ka) was also
considered as a normal distributed variable, with a coefcient of
variation of 13% [32].
The analysis of simulations results led to a safety factor cS = 1.25
for a reliability index b = 1.50.
A different analysis was performed for environmental class XC3
since it was intended to express the safety margin (propagation
stage) in terms of a reliability index. The propagation stage was
considered as the period between depassivation and crack formation due to corrosion. A lowest estimation of 45 years was considered for this period [26].
The design service life for the depassivation limit state was assessed by performing simulations using a Monte Carlo technique,
where ke was considered as a normal distributed variable with a
mean value of 9.9 and a standard deviation of 0.61 [33]. The resistance to accelerated carbonation (ka) was again considered as a
normal distributed variable, with a coefcient of variation of 13%.
Adding the propagation period to the design service life for the
depassivation limit state, considering at the worst a specied service life of 200 years and a failure criterion of concrete cracking
due to corrosion, a safety factor cS = 1.00 is computed. In terms

R. Neves et al. / Construction and Building Materials 36 (2012) 585591

of the depassivation limit state, this value for the safety factor
means a failure probability of 0.50. This failure probability corresponds to a reliability index b = 0, which is within the range of
reliability indexes for XC3 class quoted by Lay and Schiessl [23].

with the model codes partial safety factor approach. The corresponding limit state equation, in its most simplied form, with
partial safety factors already included, is:
cd P

0:5psr ToW0:446
q
0:0767
1:13  3:05  1011  RH5c 2:5 kc;d 3:08  103 R1
ACC 0:52t SL
tSL

5. Proposed method

12

As a consequence of the discussion and analysis in the previous


sections a simple method for service life design with respect to carbonation induced corrosion is proposed.
First, a nominal reinforcement cover, according to EN 1992 [11],
must be dened. Next, the resistance to accelerated carbonation, to
be assessed in the test conditions set out in Section 2, must be foreseen using the following equation:

ka

c d ke
p

cS tSL

589

11

where ka is the resistance to accelerated carbonation (mm/year1/2);


cd the reinforcement cover design value (mm), according to Eq.
(10); ke is the environmental parameter, being 9.9 for environmental class XC3 and 15.0 for environmental class XC4; cS is the safety
factor, being 1.0 for environmental class XC3 and 1.25 for environmental class XC4; and tSL is the specied service life (year).
The required accelerated carbonation resistances obtained by
applying the method for nominal reinforcement cover ranging
from 15 mm to 40 mm and service life between 10 and 120 years
are presented in Figs. 5 and 6, for environmental classes XC3 and
XC4, respectively. In fact, ka represents the inverse of resistance
to accelerated carbonation as higher values of ka correspond to less
carbonation resistant concrete.
6. Comparative analysis
The proposed method is compared with a reference method
established in the Model Code for Service Life Design [16]. The
comparison is made in terms of required accelerated carbonation
resistance using four scenarios, which vary in: reinforcement cover, service life and environmental class (Table 1). Two indicative
service life times are considered: 50 years for building structures
and 100 years for bridges [10]. The environmental classes considered are those covered by the method proposed in this paper and
the reinforcement covers are selected according to EN 1992 table
4.4N [11], taking structural class S4.
The Model Code for Service Life Design considers the depassivation limit state and three different levels for its verication: a full
probabilistic approach, a partial safety factor approach and a
deemed-to-satisfy approach. The proposed method is compared

where cd is the reinforcement cover design value (mm), according


to Eq. (10); RHc the relative humidity of concrete (%); kc,d the execution transfer parameter; R1
ACC the inverse effective carbonation
resistance ((mm2/year)/(kg/m3)); tSL the specied service life
(year); psr the probability of driving rain; and ToW is the time of
wetness.
Solving Eq. (12), an upper bound for inverse effective carbonation resistance is found. The parameters used in Eq. (12) for the
various scenarios taken from the literature on the subject [23,26]
are presented in Table 2. The resulting inverse effective carbonation resistance is to be assessed by accelerated carbonation testing. The value of inverse effective carbonation resistance is
determined using Eq. (13).

R1
ACC

x2
Cs t

13

where x is the carbonation depth (mm); CS the CO2 concentration


(kg/m3); and t is the exposure time to CO2 (year).
For the purpose of comparing the requirements of both methods, R1
ACC values resulting from Eq. (13) are converted to ka, using
the following expression:

ka

q
2  C s  R1
ACC

14

where ka is the resistance to accelerated carbonation (mm/year1/2);


CS the CO2 concentration = 0.09 kg/m3; and R1
ACC is the inverse effective carbonation resistance ((mm2/year)/(kg/m3)).
The required accelerated carbonation resistances for the considered scenarios, taken from the Model Code for Service Life Design
and the new (previous section) methods, are presented in Table 3.
For environmental class XC3 (scenarios A and B) the reference
method leads to more restrictive accelerated carbonation resistances. For environmental class XC4 (scenarios C and D) similar
accelerated carbonation resistances are required from both
methods.
The difference between the required values for scenarios A and
B is mainly due to the adoption of different reliability indexes:
b = 1.3 in Model Code for Service Life Design and b = 1.0 in the proposed method. This distinction relies on taking into account the
margin of safety provided by the slow corrosion development for
XC3 exposure conditions.

Fig. 5. Required ka for environmental class XC3.

590

R. Neves et al. / Construction and Building Materials 36 (2012) 585591

Fig. 6. Required ka for environmental class XC4.

Table 1
Scenarios for the comparative analysis.
Scenario

Cover (cd) (mm)

Service life (tSL) (year)

Environmental class

A
B
C
D

25
25
30
30

50
100
50
100

XC3
XC3
XC4
XC4

Table 2
Parameters for Eq. (12).
Scenario

cd (mm)

RHc (%)

kc,d

tSL (year)

psr

ToW

A
B
C
D

25
25
30
30

70
70
80
80

1
1
1
1

50
100
50
100

0
0
0.05
0.05

0.21
0.21

A semi-probabilistic approach that uses a safety factor in association with a reliability index was adopted.
The impact of not considering an acceptable level of deterioration where corrosion is allowed was analysed. For situations where
the relative weight of the propagation stage was high (environmental class XC3), the period was considered as an additional
safety margin, as reected in the allowance of a lower reliability
index.
A practical method was proposed, whose main features are:

Use of only one analytical model.


Analytical model calibrated with long term results.
Analytical model requiring just two parameters.
Use of a single safety factor.

Moreover, a comparative analysis with a reference method has


shown that the new proposed method is suitable to establish carbonation performance requirements.
Acknowledgments

Table 3
Required accelerated carbonation resistances (ka in mm/year1/2).
Scenario

Model code

New method

A
B
C
D

24
16
47
35

35
24
50
36

7. Conclusions
One of the processes responsible for the reinforcement corrosion problems in reinforced concrete structures is the natural carbonation of concrete, which depends on both the materials
characteristics and the surrounding environment.
Although a structures durability can generally be assured by
means of a deemed-to-satisfy approach, performance-based design
is preferable as it allows a more rational use of the available resources and a lifecycle cost analysis.
The aim of this study was to provide a simple performancebased design method to prevent carbonation-induced corrosion.
The issue of acceptable degradation level was discussed, and a
depassivation limit state which requires only one prediction model
was adopted.
A model to predict carbonation depth based on Ficks rst law
and that was calibrated with long term natural exposure conditions carbonation assessments was considered.

The authors acknowledge Brisa Auto-Estradas de Portugal, SA


for the possibility of using structures from its concessions for the
development of this work and are grateful for the support of the
ICIST Research Institute, ESTBarreiro, Polytechnic Institute of Setbal, IST, Technical University of Lisbon and FCT (Foundation for
Science and Technology).
References
[1] ACI, ACI 365. Service-life prediction, Michigan, USA; 2000.
[2] Aguiar J. Evaluation of reinforced concrete durability tests from durable
structures. Master dissertation in civil construction. Minas Gerais Federal
University; 2006 [in Portuguese].
[3] Audenaert K, Schutter G. Inuence of moisture on the carbonation of self
compacting concrete. In: Sixth CANMET/ACI international conference,
durability of concrete. Supplementary papers, Thessaloniki; 2003. p. 45165.
[4] Bakker RFM. Initiation period. In: Schiessl P, editor. Corrosion of steel in
concrete. RILEM; 1988.
[5] Baroghel-Bouny V. Durability indicators: relevant tools for performance-based
evaluation and multi-level prediction of RC durability. International RILEM
workshop on performance based evaluation and indicators for concrete
durability. Madrid; 2006. p. 330.
[6] Baroghel-Bouny V, Arnaud S, Gawsewitch J, Belin P, Olivier G, Bissonnette B,
et al. Aging of concrete in natural environments: an experiment for the XXI
century. IV Results on cores extracted from eld-exposed test specimens of
various sites as part of the rst measurement sequence. Bulletin des
Laboratoires des Ponts et Chausses 2004;249:49100.
[7] Bourguignon K. Compressive strength inuence on the carbonation of
concretes with different slag amounts. Masters dissertation in civil
engineering. Esprito Santo Federal University, Vitria, Brazil; 2004 [in
Portuguese].

R. Neves et al. / Construction and Building Materials 36 (2012) 585591


[8] CEB New approach to durability design, Bulletin dInformation N238, CEB,
Lausanne, Switzerland; 1997.
[9] CEN EN 13670:2009 Execution of concrete structures. Brussels, Belgium;
2009.
[10] CEN EN 1990:2002 Eurocode 0. Basis of structural design. Brussels, Belgium;
2002.
[11] CEN EN 1992-1-1:2004 Design of concrete structures. Part 11. General rules
and rules for buildings Eurocode 2, Brussels, Belgium; 2004.
[12] CEN EN 206-1 Concrete Part 1: Specication, performance, production and
conformity. Brussels, Belgium; 2000.
[13] Clark L. Is cover important? Concrete 2000(May):40.
[14] Ferreira RM. Probability-based durability analysis of concrete structures in
marine environments. PhD thesis in civil engineering. University of Minho;
2004.
[15] Ferreira RM, Jalali S. Probability-based approach to service life analysis of
concrete harbour structures. RILEM Workshop on Life Prediction and Aging
Management of Concrete Structures, Paris; 2003. p. 31928.
[16] b. Model Code for service life design, Lausanne, Switzerland; 2006.
[17] Galan I, Andrade C. Comparison of carbonation models. Third international
RILEM PhD student workshop on modelling the durability of reinforced
concrete. Guimares; 2009. p. 419.
[18] Gehlen C. Probabilistische lebensdauerberechnung von stahlbetonbauwerken

Zuverlssigkeitsbetrachtungen
zur
wirksamen
vermeidung
von
bewehrungskorrosion (in German), German Committee for Structural
Concrete (DafStb). Berlin: Germany; 2000.
[19] Gjorv O. Durability design of concrete structures in severe
environments. Abingdon, UK: Taylor and Francis; 2009.
[20] Hyvert N, Sellier A, Duprat F, Rougeau P, Francisco P. Dependency of CSH
carbonation rate on CO2 pressure to explain transition from accelerated tests
to natural carbonation. Cem Concr Res 2010;40(11):15829.
[21] JCSS Probabilistic assessment of existing structures. RILEM, Bagneux, France;
2001.
[22] Jones M, Dhir R, Newlands M, Abbas A. A study of the CEN test method for
measurement of the carbonation depth of hardened concrete. Mater Struct
2000;33(2):13542.
[23] Lay S, Schiessl P. LIFECON Deliverable D 3.2 Service life models, Technical
University of Munich, Munich, Germany; 2003.
[24] Liang MT, Zhao GF, Liao YS, Liang CH. Linearly unbiased estimation combined
with JC method for the carbonation contamination of some concrete bridges in
Taipei. J Mar Sci Technol 2000;8(2):7889.
[25] Lindvall A. A probabilistic, performance based service life design of concrete
structures

environmental
actions
and
responses.
Goteborg,
Sweden: Chalmers University; 2002.
[26] LNEC E-465 Concrete. Prescriptive methodology to estimate concrete
properties to achieve the design service life under environment conditions

[27]

[28]

[29]
[30]
[31]
[32]

[33]

[34]
[35]
[36]

[37]

[38]

[39]

[40]
[41]

591

XC or XS, National Laboratory of Civil Engineering (LNEC), Lisbon, Portugal;


2007 [in Portuguese].
Maes M, Troive S, Fritzsons K. Contrasting different reliability-based design
formats for the durability of concrete structures. International RILEM-JCI
Seminar on Concrete Durability and Service Life Planning: Curing, Crack
Control, Performance in Harsh Environments, Dead Sea; 2006.
Meira G, Padaratz I, Jnior J. Concrete natural carbonation: results from four
years of monitoring, XI Encontro Nacional de Tecnologia no Ambiente
Construdo, Florianpolis; 2006. p. 302937 [in Portuguese].
Molin D. Cracks in reinforced concrete structures. Master dissertation in civil
engineering. Rio Grande do Sul Federal University; 1988.
Monteiro I, Branco F, de Brito J, Neves R. Statistical analysis of the carbonation
coefcient in open air concrete structures. Constr Build Mater 2012;29:2639.
Naus D, Johnston M. International RILEM workshop on life prediction and
aging management of concrete structures. Mater Struct 2001;34(8):45866.
Neves R. Concrete air permeability and carbonation in structures. PhD thesis in
civil engineering, Instituto Superior Tcnico, Technical University of Lisbon,
Portugal; 2012 [in Portuguese].
Neves R, Branco F, de Brito J. Field assessment of the relationship between
natural and accelerated concrete carbonation resistance. Cem Concr Compos,
submitted for publication.
Page CL. Corrosion and protection of reinforcing steel in concrete. Durability of
concrete and cement composites. UK: Woodhead Publishing; 2007. p. 13686.
Parrott L. Some effects of cement and curing upon carbonation and
reinforcement corrosion in concrete. Mater Struct 1996;29(3):16473.
Polder R, van der Wegen G, Boutz M. Performance based guideline for service
life design of concrete for civil engineering structures a proposal discussed in
the Netherlands. International RILEM workshop on performance based
evaluation and indicators for concrete durability. Madrid; 2006. p. 319.
Rajabipour F, Weiss J, Abraham D. In situ electrical conductivity measurements
to assess moisture and ionic transport in concrete. In: International RILEM
symposium on advances in concrete through science and engineering,
Evanston; 2004.
Saetta A, Vitaliani R. Experimental investigation and numerical modeling of
carbonation process in reinforced concrete structures. Part I: Theoretical
formulation. Cem Concr Res 2004;34(4):5719.
Smolczyk H. Discussions to M. Hamadas paper: Neutralization (carbonation)
of concrete and corrosion of reinforcing steel. In: Fifth international
symposium on cement chemistry. Tokyo; 1968. p. 36983.
Tuutti K. Service life of structures with regard to corrosion of embedded steel.
ACI Special Publication 1980;65:22336.
Verbeck G. Carbonation of hydrated Portland cement, PCA Bulletin 87,
Portland Cement Association, Illinois, USA; 1958. p. 1736.

You might also like