Concrete Repair, Rehabilitation and Retrofitting II – Alexander et al (eds)

© 2009 Taylor & Francis Group, London, ISBN 978-0-415-46850-3

Reflections on future needs in concrete durability research

and development

Y. Ballim
School of Civil & Environmental Engineering, University of the Witwatersrand, Johannesburg, South Africa

M.G. Alexander, H.-D. Beushausen & P. Moyo

Department of Civil Engineering, University of Cape Town, Cape Town, South Africa

ABSTRACT: There is no doubt that, over the past two decades, we have made enormous advances in the
understanding and practice of concrete durability. Spurred by the often experienced early deterioration of rein-
forced concrete structures, with high capital investment for repair and rehabilitation, conceptions of design for
durability have gained an increasingly higher level of importance in recent years. Durability design is beginning
to be considered of equal importance as design for safety and serviceability. Nevertheless, durability concerns
remain and early deterioration still attracts much expenditure. This paper is aimed at identifying important
developments made in the field of concrete durability during the past two decades. Based on current design
practice and current knowledge, future research and development needs are discussed, focusing on the influ-
ences of constituent materials, deterioration prevention methods, service life modelling of reinforced concrete
structures, and performance-based test methods.

1 INTRODUCTION far from settled and the need for continuing research
and development remains.
It is not difficult to recognise or to acknowledge the This paper therefore presents some reflections on
significant changes that have taken place in the field the future research needs in concrete durability. Much
of concrete durability over the last 25 years. This of this discussion relies on observations of the extent
period has seen an enormous growth in our under- to which researchers have addressed some of the
standing of the mechanisms and processes of deterio-
ration, the ways in which to protect concrete against
early deterioration and the effective utilisation of
concrete-making materials that will enhance durabil-
ity. Equally important has been the extent to which the
advocacy of durability—as a fundamental property of
concrete—has resulted in a conceptual change in the
minds of those involved in the design and construc-
tion of concrete structures. For practitioners, concepts
of durability are beginning to be woven into the eve-
ryday discourse of concrete technology.
In large measure, this change has been driven by
the efforts of researchers throughout the world, both in
developing our understanding of the theoretical issues
and in making the case for greater attention to dura-
bility in the cement and concrete sector in general. In
the best tradition of the path of knowledge develop-
ment, the debates in the field of concrete durability
have certainly been vigorous and many misconcep- Figure 1. Reinforcement corrosion remains the most
tions or flawed theories lie on the sides of this path. significant threats to the durability of reinforced concrete
However, in many areas, the issues and concerns are structures.

73
fundamental questions of concrete durability, influ- structures in relation to concrete mixture constituents.
enced the practice of concrete technology, or attended This will act as a touchstone for researchers to assess
to needs of technology transfer in this field. While the the suitability of proposed models against a historical
issues raised are intended to have general application, knowledgebase of recorded performance. There is a
the reader will find something of a South African bias. special concern here regarding the potential problems
with new materials such as high-performance con-
crete, for which long-term behaviour is unrecorded.
2 MATERIALS In these materials, the likely effects of large amounts
of unhydrated cement on durability (as well as long-
2.1 Influence of material properties term deformations) remain a concern.
on reinforcement corrosion
Significant advances have been made in our under- 2.2 Other deterioration mechanisms
standing of the influence of concrete mixture con- While deterioration due to reinforcement corrosion
stituents on the likelihood and rate of reinforcement remains the most significant global concern in this
corrosion. Much of this work has focused at under- area, other deterioration mechanisms are also in need of
standing the role of cement extenders and admixtures research attention. Some examples of such issues are:
in modifying the ionic and fluid transport properties
of the concretes that we use in chloride-laden envi- − Our understanding of the mechanisms and process
ronments. Greater understanding of the notion of the of sulphate attack remains hazy and there is much
‘threshold’ level of chloride content in relation to the confusion and sometimes contradiction on this
binder type has also helped to clarify much of the phenomenon (Neville, 2006).
complexity of empirical data. − In the case of alkali silica reaction, our knowledge
Our sense is that continued work is necessary in of the strategies for prevention is fairly advanced.
developing a more fundamental understanding of the However, more work is clearly needed in the repair
physico-chemical process related to the interactions and rehabilitation of concrete damaged by this
between the products of hydration and corrosion- mechanism of attack.
inducing species such as chlorides or carbon dioxide. − Soft water attack and other dissolution proc-
These studies should be focused on ingress, binding esses also requires further research, particularly
and flow of these materials in concrete. This is neces- to develop our understanding of the strategies for
sary in light of the rapidly changing nature of cemen- protecting concrete against such deterioration.
titious materials that are being used in concrete. Our
Much of the comments related to the materials
reliance on models developed from empirical data,
research needs for reinforcement corrosion apply
without fundamental understanding, limits our ability
here in equal measure.
to predict the long-term performance of new mate-
rials used in concrete. Furthermore, the increasing
use of combinations of supplementary cementitious 3 SERVICE LIFE MODELING
materials—so called ternary blends—presents par- OF REINFORCED CONCRETE
ticular challenges in this context. STRUCTURES
The past five years have also seen enormous
growth in the materials and technology of concrete
3.1 General
admixtures. However, there is little information on
the effects of these new materials on the general For reinforced concrete structures, the most impor-
durability and deterioration of concrete. In particular, tant durability considerations concern reinforce-
there is a serious lack of understanding of the effects ment corrosion. As a result, the modeling of the
of these materials on concrete exposed to the poten- ingress of aggressive agents such as chlorides and
tial of damage due to reinforcement corrosion. carbon dioxides has received considerable attention
In the areas of limiting the effects of corrosion of on recent years. However, many questions yet have
reinforcing steel, the following are some of the issues to be answered in determining reliable and practical
that will require attention: design procedures for reinforced concrete members
subjected to chloride ingress or carbonation.
− Corrosion inhibitors: future research is needed to
Design approaches for durability can be divided
establish long-term performance and applicability
into prescriptive concepts, also termed deemed to sat-
− Service life of protective coatings
isfy concepts, and performance-based concepts. Pre-
− The development of corrosion-resistant reinforcement.
scriptive concepts are based on material specification
There is certainly a need for an international shared from given parameters such as exposure classes and
database on long-term performance of concrete life span of the structure. Following this approach,

74
durability specifications in most existing codes and The full probabilistic approach, which is intended
standards are based primarily on establishing con- to be used for exceptional structures only, should be
straints to the mix proportions of the concrete as a based on probabilistic models that are sufficiently
function of the severity of the exposure. validated to give realistic and representative results
Durability specifications in the South African stand- of deterioration mechanisms and material resistance.
ards (SANS 2005) and the new European standards The basis of this approach is formed by appropriate
(BS EN 2004), for example, follow the prescriptive test methods and statistical evaluation models.
concept and are of the ‘recipe’ type, setting limits on In the partial safety factor approach, the probabil-
w/c ratios, cement contents, and compressive strength istic nature of the problem (scatter of material resist-
for different exposure classes. The design for dura- ance and load) is considered through partial safety
bility includes the correct choice of exposure class factors. It is based on the same models as for the full
and compliance with material requirements, concrete probabilistic approach and intends to present a practi-
cover specifications, and curing procedures. cal, yet statistically reliable design tool.
However, durability is a concept that incorporates The deemed to satisfy approach is comparable
material properties, processing technology and envi- to the durability specifications given in most cur-
ronmental exposure conditions and, as such, it cannot rent codes and standards, i.e. specifications based
easily be assessed through intrinsic material properties. on a selection of certain design values (dimension-
The prescriptive approach ignores, to a large extent, ing, material and product selection, execution proce-
the different performance of the different cement types dures) based on environmental classes. The difference
and of the mineral components added to the cements between the deemed to satisfy approach envisaged by
or to the concrete itself, as well as the influences of on- the fib Model Code and traditional service life design
site practice during the construction process. rules is that the latter are commonly not based on
Performance concepts, on the other hand, are based physically and chemically correct models but largely
on quantitative predictions for durability from expo- on practical experience, whereas the fib method is
sure conditions and measured material parameters. The intended to be calibrated against the full probabilistic
resistance of the structure, measured through durabil- approach.
ity parameters of the actual concrete used, is compared The fourth level of service life design (avoidance
against the environmental load, taking the influences of deterioration) is based on the use of non-reactive
of time into account. On this basis, the probability of materials such as stainless steel, or protection sys-
damage occurring to the structure during its lifetime is tems, such as coatings.
calculated using appropriate deterioration models. In the presentation of the above design options, fib
Different levels of sophistication may be applied to has taken a major step towards designing concrete
performance-based design for durability, including the structures for durability. Fundamental principles
use of durability indexes, the application of analytical and design options for concrete durability have been
deterioration models, and full probabilistic methods. At clearly laid out. For the successful application of the
the moment, various approaches are being developed various design options, however, further work is nec-
worldwide, however yet with very limited application in essary in the following fields:
real projects. Of course, the added challenge of assess-
ing such models against long-term structural perform- − Defining rational limit state criteria,
ance remains as an important future research need. − Testing actual material performance against rel-
evant material deterioration models,
− Calibrating service life models against uncertain-
3.2 Current international developments ties in material properties and environmental influ-
The recently published fib Model Code for Service ences (probabilistic methods),
Life Design (fib, 2006) proposes a design approach − Identifying suitable test methods for the characteri-
to avoid deterioration caused by environmental action zation of relevant concrete properties. For existing
comparable to load design. Based on quantifiable test methods, clear guidance needs to be provided
models for the load side (environmental actions) and on how to interpret test results and on how to incor-
the resistance side (resistance of the concrete against porate them into service life models,
the considered environmental action), the following − Identifying efficiency and durability of protection
design options are presented: systems, such as coatings.

− Option 1: Full probabilistic approach Many international research efforts are and have
− Option 2: Semi probabilistic approach (partial been underway in resolving the above listed issues.
safety factor design) However, most countries struggle with the implemen-
− Option 3: Deemed to satisfy rules tation of durability design guidelines, due to the many
− Option 4: Avoidance of deterioration. questions yet to be answered.

75
 In contrast, in South Africa, durability design concretes using these tests, studying in-situ concrete
guidelines have been implemented in the past dec- performance, and applying the results to practical
ade and are nowadays frequently applied in concrete construction.
construction. The approach has been developed to The approach links durability index parameters,
a point where it can be used with some confidence; service life prediction models, and performance
however the development of the approach is ongoing specifications. Concrete quality is characterized in-
as remaining uncertainties still need to be addressed. situ and/or on laboratory specimens by use of dura-
Based on the South African durability design bility index tests, covering oxygen permeation, water
approach, some of the most common shortcomings absorption, and chloride conduction (Alexander et al
associated with durability design are discussed in the 2001, Beushausen et al 2003, Alexander and Stanish
following. 2005). The service life models in turn are based on
the relevant DI parameter, depending on whether the
design accounts for carbonation-induced or chloride-
3.3 The South African approach for durability
induced corrosion. Designers and constructors can
design
use the approach to optimize the balance between
The philosophy of the South African durability design required concrete quality and cover thickness for a
approach involves the understanding that durability given environment and binder system.
will be improved only when unambiguous measure- A framework for the development and application of
ments of appropriate cover concrete properties can performance-based specification methods for concrete
be made. Such measurements must reflect the in situ durability is illustrated schematically in Figure 2.
properties of concrete, influenced by the dual aspects In view of the various options for service life design
of material potential and construction quality. Key presented by fib (2006), the South African Durabil-
stages in formulating this approach were develop- ity Index approach principally adopted the strategy
ing suitable test methods, characterizing a range of of providing deemed to satisfy rules, which limit

MATERIAL INDEXING
Characterization of concrete (surface layer)
using easily measured physical properties,
such as permeability, sorptivity and possibly QUALITY CONTROL
strength.

Correlations
E
N
DIRECT DURABILITY TESTING V
I
Suite of accelerated tests Long-term tests (lab R
(lab): Modelling of envi- or site-based). As O
Correlations
ronment and mechanisms close to ‘real’ condi- N
of deterioration tions as possible M
E
N
T
PREDICTION
STRUCTURAL PERFORMANCE
Correlations Evaluation of structural performance; Con-
sequences of deterioration; Management of
economic strategies

Micro- and macro- environ-

ment. Presence of deteriorat-
ing agents; measurement of
FUNDAMENTAL MECHANISTIC STUDIES aggressiveness of environ-
ment

Figure 2. Framework for carrying out durability studies in South Africa.

76
durability index values and cover depths for given regions in South Africa. This project is described in
environmental classes and selected binder types. more detail in another paper presented at ICCRRR
Limiting durability index values are based on serv- 2008 (Heiyantuduwa, 2008).
ice life models, which in turn relate to partial fac- The design against carbonation using oxygen per-
tor design and partly to full probabilistic modelling meability indexes is based on empirical relationships
of relevant parameters. The South African approach between 28-day OPI values and carbonation depth
therefore aims at combining Options 1, 2, and 3, as measurements on actual structures and laboratory-
presented by fib, to provide a practical tool for dura- cured samples (under accelerated carbonating condi-
bility design and quality control. tions). Carbonation depth measurements, particularly
The durability index approach is currently being on real structures, show large variations, making sta-
applied in a number of large scale construction tistical evaluations of test results difficult. It is there-
projects in South Africa. This is considered a large fore important to collect more data to calibrate the
step towards improvement of concrete quality and service life models used for the prediction of carbon-
structural durability. However, much work remains ation. This needs to be done taking various climatic
to be done, in particular generating correlations conditions into account and researching fundamental
between indexes and actual structural performance. influences on carbonation of concrete.
Future research needs in the field of durability index
design and service life modelling are discussed in the 3.4.2 Definition of environmental classes
following. The environmental classes used in durability design
in South Africa are related to the EN 206 classes as
modified for South African conditions (Table 1). As
3.4 Future needs in durability design discussed in the previous section, it needs to be estab-
and service life modelling lished whether the current environmental classes are
3.4.1 Calibration of service life models and sufficient in describing potential deterioration mech-
test methods against field performance anisms, or whether a more refined approach needs to
of concrete structures be developed that classifies environmental exposure
The design against chloride ingress, using chloride based on the prevailing climate conditions.
conductivity indexes, is based on service life models
that were developed at the University of Cape Town
(Mackechnie, 2001). The relationship between con-
ductivity index values and the potential field perform- Table 1. Environmental Classes adopted in South Africa
ance of reinforced concrete structures was established (Natural environments only) (after EN206).
from 2 different techniques:
Carbonation-Induced Corrosion
− Correlation between 28-day conductivity index
values and chloride ingress in structures in the Designation Description
Western Cape Province.
− Laboratory-based experimental correlations bet- XC1 Permanently dry or permanently wet
ween 28-day conductivity index values and chlo- XC2 Wet, rarely dry
ride diffusion coefficients. XC3 Moderate humidity (60–80%)
(Ext. concrete sheltered from rain)
Further work is necessary to test 28-day conduc- XC4 Cyclic wet and dry
tivity index values against chloride ingress in vari-
ous marine environments in South Africa, taking into Corrosion Induced by Chlorides from Seawater
account that chloride ingress is dependent on envi-
Designation Description
ronmental conditions such as water and air tempera-
ture and relative humidity. In a current study at the XS1 Exposed to airborne salt but not in
University of Cape Town, chloride ingress into vari- direct contact with seawater
ous different types of concrete is being investigated, XS2a* Permanently submerged
based on site exposure in the Cape Town and Dur- XS2b* XS2a + exposed to abrasion
ban harbours, which, from a temperature and rela- XS3a* Tidal, splash and spray zones
tive humidity point of view, represent very different Buried elements in desert areas exposed
environmental conditions. Measurements taken on to salt spray
site-exposed samples are correlated to laboratory- XS3b* XS3a + exposed to abrasion
based measurements of chloride conductivity index
values and diffusion coefficients, in order to refine * These sub clauses have been added for South African
existing service life models for application in various coastal conditions.

77
3.4.3 Development of probabilistic models Level of deterioration
The natural variability in the concrete material makes
it inevitable to use probability theory in formally
Initiation period Propagation period 4
including the uncertainties in the service life predic-
tion model parameters.
A framework for the application of probabilistic
models in durability design of reinforced concrete
(RC) structures in South African marine conditions 3
has been developed in a research project at the Uni-
versity of Cape Town. This involved incorporating the 2
steady state diffusion coefficient derived theoretically
1
from the chloride conductivity test results in a proba-
bilistic model as discussed in Muigai et al (2008). Time of exposure [years]
Statistical information for each parameter in the
model was applied in providing improved estimates of
the predicted service life. The research study also dem- 1 Depassivation of the reinforcement Initiation period
onstrated the use of the probabilistic model in specify- 2 Formation of cracks
ing limiting values for chloride conductivity based on 3 Spalling of concrete cover
initiation limit state target probability values given in 4 Propagation
fib Model Code for Service Life Design (fib, 2006). Failure of the structure through
bond failure or reduction in cross period
However, additional data sets still need to be
section of the load bearing
acquired for each of the quantified model parameters reinforcement
to improve on the accuracy of the model. Further
research in this field also requires the application of Figure 3. Deterioration process of reinforcement corro-
the oxygen permeability test in the carbonation serv- sion: 2-phase model for service life (Tuuti 1982, fib 2006).
ice life prediction model based on the same probabi-
listic framework.

3.4.4 Definition of limit state criteria 3.4.5 Investigating the effects of cracking
A number of service life models for reinforced con- on concrete durability
crete structures exist. Many of these adopt the two- Most service life prediction models cover the ingress
stage service life model first proposed by Tuutti of harmful substances into uncracked concrete only.
(1982), in which the deterioration is split into two As a result, transport mechanisms for chlorides and
distinct phases, namely the initiation period and the carbon dioxide are primarily assumed to relate to
propagation period, as shown in Figure 3. established diffusion models. However, cracks do
Most service life models assume that the end of frequently occur in concrete structures, especially
the initiation period denotes the end of service life. under the influence of load-induced stresses. Trans-
However, since a concrete structure does not imme- port mechanisms and corrosion cell development in
diately lose its strength or functionality at the onset cracked concrete may be very different, in compari-
of the propagation period, a more sensible approach son to uncracked concrete.
would make use of a detailed maintenance strategy Further research needs to be done to establish the
for corrosion-damaged structures. The definition of a influence of cracks on the corrosion of steel rein-
suitable maintenance strategy depends on knowledge forcement. It appears ineffective to model the ingress
of the following aspects: of harmful substances into uncracked concrete when
existing cracks may accelerate the deterioration proc-
− influence of steel reinforcement corrosion on the
ess. In particular, the influences of various crack
load-bearing capacity of structural members
widths on the ingress of harmful substances into
− performance (and durability) of materials and
concrete needs to be established and considered in
systems for repair and protection of corrosion-
the modeling of service life.
damaged structures.
In another paper presented at ICCRRR 2008,
Both of the above aspects still need further research Otieno et al (2008) discuss the influence of crack
to be fully understood in the context of service-life widths on reinforcement corrosion in more detail.
modeling. Two research projects are currently being
undertaken at the University of Cape Town, address- 3.4.6 Remedial measures for structures that
ing the above 2 aspects. Aspects relating to the struc- do not meet durability specifications
tural capacity of corrosion-damaged reinforced beams The South African Durability Index approach enables
are discussed by Malumbela et al (2008). engineers to specify certain durability parameters

78
(indexes) in relation to the anticipated service life, From a technical point of view, it needs to be
environmental conditions, binder types and cover established, which coatings can be used to either pre-
depth requirements. Durability specifications com- vent or slow down the ingress of chlorides or carbon
monly comprise limiting values for the thickness and dioxide sufficiently. The performance of protective
penetrability of the concrete cover. When limiting coatings commercially available can commonly be
values, obtained on the as-built structure, meet the shown to be promising in short-term tests. However,
specified requirements, the structure is considered to there is a lack of data available on the durability of
be inherently durable. such coatings. Future research is needed to fill this
However, a clear design methodology for concrete gap of knowledge. A current project at the University
structures that do not meet the specified requirements of Cape Town is dealing with these issues.
needs to be established. If limiting durability index Another promising protection method for rein-
values have not been achieved, the owner of the struc- forced concrete structures is the application of corro-
ture principally has the following options: sion inhibitors. However, also for these materials the
long-term efficiency still needs to be established.
− Demolish and rebuild the structure
− Accept that the anticipated service life duration
may not be reached
− Protect the structure against the ingress of harmful 4 CLOSURE AND OUTLOOK
substances, such as carbon dioxide and chlorides
− Accept that harmful substances can reach the rein- In this paper, we have tried to undertake a particularly
forcement, but protect the reinforcement against difficult task: to define what future users of concrete
corrosion. will need from the research that we are currently
For most projects, the first of the above options will undertaking. Nevertheless, we feel that the issues we
be undesirable, for obvious reasons. The second option have raised are practical and draw directly from our
involves a re-evaluation of the original design param- experience of the knowledgebase of concrete deterio-
eters and may in many cases also not be acceptable. ration and durability research.
Probably in most cases, the third or fourth option will It is important to acknowledge that the global con-
be aimed at, i.e. protecting the structure against dete- crete research community has made remarkable con-
rioration to ensure that the design service life can be tributions to our understanding of concrete durability
reached. Such methodology may for example include and deterioration. In the same breath, this paper has
the application of protective surface coatings or corro- tried to emphasise that there is much understanding
sion inhibitors. Depending on the quality of the structure that has yet to be developed through research. The par-
(by how much did it not reach the limiting durability ticular problem of deterioration due to reinforcement
design parameters?), a once-off application may be suf- corrosion will continue to demand much research
ficient, whereas in other cases a detailed maintenance time and effort before we can say that international
plan may need to be established, taking repeated appli- concerns in this area have largely been allayed.
cation of protective measures into account. This paper has also made a strong case for a more
The decision of appropriate repair and mainte- rigorous approach to service life modelling, durabil-
nance strategies needs to be based on an evaluation ity design and specification. In this regard, the pro-
of the expected service life. For this, the measured posals presented in this paper are intended as a basis
durability index value needs to be used as an input for the development of a framework towards address-
parameter in the service life model, with which the ing this concern.
original design parameter was established. This will Finally, the important challenge of technology
allow an estimation of the actual service life duration transfer should not be neglected. Our success in these
that can be expected. This, in turn, will give the infor- areas of research endeavour will be measured by the
mation of how many years of additional service life extent to which we positively influence the practice
the protective measure needs to provide. of concrete technology to produce more durable con-
Based on this, it can for example be argued that a crete structures.
coating, which prevents the ingress of harmful sub-
stances over that required duration, presents a suitable
protective measure, bringing the structure back to its REFERENCES
original service life. However, a clear philosophy
needs to be developed, based on which the design of Alexander, M.G., Mackechnie, J.R. and Ballim, Y. Guide to
appropriate protective measures can be carried out. the use of durability indexes for achieving durability in
The design engineer and the owner of the structure concrete structures, Research Monogram No. 2, Depart-
need to be given clear guidance on what steps to fol- ment of Civil Engineering, University of Cape Town,
low and on what options are available. March 1999, 35 pp.

79
Alexander, M.G. and Stanish, K. (2005), ‘Durability design Mackechnie, J.R. Predictions of Reinforced Concrete
and specification of reinforced concrete structures using Durability in the Marine Environment, Research Mono-
a multi-factor approach’. Mindess Symposium, Third graph No. 1, Department of Civil Engineering, Univer-
Int. Conference on Construction Materials, Vancouver, sity of Cape Town, 2001, 28 pp.
August 2005, CD ROM, University of British Columbia, Malumbela, G., Alexander, M.G. and Moyo, P. Structural
Vancouver, 2005, 10 pp. behaviour of beams under simultaneous load and steel
Beushausen, H., Alexander, M.G. and Mackechnie, J. Con- corrosion, International Conference on Concrete Repair,
crete durability aspects in an international context, Con- Rehabilitation and Retrofitting, Proceedings ICCRRR
crete Plant and Precast Technology BFT, vol. 7, 2003, 2008, Cape Town, 24–26 November 2008.
Germany, pp. 22–32. Neville, A. Concrete: Neville’s Insights and Issues, Thomas
European Committee for Standardization (2004), BS EN Telford, London, 2006. 314 pp.
1992-1-1 Eurocode 2: Design of Concrete Structures— Otieno, M., Alexander, M.G. and Beushausen, H. Corrosion
Part 1-1: General rules and rules for buildings (European Propagation in Cracked and Uncracked Concrete, Inter-
standard prEN1992-1-1). CEN, Brussels, Belgium. national Conference on Concrete Repair, Rehabilitation
fib bulletin 34: Model Code for Service Life Design, Swit- and Retrofitting, Proceedings ICCRRR 2008, Cape
zerland, 2006, 110 pp. Town, 24–26 November 2008.
Heiyantuduwa, R. and Alexander, M.G., Studies on pre- SANS 10100-2 (2005), Structural use of concrete Part 2—
diction models for concrete durability, International materials and execution work, 3rd Edition.
Conference on Concrete Repair, Rehabilitation and Tuutti, K. Corrosion of steel in concrete, Stockholm: Swed-
Retrofitting, Proceedings ICCRRR 2008, Cape Town, ish Cement and Concrete Research Institute. In: CBI
24–26 November 2008. Research, Report No. 4:82, 1982.

80