BENCHMARK FOR DEEMED-TO-SATISFY RULES (XD, XS)

FOR fib-CONGRESS February 2014 MUMBAI

Greve-Dierfeld, S. von
former Materials Science and Testing (MST), TU Muenchen, Baumbachstr. 7, 81245 Muenchen,
Germany.

Gehlen, Ch.
Chair of Materials Science and Testing (MST), TU Muenchen, Baumbachstr. 7, 81245
Muenchen, Germany.

Abstract
The current system for specifying and ensuring durability of new concrete structures in standards is
commonly of a prescriptive type. In prescriptive specifications, durability is guaranteed indirectly
by ensuring compliance with limiting values for concrete composition and construction details.
These empirical provisions have typically evolved from local experience and the local
availability of concrete constituents. They are based on the individual preferences on safety without
any type of mathematical or scientific verification. One of the practical results is that there is an
enormous variation in requirements between the various countries all over the world and even in
Europe when close regional proximity is given. However, the different national provisions cannot
be explained on a rational basis and it is likely that they do not lead to a consistent exposure
resistance.
The aim of this work is to perform a benchmark for deemed-to-satisfy rules for the exposure
classes XD and XS. Within the benchmark it is determined which reliabilities against chlorideinduced depassivation of rebars can be expected if the deemed-to-satisfy rules of different countries
are considered. This includes not only calculations mainly based on short term laboratory data, but
also an independent assessment of existing structures.
The calculated reliability ranges determined are compared with the target reliabilities proposed by
current specifications and, based on the above comparison, a proposal for improving deemed-tosatisfy rules and specifications is made.

Keywords: Durability, deemed-to-satisfy rules, performance based rules, chloride, reliability

design, assessment of existing structures

Introduction

In prescriptive specifications durability with respect to chloride induced corrosion is guaranteed by

ensuring compliance with limiting values for:
- maximum w/c-ratio, minimum cement content, limits on the permitted types of cement and
their components and / or limits for the compressive strength, together with
- a minimum concrete cover mostly combined with an allowed tolerance, both in dependence
of
- exposure class, giving the classified degree of severity of the environmental load.
Although globally these limiting values vary over a wide range of values and although these
values are often combined differently, they all follow the same goal: provide enough resistance in
order to avoid chloride induced corrosion with sufficient reliability.

Hence, one aim of the benchmarking was to determine whether the limiting values and their
various combinations fulfil the target reliabilities of current specifications. In order to verify the
reliability provided by the deemed-to-satisfy rules, the initial goal was to
- identify those prescriptive values that mainly affect the material performance and to relate
material performance (under constant, standardized conditions) to the prescriptive
specifications.
- transfer the definitions of the exposure classes in meso-climatic environmental loads in order
to describe the material performance under exposure.

Background: model, parameters and data used for the benchmark

task

2.1 Mathematical model

Diffusion, permeation and convection are the means by which chlorides penetrate concrete.
Assuming, diffusion to be the prevailing transport process, the following mathematical expression
is used to describe the rate at which chlorides penetrate the concrete, Eq. (1).

C
x(C,t) = 2 Dapp ( t ) erf -1 1 t
C
s

t
with Dapp (t) = D0 0
t

(1)

Here x in [m] is the depth with a defined critical chloride content (C = Ccrit) at time t, Cs is the
chloride surface concentration [wt.%/cement], Dapp(t) in [m2/s] is the apparent diffusion coefficient
at time t [s], D0 is the diffusion coefficient measured at a reference time t0 and in [-] is the aging
exponent, giving the decrease over time of the apparent diffusion coefficient.
This is a simple transformation of the well-known expression, where the chloride content
C(x,t) [wt.%/cement] at depth x and time t is given by Eq. (2).

x
C( x , t ) = C s erf

2 Dapp ( t ) t

(2)

The initial chloride content (Ci) can be included in both the above equations (ISO 16204, 2012).
One can conclude that the rate at which chlorides penetrate concrete is governed by the diffusivity
of the concrete (material) and concentration of the chloride load (environment).
2.2 Data
Material
The apparent chloride diffusion coefficient is used to describe the diffusivity of the concrete. In
saturated concrete, the apparent diffusion coefficient is mainly controlled by the pore structure
(amount, size and tortuosity) and is less affected by ionic interactions.
Usually, the chloride diffusion coefficients for specific concrete compositions are either
determined in accelerated tests or in short term tests both under different experimental condition.
Mostly, non-steady-state diffusion coefficients are determined. In the accelerated tests methods,
migration coefficients are determined by accelerating the chlorides in an electric field, where
transport by diffusion is negligible (e.g. NT BUILD 492, 1999), or the diffusion coefficients are
determined with increased chloride concentrations (e.g. NT BUILD 443, 1995). In short term tests,
diffusion coefficients are determined under natural chloride load (e.g. DD EN/TS 12390-11,
2010). The migration coefficient is assumed to correlate well with the diffusion coefficients at early
ages (Tang et al. 2010, Gehlen, 2000, Tang, 1996).
In Fig. 1 (left) a summary of migration coefficients (t0 = 28 d) in dependence of the type of
cement and w/c-ratio is given since these two parameters strongly affect the pore structure of the

30

CEM II/A,B-LL

25

CEM I

20
CEM II/A,B-V

15
10

CEM II/B-S
CEM II/A-D

CEM III/A,B

0
0.4

0.45

0.5
w/c-ratio [-]

0.55

0.6

30
Dapp(t0) [10-12 m/s]

DRCM [10-12m2/s]

concrete. The amount of binder, providing a minimum for sufficient compaction is guaranteed, as
well as by the type, shape and size of aggregate are less important (Lay & Schiel, 2002). The
migration coefficients where derived from (Bjegovic et al., 2012, Heinz et al., 2011, Mller &
Severins, 2009, Mller et al., 2009, Visser & Nijland, 2009, Gruyaert et al., 2009, Wiens, 2005,
Lay & Schiel, 2002, Gehlen, 2000, Chrisholm & Lee, 2001). In Fig. 1 (right) the correlation
between chloride migration coefficient and chloride diffusion coefficient according to (NT BUILD
443, 1995) and (DD EN/TS 12390-11, 2010) is shown (Tang et al., 2010).
prEN/TS 12390-11
NT BUILD 443

25
20
15
10
5
0
0

5 10 15 20 25 30
DRCM(t0) [10-12 m/s]

Fig. 1 Chloride diffusion coefficient at time t0 = 28 d (left), correlation between chloride migration and
chloride diffusion coefficient (Tang et al, 2010) (right).

In Fig. 1 the symbols represent the mean value of the chloride migration coefficients determined by
various authors. An exponential function was fitted to the mean values. Due to the random scatter
for CEM II/A,B-LL, CEM II/A,B-V and CEM III/A,B (A: full symbols, B: empty symbols) one
curve was fitted to both cement clinker contents (A and B).
In Fig. 1 (left) it can be seen that chloride migration coefficients increase exponentially with
w/c-ratio. At the same w/c-ratio, concretes with high contents of ground granulated blast-furnace
slag (GGBS) show low chloride migration coefficients due to their low porosity (CEM III).
Generally, the chloride migration coefficients decrease continuously from 30 wt.% GGBS content
up to 50 wt.% GGBS (Lay & Schiel 2002). Concretes with silica fume (SF) (CEM II/A-D)
possess low chloride migration coefficients mainly due to the beneficial filler effect of SF. The
same holds for concretes with fly ash (FA) (CEM II/A-V, CEM II/B-V), but with less effect. The
highest chloride migration coefficients are possessed by concretes with limestone (LS) which is
due to the increased porosity - especially if more than 15 wt.% is used (Lay & Schiel, 2002).
CEM I concretes show slightly lower migration coefficients than CEM II/A,B-LL concretes.
The chloride diffusion coefficient decreases due to ongoing hydration and reaction products
that lead to changes in the pore structure. The decrease in chloride diffusion coefficient is
represented by the aging exponent.
The aging exponent under constant saturated conditions is mainly affected by the type of
cement and its components (Gehlen, 2000). Ordinary Portland cement concretes (OPCC), and
concretes with components like LS (CEM II/A, B-LL), SF (CEM II/A-D) or low contents of GGBS
(CEM II/A, B-S) have the lowest aging exponents ( ~ 0.30), due to the negligible ongoing
hydration and the low binding capacity. Higher aging exponents are possessed by concretes with
higher amounts of GGBS (CEM III) due to the ongoing latent hydraulic reaction and the higher
binding capacity ( ~ 0.45). Concretes with fly ash (CEM II/A, B-V) show the highest aging
exponent due to the ongoing puzzolanic reaction and the high binding capacity ( ~ 0.60). The
values for the ageing exponents given are derived from (Caballero et al., 2010, Polder et al., 2010,
Markeset & Skjlsvold, 2010, Nokken et al., 2006, Stanish & Thomas, 2003, Gehlen, 2000).
Taking into consideration aging behaviour of different types of cement: concretes with lower
resistances (with higher migration coefficients or diffusion coefficients, respectively) determined in
short-term tests may have a higher long-term resistance. Thus, from the migration coefficients of
Fig. 1, the apparent diffusion coefficients at time t = 50 years where calculated with the aging

Dapp(t=50) [10-12m2/s]

function (Eq. (1) right) and the aging exponents for the specific type of cement introduced above,
see Fig. 2.
5

CEM II/A,B-LL

CEM I

3
2

CEM II/B-S
CEM II/A-D
CEM II/A,B-V

CEM III/A,B

0
0.4

0.45

0.5
w/c-ratio [-]

0.55

0.6

Fig. 2 Apparent chloride diffusion coefficients at time t = 50 years, calculated for the cement types of Fig.
1 (left) using Eq. (1) (right) and the aging exponents introduced above.

Environment
The chloride surface concentration strongly affects the concentration gradient. The chloride surface
concentration under saturated conditions depends on the chloride concentration of the ambient
solution as well as on the adsorption and binding capacity of the concrete. Especially under
immersed conditions, the chloride surface concentration is affected by the type of cement. The
surface concentration of concretes with GGBS or FA is increased due to their higher binding
capacity (higher C3A, C4AF contents). For concretes with OPC or additives like SF, the chloride
surface concentration is decreased due to the lower binding capacity. Furthermore, the chloride
surface concentration increases with increasing w/c-ratio (adsorption). The amount of binder has
less effect.

XD3

Chloride surface
concentration [wt.%/c.]

Chloride surface
concentration [wt.%/c.]

The environmental load is loosely defined in the exposure classes by the description of the chloride
source and humidity conditions. The effect of chloride load on material performance is governed
by the chloride surface concentration Cs. The effect of moisture content is included in the surface
concentration Cs and the aging exponent .
The chloride surface concentration under real exposure conditions is subject to systematic
spatial variations documented in dependence of the distance of the concrete surface to the chloride
source in urban (DARTS 2004) as well as marine environments (Markeset & Skjlsvold, 2010,
Helland et al., 2010, Wall, 2007, Nokken et al., 2006, Ghods et al., 2005, Fluge et al., 2001), Fig. 3.
Herein, our own data (black symbols) is included derived from four car parks in Munich (XD3) and
from exposed specimens in Helgoland (XS3). In Fig. 3, the chloride surface concentration
determined from field exposure (meso-climatic conditions) is shown in dependence of the distance
to the chloride source to the concrete surface and related to the exposure classes.
XD1

XD2

4
3

horizontal distance
from road:

car
prarks

0.5 m

2.0 m

0
0

2
3
4
Distance to chloride source [m]
(height above road)

XS3

XS2

XS1

5
CEM I+FA

CEM I

CEM I

3
2
1
0
-5

< MLT

10

15

20

25

~ wave crest

Distance to chloride source [m]

Fig. 3 Chloride surface concentration in dependence of the distance to chloride source for urban (DARTS,
2004) (left) and marine environments (Markeset & Skjlsvold, 2010, Helland et al., 2010, Wall, 2007,
Nokken et al., 2006, Ghods et al., 2005, Fluge et al., 2001) (right) meso-climatic conditions both in wt. % per
cement.

In Fig. 3 (left) for wet, rarely dry conditions (XD2) no systematic spatial variability is considered.
For cyclic wet and dry conditions (XD3) and in moderate humidity (XD1), a constant decrease with
the distance to the chloride source occurs.
In Fig. 3 (right) for permanently submerged conditions (XS2) no systematic spatial variability
is reported. The chloride surface concentration corresponds with the salinity of the ambient
solution. For tidal and splash zone (XS3), the chloride surface concentration decreases rapidly with
the distance to the chloride source. In the spray zone (XS1), a further decrease in chloride surface
concentration occurs. Furthermore, it can be seen that differences due to chloride binding and
adsorption in the exposure classes XS3 and XS1 disappear against the background of random
spatial variability.

1.0

XD2, XD3, XS2, XS3

0.8
0.6
0.4
0.2
0.0
CEM I
(+ LL, S, D)

CEM I
(+ FA)

CEM III

age exponent [-]

age exponent [-]

The effect of humidity is empirically included in the lower surface concentration and is partly due
to the reduced adsorption. Furthermore, the diffusion process is reduced in partly saturated concrete
(spray zone: XD1, XS1), which results in an increased aging exponent (Fig. 4). The degree of
saturation in dependence of relative humidity decreases with increasing pore radius. Because of
this, higher differences in the aging exponents for OPCC (more dry) concretes are reported, smaller
differences are reported for FA concretes (less dry) see Fig. 4 (cf. left with right). The aging
exponents in Fig. 4 are taken from (Caballero et al., 2010, Polder et al., 2010, Markeset &
Skjlsvold, 2010, Nokken et al., 2006, Stanish & Thomas, 2003, Gehlen, 2000).
1.0

XD1, XS1

0.8
0.6
0.4
0.2
0.0
CEM I
(+ LL, S, D)

CEM I
(+ FA)

CEM III

Fig. 4 Aging exponents for nominally saturated conditions (XS2, XS3, XD2, XD3) (left) and partly
saturated conditions (XS1, XD1) (right) from (Caballero et al., 2010, Polder et al., 2010, Markeset &
Skjlsvold, 2010, Nokken et al., 2006, Stanish & Thomas, 2003, Gehlen, 2000).

Furthermore, it is known, that temperature affects binding capacity and ion mobility and hence the
diffusion process. The effect of temperature on the diffusion coefficient is accounted for by the
Arrhenius function with the regression parameters given in (fib, 2006).
At least, deviations from Ficks second law of diffusion for example due to leaching,
capillary suction and interactions with effects like carbonation in the exposure classes XD3, XS3
have to be taken into consideration. One possibility is to use an apparent chloride surface
concentration calculated for the surface at x = 0. Another possibility is to use the highest chloride
concentration (Cs,x) in the depth x and to deduct the depth x from x, as applied here.
Summary
The prescriptive values minimum w/c-ratio and recommendations with respect to the type of
cement and its components are the decisive parameters that govern material resistance. The
cement content, provided a minimum for sufficient compaction is guaranteed, is of less
importance. The effect of compressive strength is already given by the w/c-ratio and may be
regarded as supplementary control measure.
An increase in w/c-ratio leads to an increased chloride diffusion coefficient and higher
chloride surface concentrations. In contrast, types of cement which yield low diffusion coefficients
lead to high chloride surface concentrations and vice versa. Furthermore, the effect of cement type

on material performance is strongly time dependent. Type of cement with low diffusion
coefficients at young ages may possesses higher diffusion coefficients at higher ages.

Benchmarking

3.1 Selection of countries

Regarding material performance under immersed conditions (more or less constant), the w/c-ratio
and the type of cement are the decisive parameters that affect material performance. Regarding
material performance under exposure, the decisive parameters are the location (distance to chloride
source, humidity condition and the chloride content of the ambient solution) to be taken into
account when selecting a country.
Thus for the selection of countries, i.e. the respectively national specifications, the following was
considered:
I.
diversity in deemed-to-satisfy rules: w/c-ratio, cover, types of cement and the diversity in
construction practice given by the local availability of specific types of cement
II.
diversity in environmental load (exposure class definition, salinity of seawater,
temperature conditions and frequency of de-icing salt application)
Point I: a significant diversity in deemed-to-satisfy rules in comparison with EN 1992-1-1/EN 2061 (EN 1992-1-1, 2002, EN 206-1, 2001) may be found when considering Australia (AUS), United
States of America (USA), Portugal (P), Spain (E) and Great Britain (GB), the Netherlands (NL),
Germany (D), Denmark (DK) and Norway (N). The variety of permitted types of cements and their
combination with w/c-ratio and / or cover are summarized in Table 1.

Table 1
Diversity in deemed-to-satisfy rules for structures exposed to chlorides
Country
Spain (E)

Diversity in deemed-to-satisfy rules and permitted types of cement

CEM III, CEM IV,
Others
CEM II/B-S, B-P, B-V, A-D combined
combined with a higher cover for
with a lower cover for one w/c-ratio
one w/c-ratio
Portugal (P)
CEM II/A-D, CEM II/B, CEM III/A, CEM
CEM I, CEM II/A
III/B, CEM IV/A, CEM IV/B, CEM V combined with a lower w/c-ratio for
combined with a higher w/c-ratio for one
one cover
cover
Great Britain (GB)
Groups of types of cement combined with w/c-ratio correlated with cover
The Netherlands (NL)
All types of cement; One cover and w/c-ratio within each exposure class
Germany (D)
All except: CEM II/B-LL, CEM II/B-L, CEM II/A-W, CEM II/B-W, CEM
IV/A, CEM III/C{1}, one cover and w/c-ratio within each exposure class
Denmark (DK)
XD1, XS1, XS2:
XS3, XD2, XD3:
CEM I, CEM II/A-L, CEM II/A-LL and
CEM I and CEM II/A-V
CEM II/A-V
One cover and w/c-ratio within each exposure class
Norway (N)
XD1, XS1:
XD2, XD3, XS2, XS3:
CEM I, CEM II/A-S, CEM II/B-S, CEM CEM I + 4 M.% SF/c., CEM II/AII/A-D, CEM II/A-V, CEM II/B-V, CEM S, CEM II/B-S, CEM II/A-D, CEM
III/A
II/A-V, CEM II/B-V, CEM III/A
One cover and w/c-ratio within each exposure class
United States of America
All, except LS or GGBS contents > 70 % (Beushausen & Fernandez, 2011)
(USA)
One cover and w/c-ratio within each exposure class
Australia (AUS)
No recommendations on type of cement and w/c-ratio: durability is verified
with respect to compressive strength combined with cover
{1}Not considered in exposure classes XD2 and XS2

In the Netherlands traditionally concretes with BFSC are preferred. In Germany preferably CEM II
concretes are used. In Norway concretes with additives like SF of FA are favoured.
Point II: while in P, GB, NL, D, DK and N the exposure classes (European definition: XD and XS)
are applied as defined in (EN 1992-1-1, 2002), in AUS, USA and E exposure classes are defined
individually. Especially in AUS the class definition is not identical with the European definition,
which leads to an overlap of the classes. The salinity of oceans is similar with ~35 g/L (except the
Baltic Sea with a salinity of ~8 g/L). Temperature affects the mobility of ions and hence the
diffusion rate, as well as the frequency of usage of de-icing salts. The temperature ranges from
~15C (E and P Southern Europe, AUS), ~10C (GB, NL, D Central Europe, USA) to ~8C
(DK, N Northern Europe) mean annual temperature.
Thus the following countries are considered, see Table 2. Herein, additionally the respective
specifications are summarized.

Table 2
Selected countries and specifications taken into account
Country
Considered specifications
Spain (E)
EHE-08 (2008)
Portugal (P)
LNEC E 464 (2007), NP EN 206-1 (2007)
Great Britain (GB)
BS 8500-1 (2006), BS 8500-2 (2006), EN 206-1 (2000)
The Netherlands (NL)
NEN 8005 (2008), EN 206-1 (2000)
Germany (D)
DIN 1045-2 (2008), EN 206-1 (2000)
Denmark (DK)
DS/EN 1992-1-1 DK NA: 2011, DS 2426 (2011), EN 206-1 (2000)
Norway (N)
NS-EN 206-1 (2007), NS-EN 1992-1-1 (2008), NS EN 13670 (2010)
United States of America (USA)
ACI-318-08 (2008)
Australia (AUS)
AS 3600 (2009)

In all countries, except USA, deemed-to-satisfy rules are provided for a design service life of
around 50 years. In some countries additionally rules for a design service life of 100 years are
given. In this publication resistances and reliabilities provided from deemed-to-satisfy rules were
analysed for a design service life of 50 years.
3.2 Procedure
Based on Table 1 and Section 2 the following may be concluded.
I.
All prescriptive specifications give freedom in choice of cement type. Some prescriptive
specifications give freedom in the choice of combined limiting values (E, P, GB). Hence,
within one exposure class, a concrete composition with a lower resistance (low-resistance
concrete) or a concrete composition with a higher resistance (high-resistance concrete)
may be chosen (Fig. 2).
II.
Within one exposure class a high or low chloride load is possible (Fig. 3).
Both I. and II. will result in different structural reliabilities (spectrum of reliabilities) within
one exposure class. Therefore, the goal of this work was to determine the spectrum of reliabilities
provided by choosing unfavourable design situations (lower reliability level) and favourable design
situations (upper reliability level). Unfavourable design situations are characterized by a lowresistance concrete composition (type of cement and w/c-ratio) in combination with a high chloride
load. Favourable design situations are characterized by a high-resistance concrete composition in
combination with a low chloride load.
The procedure to determine the levels of reliability is explained with the example of Germany
exposure class XS3 in detail. The limiting values for the exposure class XS3 as stated in DIN 10452, 2008 are summarized in Table 3.

Table 3
Limiting values (DIN 1045-2, 2008)
EnvironMaterial performance
mental load
Exposure Minimum cement Minimum
Type of cement [-]
class
content [kg/m] w/c-ratio [-]
XS3

320

0.45

Structural
performance
Nominal concrete
cover cnom = cmin + c
[mm]
All except CEM II/B-LL, CEM II/B-L,
55 = 40 + 15
CEM II/A-W, CEM II/B-W,
CEM IV/A, CEM III/C

CEM II/A,B-LL

CEM I

3
unfavourable

2
1

CEM II/B-S
CEM II/A-D
CEM II/A,B-V
CEM III/A,B

favourable

0
0.4

0.45

0.5
w/c-ratio [-]

0.55

0.6

Chloride surface
concentration [wt.%/c.]

Dapp(t=50) [10-12m2/s]

The nominal concrete cover is stated in the specification. The difficulty which arises is the question
as to which type of cement in combination with maximum w/c-ratio yields the low-resistance
concrete or the high-resistance concrete? Which chloride loads in combination should be selected?
6

XS3

XS2

XS1

unfavourable

4
3

favourable

2
1
0
-5

< MLT

10

15

20

25

~ wave crest

Distance to chloride source [m]

Fig. 5
Material performance in dependence of type of cement and w/c-ratio (left), chloride surface
concentration in dependence of the exposure class (right).

In Fig. 5 (left) it can be seen that a low-resistance concrete is provided by CEM II/A-LL. A highresistance concrete is provided by CEM III/B, both with a minimum w/c-ratio of 0.45.
In Fig. 5 (right) obviously a high chloride load occurs at a short distance from the chloride
source and a low chloride load further from the chloride source.
Based on the above considerations the unfavourable design situation was chosen for a CEM
II/A-LL and a chloride surface concentration of 4.0 wt.%/c. The favourable design situation was
chosen for a CEM III/B and a chloride surface concentration of 2.0 wt.%/c. Both with a w/c-ratio
of 0.45.
All the parameters, which have been described in detail in Section 2 with respect to their
effect on material performance, are listed in in Table 4. In addition to the explanations in Section 2,
all these variables are linked with uncertainties due to the inhomogeneity of material, execution
quality, random temporal and spatial variability of the environmental conditions etc. In order to
take these uncertainties into account each variable was described as a random variable with a
distribution function and its distribution parameters (mean and standard deviation). These
uncertainties were quantified as follows.
Chloride diffusion coefficients are sufficiently described by a normal distribution (Gehlen,
2000). The reproducibility is taken into account with a constant coefficient of variation of 0.4.
According to (Gehlen, 2000), the aging exponent may be sufficiently described using a beta
distribution with the boundary conditions 0 < < 1. For saturated conditions (XD2, XD3, XS2,
XS3) the standard deviation was set to 0.12, 0.15 and 0.20 for CEM I, CEM II and CEMIII
concretes, respectively. For dry conditions (XD1, XS1) the standard deviation was set to 0.12.
The chloride surface concentration varies randomly due to variations in the chloride content
of the ambient solution, frequency of use of de-icing salts, temporal and spatial variations in the
humidity conditions of the concrete etc. The chloride surface concentration may be sufficiently
described with a lognormal distribution. For urban environments (XD1, XD2 and XD3) a CoV of
75% is documented in DARTS, 2004 because of the huge variability in frequency and amount of
used de-icing salts. For submerged conditions in marine environments (XS2) a low CoV of 25 % is
given which is mainly due to the inhomogeneity of concrete surfaces and the variation in
composition of the ambient solution. With increasing distance to the chloride source (XS3, XS1),

the random spatial variations increase from a CoV of 45 % (DARTS, 2004) up to 65% (Wall,
2007).
The concrete cover may be sufficiently described with a normal distribution. The mean
concrete cover is the nominal concrete cover a which is the sum of minimal cover and an allowed
tolerance. The standard deviation is independent of the mean value and is given by 9 mm (Gehlen,
2000, JCSS, 1999).
Information on the critical chloride content may be found for example in (Breit et al., 2011,
Angst et al., 2009). The critical chloride content may be described by a beta distribution with mean
of 0.6 and a standard deviation of 0.15 with a lower boundary of 0.2 and an upper boundary of 2.0.
The reliability index was calculated according to Eq. 3, where the depth with a critical chloride
content (C = Ccrit) is compared with the concrete cover a. Both are described as random variables
(Ditlevsen & Madsen, 2005).
(3)
( ) = pdep. = p{x( Ccrit ,t ) a} < p0 < ( 0 )

Table 4
Model parameters for the unfavourable and favourable design situation (D), XS3

Variable

Symbol
DRCM,0

Dapp(t)

ke (Gehlen, 2000)
Depth with deviation
from Fick`s second
law of diffusion
(Gehlen, 2000)
Chloride concentration at depth x
Critical chloride
content
(Gehlen, 2000)
Cover
Design service life

Unit
10

-12

Distribution
2

m /s

Normal
Beta
a=0.0 b=1.0
Constant
Constant
Normal
Normal

unfavourable
(CEM I)
Standard
Mean
deviation
10.0
4.0

favourable
(CEM III/B)
Standard
Mean
deviation
1.9
0.8

0.30

0.12

0.45

0.20

0.0767
293
283
4800

5
700

0.0767
293
283
4800

5
700

t0
Tref
Treal
be

a
K
K
K

mm

Beta
a=0 b=50

10

10

Cs,x

wt.%/c

Lognormal

4.0

1.8

2.0

1.3

ccrit

wt.%/c

Beta
a=0.2 b=2.0

0.60

0.15

0.60

0.15

a
tSL = t

mm
a

Normal
Constant

55
50

9
-

55
50

9
-

The reliability index as a function of time for the unfavourable and the favourable design situation
is shown in Fig. 6. The area between the unfavourable and the favourable design situation
represents the spectrum of expected reliabilities. The spectrum of reliability at the end of design
service life (50 years) is the one provided when applying the deemed-to-satisfy rules of the German
specification for the exposure class XS3.

Reliability index [-]

4
3

CEM III/B
favourable

2
1

reliability
spectrum

0
-1

CEM I
unfavourable

-2
0

10

20

30

40

50

Time t [a]

Fig. 6 Reliability index versus design service life for the unfavourable and the favourable design situation.
Reliability spectrum provided when following the recommendations of deemed-to-satisfy rules of German
specification for XS3.

3.3 Reliability spectra

w/c-ratio [-]

cmin [mm]

w/c-ratio [-]

cmin [mm]

w/c-ratio [-]

cmin [mm]

This methodological approach was applied for each country and exposure class.
In Fig. 7 a and b (left) the w/c-ratio and the concrete cover is given for each country and
exposure class. Additionally, the favourable and unfavourable types of cement are stated which
may be selected when taking into consideration the permitted types of cement of Table 1. In Fig. 7
a and b (right) the reliability spectrum (bars) derived from the unfavourable and favourable design
situation at the end of design service life (50 years) is shown for each country and exposure class.
70
60
50
40
30
20
0.7
0.6
0.5
0.4
0.3
70
60
50
40
30
20
0.7
0.6
0.5
0.4
0.3
70
60
50
40
30
20
0.7
0.6
0.5
0.4
0.3

Reliability index [-]

4

CEMI

CEM III

XD1

w/c = 0.60

w/c = 0.45
unfavourable CEM I, II/A-LL

CEMIII

favourable CEM III/B, II/A-V

1
0
-1

CEMI

-2

CEMI

CEM III

w/c = 0.55

w/c = 0.4

XD2

1
unfavourable CEM I, II/A-LL

CEMIII

favourable CEM III/B, II/A,B-V

0
-1

CEMI

-2

CEMI
w/c = 0.5
w/c = 0.35

CEM III

XD3

3
2
unfavourable CEM I, II/A-LL, II/A-S
favourable CEM III/B, II/A,B-V

1
0

CEMIII

-1
CEMI

GB

NL

DK

USA AUS

-2

GB

NL

DK

USA AUS

Fig. 7a Favourable and unfavourable type of cement, maximum w/c-ratio and minimum cover (left),
reliability spectra (right).

cmin [mm]
w/c-ratio [-]
cmin [mm]
w/c-ratio [-]
cmin [mm]
w/c-ratio [-]

70
60
50
40
30
20
0.7
0.6
0.5
0.4
0.3
70
60
50
40
30
20

Reliability index [-]

4
CEMI

w/c = 0.55

CEM III

w/c = 0.35

2
unfavourable CEM I, II/A-LL

CEMIII

1
0
-1

CEMI

-2

4
CEMI

w/c = 0.55

CEM III

w/c = 0.4

0.7
0.6
0.5
0.4
0.3

70
60
50
40
30
20
0.7
0.6
0.5
0.4
0.3

favourable CEM III/B, II/A-V

XS1

unfavourable CEM I, II/A-LL

CEMIII

favourable CEM III/B, II/A,B-V

XS2

1
0
-1

CEMI

-2
4

CEMI
w/c = 0.5
CEM III

2
unfavourable CEM I, II/A-LL, II/A-S
favourable CEM III/B, II/A,B-V

CEMIII

1
0
-1

CEMI

XS3

w/c = 0.35

GB

NL

DK

USA AUS

-2

GB

NL

DK

N USA AUS

Fig. 7b Favourable and unfavourable type of cement, maximum w/c-ratio and minimum cover (left),
reliability spectra (right), AUS: w/b ratios appraised from concrete strength requirements, exposure class C1
was assumed to be equivalent to XS1, USA: concrete cover in dependency to bar diameter, cover tabled
refers to bars of diameter ds > 19 mm

The lower reliability level (bottom of bars) results when selecting CEM I, CEM II/A-LL,
CEM II/A-S cements. The upper reliability level (top of bars) results when selecting CEM III or
CEM II/A,B-V cements.
In all exposure classes, an increased cover (E) for low-resistance types of cements or the
exclusion of low-resistance types of cement (N) leads to a narrower reliability spectrum. The same
holds for a reduced w/c-ratio (P) for low-resistance types of cement, but with less effect. An
increase in w/c-ratio of about 0.1 in combination with an increase of cover of about 10 mm yields
no significant change in reliability spread (GB).
The above reliability spectra were calculated with the listed data in Section 2. These
calculations were subject to simplifications and effects that, strictly, are not taken into proper
consideration. Some effects and simplifications are due to:
- the variety of specific surface effects ,
- neglecting the effect of discrete changes in diffusion coefficient during the design service life,
- neglecting the time-dependent development of chloride surface concentration in the first five
to ten years of exposure (Bioubakhsh, 2011, Markeset & Skjlsvold, 2010)
- neglecting the effect of binding on the concentration gradient (Baroghel-Bouny et al., 2012)
In order to quantify the effect of these simplifications, the reliabilities of some existing structures
where checked.

3.4 Verification by assessing existing structures

By assessing existing structures, the in-field performance of the material was determined. In-field
performance is given by measured chloride profiles and is used to
- analyse whether the predicted chloride concentrations reflect the real behaviour.
- verify the reliability spectra determined by reliability design with field reliability.

Methodology
In order to verify the model predictions all measured chloride profiles for individual structures
where compared with the predicted performance represented by the calculated chloride profiles.
Chloride contents were calculated for the specific design situation using Ficks second law (Eq.
(2)). The model parameters where varied randomly in a Monte Carlo approach to yield chloride
concentration distributions (50 % and 95% quantiles) at discrete depths (Ditlevsen & Madsen,
2005).
In order to verify the calculated reliability spectra in Figs. 7, the measured depth-dependent
chloride contents where used to calculate the reliability at the end of design service life within a
Bayesian update as is well-documented in (Ditlevsen & Madsen, 2005). This requires that the
material composition on-site fulfils the requirements stated in the specifications.
Field investigations
Chloride profiles from structures older than 5 years where provided by the members of fib-TG
5.11. In Table 5 the age of the investigated structural elements, the country, the related exposure
class, the description of the location, the concrete composition (w/c-ratio and type of cement) as
well as the cover required by the specific specification are summarized.

Table 5
On-site investigations from fib-TG 5.11
No.
1
2
3
4
5
6
7
8
9
10
11

age Coun- Exposure

Description of location
Type of cement w/c-ratio cmin
[a]
try
class
[-]
[mm]
5
P
XS3
Setenave shipyards, Atlantic Ocean
CEM I + 5 % SF
0.35
45
18
NL
XS3
Box girder bridge wall, North Sea
CEM III/B
0.45
40
6
D
XS2
Specimens exposed to Eckenfrde, Baltic
CEM I
0.47
40
Sea (exposed specimen)
6
XS2
CEM III/A
0.47
40
6
XS3 Specimen exposed to wall of Eider barrage,
CEM I
0.45
40
North Sea (exposed specimen)
6
XS3
CEM III/A
0.45
40
30, 40
XS3
Wall of Eider barrage, North Sea
CEM I + 5 % trass 0.45
40
(Osterminski & Gehlen, 2009)
12
DK
XS3
Bridge pile
CEM I + FA
0.38
35
32
XS3
Bridge pile
CEM I
0.40
35
8
N
XS2
Floating structure Troll B
CEM I +7 % SF
0.35
50
(Helland et al., 2010)
5, 9
XS3
Floating structure Heidrun
CEM I + 5 % SF
0.39
50
(Helland et al., 2010)

Fig. 8 shows some of the locations.

Fig. 8
Some of the locations investigated by fib-TG 5.11 (www.somague.pt, www.shz.de, www.
eiderstedt.net, Helland et al., 2010).

Results
In Figs. 9 (left) the measured chloride profiles from each field investigation are compared with the
calculated chloride profiles. In Figs. 9 (right) the evolution of the reliability indices is shown as
determined by reliability designs presented in Section 3.2 and 3.3. Furthermore, the field
reliabilities are given. Here the field reliabilities are based on the measured chloride profiles and
the concrete cover required by the standards.

Fig. 9a Measured chloride concentrations and calculated chloride concentrations (left), reliability index
versus time determined by reliability design and by assessing existing structures (right) Southern Europe.

3
2
1
0
6

No.3

5
4
3
2
1
0

6
No.4

5
4
2
1
0
6

No.5

5
4
3
2
1
0
6

No.6

3
2
1
0
-1
-2
4
3
2
1
0
-1
-2

4
3
2
1
0
-1
-2
4
3
2
1
0
-1
-2
4 0
3

-1

0
6

No. 7
tinsp = 30 years

4
2
0
6

tinsp = 40 years

4
2
0
0

10

20

30

40

50

Depth x [mm]

60

70

80

10

20

30

40

50

10

20
30
Time
Timet [a]
t [a]

40

50

-2
4
Reliability index b [-]

Chloride content [wt.%/c.]

Reliability index [-]

Reliability index [-]

No.2

Reliability index [-]

Reliability index b [-] Reliability index [-]

3
2
1
0
-1
-2
0

Fig. 9b Measured chloride concentrations and calculated chloride concentrations (left), reliability index
versus time determined by reliability design and by assessing existing structures (right) Central Europe.

6
4
3
2
1
0
6 0
5

10

20

30

40

50

60

70 80
No.9

Reliability index b [-]

3
2
1
0
-1
-2
4

2
1
0
6
No.10

5
4
3
2
1
0
6

No.11
tinsp = 5 years

4
2

Reliability index [-]

Reliability index b [-]

Chloride content [wt.%/c.]

[M.%/c.]

No.8

0
-1
-2
4
3
2
1
0
-1
-2

4
3
2

0
6

tinsp = 9 years

4
2

1
0
-1

-2

10

20

30

40

50

Depth x [mm]

60

70

80

10

20
30
Time
t [a]t [a]
Time

40

50

Fig. 9c Measured chloride concentrations and calculated chloride concentrations (left), reliability index
versus time determined by reliability design and by assessing existing structures (right) Northern Europe.

In Figs. 9 it can be seen that the measured chloride profiles agree very well with the calculated
chloride profiles. Furthermore, the field reliabilities are mostly higher than the ones determined by
reliability design (No. 1, 2, 3, 4, 5, 6, 7, 10, 11). The field reliability is only lower for No. 8 and 9,
where high chloride contents where measured. For structure No. 8, the performance of FA was
overestimated because no information about the actual FA content existed. In the case of No. 10 the
performance of SF was underestimated.
3.5 Comparison and discussion
In Fig. 10 the reliability spectra determined by the reliability design shown in Figs 8 (right) is
reproduced. Here, the field reliabilities are added using symbols. The ranges of all reliabilities are
marked by the grey beams.

Reliability index [-] Reliability index [-]

XD2

XD1

XD3

2
1
0
-1
-2
4

SF

XS1

XS2

XS3

CEM III

CEM III

SF

SF

CEM III

1
0

CEM I
CEM I

-1
-2

P GB

NL

DK N USA AUS

P GB

NL

DK N USA AUS

P GB

NL

DK N USA AUS

Fig. 10 Spectra of reliabilities provided by deemed-to-satisfy rules for a design service life of 50 years
determined by reliability design (bars) and by assessing existing structures (symbols).

In Fig. 10 it can be seen that the reliability spectrum is always broad, independent of the exposure
class. The lower reliability level determined when selecting low-resistance types of cements is
confirmed by the field investigations (CEM I). The same holds for the upper reliability level
determined when selecting high-resistance types of cement (CEM III, CEM II/A-D). Therefore, it
can be concluded that the wide spread of reliability within each exposure class results from the
huge differences in material performance.
In Fig. 10 it can be seen that the reliability level (grey beams) is systematically higher for the
exposure classes XD1 and XS1 and systematically lower for the exposure classes XD2, XD3, XS2,
XS3. The lowest reliability level (bottom bar) for the exposure classes XD1 and XS1 is = 1.5, for
the exposure classes XD2, XD3 and XS2 is -1.0 and for the exposure class XS3 is -1.5.
In current standards and supplementary documents a variety of reliability levels 0 are
proposed for durability limit states for example:
0 = 0.5 (Gehlen et al., 2008, DAfStb, 2008, Zilch & Schiel, 2001), 0 = 0.8 (Teply & Novak,
2012), 0 = 1.0 (Sarja, 2005), 0 = 1.2 (LNEC E 465, 2007), 0 = 1.3 (fib, 2006, Gulvanessian et
al., 2002, JCSS, 2000), 0 = 1.5 (EN 1990, 2002), 0 = 1.7 (JCSS, 2000), 0 = 1.8 (NEN 6700,
2005), 0 = 2.0 (LNEC E 465, 2007, Zilch & Schiel, 2001) and 0 = 2.3 (Sarja, 2005,
Gulvanessian et al., 2002, JCSS, 2000). Except for the exposure classes XD1 and XS1, the
reliability level provided by deemed-to-satisfy rules is, at least for unfavourable design situations,
below all the proposed target reliabilities.
Therefore it can be concluded that the major consequences of the current prescriptive approach are:
- lack of safety for specific design situations,
- lack of economic viability of prescriptive designed structures.
This results mainly from the lack of reliable information on the durability properties of the concrete
(type of cement) which makes it difficult to evaluate concrete quality and performance in the
respective environmental load. In order to overcome the current problems an adaption of deemedto-satisfy rules is required, where the limiting values are derived from scientifically verified
material resistances (performance based deemed-to-satisfy rules). Therefore, a system with
exposure resistance classes is proposed (Leivestad, 2013).

Proposal for performance based deemed-to-satisfy rules

The methodological approach for ensuring durability based on exposure resistance classes in
analogy to the approach for ensuring load-bearing capacity based on strength classes is shown in
Fig. 11.

Fig. 11 Methodological approach for ensuring durability based on exposure resistance classes in comparison
with the approach for ensuring load-bearing capacity based on strength classes (RSD for Exposure Resistance
Class Sea/De-Icing Salts, RC for Exposure Resistance Class Carbonation).

Exposure resistance classes classify concrete performance. Compliance of a specific concrete

composition with the exposure resistance class is tested under standard conditions using a certain
test specification. The class definition gives the characteristic material performance and the
environmental action. For example the exposure-resistance class RSD45 stands for a characteristic
chloride ingress depth (depassivation will occur with a probability of 10%) of 45 mm (chloride
resistance class). Based on the characteristic material performance the cover required to withstand
the specific environmental load (exposure class) is derived from (reliability) design. This approach
is comparable with the approach for ensuring load-bearing capacity based on strength classes.
Here, material performance is classified and characterized by strength classes. Compliance is tested
under standard conditions. The dimensions of the structural component required to withstand the
actual load are derived from (reliability) design based on the characteristic strength of the concrete.
In order to maintain the prescriptive concept deemed-to-satisfy rules have to be determined.
Deemed-to-satisfy rules have to be laid down for the concrete composition parameters (w/c-ratio,
type of cement etc.) for which a reliable relationship with the relevant material performance
(exposure resistance, e.g. chloride ingress depth) is documented. For the relevant material
performance, the different cover required to withstand the conditions of the individual exposure
classes may be given in design tables for a specified design service life. Thus, for common design
situations, no reliability design is to be performed. For the carbonation resistance classes (RC),
deemed-to-satisfy values have been developed and introduced in (Greve-Dierfeld & Gehlen, 2014).
Those rules are also based on a benchmark for deemed-to-satisfy rules performed for the exposure
classes XC1, XC2, XC3 and XC4 (Gehlen & Greve-Dierfeld, 2013).

XS3

3
2
1
0
-1
-2

P GB

NL

DK N USA AUS

Reliability index [-]

Reliability index [-]

The advantage of this new approach is that reliable information on the durability properties of the
concrete composition is available. This results in (see Fig. 12)
- an increased safety margin for specific design situations and
- an improved economic viability of the structures.
4

XS3

3
2
1
0
-1
-2

P GB

NL

DK N USA AUS

Fig. 12 Reliability spectra provided by current deemed-to-satisfy rules (left), reliability spectra provided
when specifying cover in dependence of exposure resistance class (right).

Acknowledgement
The authors wish to acknowledge the assistance and support of the fib-TG 5.11.

References
ACI-318-08 (2008), Building code requirements for structural concrete. American Concrete
Institute, United States of America.
Angst, U., Elsener, B., Larsen, C., Vennesland, O. (2009), Critical chloride content in reinforced
concrete a review. Cement and Concrete Research 39, pp. 1122-1138.
AS 3600 (2009), Concrete structures. Standards Australia GPO Box 476, Sydney, NSW 2001,
Australia.
Baroghel-Bouny, V. Wang, X., Thiery, M., Saillio, M., Barberon, F. (2012), Prediction of chloride
binding isotherms of cementitious materials by analytical model or numerical inverse
analysis. Cement and Concrete Research, Vol. 42, pp. 1207-1224.
Beushausen, H., Fernandez, L. (2011), Performance based specification and control of concrete
durability. WG3 State-of-the-Art report on prescriptive durability specifications.
Bioubakhsh, S. (2011), The penetration of chloride in concrete subject to wetting and drying:
Measurement and modelling. Ph.D. thesis, UCL University College London, Great Britain.
Bjegovic, D., Stirner, N., Serder, M. (2012), Durability properties of blended cement concrete.
Materials and Corrosion, Vol. 63, Issue 12, pp. 1087-1096.
Breit, W., Dauberschmidt, C., Gehlen, Ch., Sodeikat, C., Taffe, A., Wiens, U. (2011), Zum Ansatz
eines kritischen Chloridgehalts bei Stahlbetonbauwerken. Beton- und Stahlbetonbau, Vol.
106, Issue 5, pp. 290-298.
BS 8500-1 (2006), Concrete complementary British Standard to BS EN 206-1 Part 1: Method
of specifying and guidance for the specifier. London, Great Britain.
BS 8500-2 (2006), Concrete complementary British Standard to BS EN 206-1 Part 2:
Specification for constituent materials and concrete. London, Great Britain.
Caballero, J., Polder, R.B., Leegwater, G., Fraaij, A. (2010), Chloride penetration into cementitious
mortar at early ages. 2nd International Symposium on Service Life Design for Infrastructure,
October 2010, Deft, The Netherlands.
Chrisholm, D.H., Lee, N.P. (2001), Actual and effective diffusion coefficients of concrete under
marine exposure conditions. 20th Biennial Conference of the Concrete Institute of Australia,
Perth, Australia.
DAfStb (2008), Positionspapier zur Umsetzung des Konzepts von leistungsbezogenen
Entwurfsverfahren unter Bercksichtigung von DIN EN 206-1, Anhang J. Beton- und
Stahlbetonbau, Vol. 103, Issue 12, pp. 837-839.
DARTS (2004), Durable and reliable tunnel structures DATA. Project with financial support of
the European Commission under the Fifth Framework Program, GROWTH 2000 Project
GRD1-25633, Contract G1RD-CT-2000-00467.
DD CEN/TS 12390-11 (2010), Testing hardened concrete Part 11: Determination of the chloride
resistance of concrete, unidirectional diffusion.
DIN 1045-2 (2008): Tragwerke aus Beton, Stahlbeton und Spannbeton - Teil 2: Beton - Festlegung,
Eigenschaften, Herstellung und Konformitt - Anwendungsregeln zu DIN EN 206-1.
Ditlevsen, O., Madsen, H.O. (2005), Structural reliability methods. Coastal, maritime and structural
engineering, department of mechanical engineering, Technical University of Denmark, ISBN
0 471 96086 1.
DS 2426 (2011), Concrete materials rules for application of EN 206-1 in Denmark. Danish
Standards Association, Charlottenlund, Denmark.
DS/EN 1992-1-1 DK NA: 2011: National Annex to Eurocode 2: Design of concrete structures
Part 1-1: General rules and rules for buildings.
EHE-08 (2008), Code on structural concrete. Ministerio de Fomento, Spain.

EN 1990 (2002), Eurocode - basis of structural design.

EN 1992-1-1 (2002), Eurocode 2: design of concrete structures - Part 1: General rules and rules for
buildings.
EN 206-1 (2001), Concrete - Part 1: Specification, performance, production and conformity.
fib (2006), Model code for service life design. Bulletin of the fib. Lausanne, Switzerland.
Fluge, F., Helland, S., Maage, M., Smeplass, S. (2001), Marine chlorides a probabilistic approach
to derive durability related provisions for NS-EN 206-1. Report No. 19 / BP1 B4 of the R&D
project Betongkonstruksjoners liveslop, Oslo, Norway.
Gehlen, Ch. (2000), Probabilistische Lebensdauerbemessung von StahlbetonbauwerkenZuverlssigkeitsbetrachtung zur wirksamen Vermeidung von Bewehrungskorrosion.
Doctoral thesis, RWTH Aachen, Germany.
Gehlen, Ch., Schiel, P., Schiel-Pecka, A. (2008), Hintergrundinformationen zum Positionspapier
des DAfStb zur Umsetzung des Konzepts von leistungsbezogenen Entwurfsverfahren unter
Bercksichtigung von DIN EN 206-1, Anhang J, fr dauerhaftigkeitsrelevante
Problemstellungen. Beton- und Stahlbetonbau, Vol. 103, Issue 12, pp. 840-851.
Gehlen, Ch., Greve-Dierfeld, S. von (2013), Lebensdauer von Stahlbetonbauteilen - Empfehlungen
fr eine modifizierte deskriptive Bemessung. Band 1, Kapitel III. In: Konrad Bergmeister,
Frank Fingerloos, Johann-Dietrich Wrner (Hrsg.): Beton-Kalender 2013. Ernst & Sohn
GmbH & Co. KG, pp. 223-270.
Ghods, P., Chini, R., Alizadeh, M., Hoseini, M., Shekrachi, A.A., Ramezanianpour (2005): The
effect of different exposure conditions on the chloride diffusion into concrete in the Persian
Gulf region. ConMAT Conference. Vancouver, Canada.
Greve-Dierfeld, S. von, Gehlen, Ch. (2014), Performance based deemed-to-satisfy rules.
Proceeding of the fourth International fib congress 2014 - Improving performance of concrete
structures, February 2014, Mumbai, India.
Gruyaert, E., Van den Heede Ph., De Belie, N. (2009): Chloride ingress for concrete containing
blast-furnace slag related to microstructural parameters. RILEM Workshop on Concrete
durability and service life planning, Concrete life 09, Haifa, Israel.
Gulvanessian, H., Calgaro, J-A., Holicky, M. (2002), Designers guide to EN 1990 Eurocode: basis
of structural design. Thomas Telford Ltd., London, Great Britain, ISBN: 0 7277 3011 8.
Heinz, D., Urbonas, L., Gbel, M., Schubert, J. (2011), Praxisgerechte flugaschereiche Betone mit
Hochleistungsfliemittel. BetonWerk International, Vol. 51, pp. 36-45.
Helland, S. Aarstein, R., Magne, M. (2010), In-field performance of North Sea offshore platforms
with regard to chloride resistance. Structural Concrete, Vol. 11, Issue 2.
ISO 16204 (2012), Durability Service life design of concrete structures. ISO FDIS 16204:
updated 2012-08-08, ISO TC 71/SC 3.
JCSS (1999), JCSS probabilistic model code Part 3: Resistance models. Joint committee on
structural safety.
JCSS (2000), JCSS probabilistic model code Part 1: Basis of design. Joint Committee on Structural
Safety.
Lay, S., Schiel, P. (2002), Dauerhaftigkeitsbemessung von Stahlbetonkonstruktionen.
Forschungsbericht AiF/DBV-Nr. 12525/225, Technical University Muenchen, Germany.
Leivestad, S. (2013), Durability exposure resistance classes, a new system to specify durability in
EN 206 and EN 1992. Memo durability classes, JWG 250/104 N19C.
LNEC E 464 (2007), Betoes metodologia prescritiva para uma vida util de projecto de 50 e de 100
e face as accoes ambientais. Laboratorio naional de engenharia civil, Portugal.
LNEC E 465 (2007), Concrete - methodology for estimating the concrete performance properties
allowing to comply with the design working life of reinforced or pre-stressed concrete
structures under environmental exposures XC and XS. MOPTC - Laboratrio Nacional de
Engenharia Civil, Portugal.
Markeset, G., Skjlsvold, O. (2010), Time dependent chloride diffusion coefficient field studies
of concrete exposed to marine environment in Norway. 2nd International Symposium on
Service Life Design for Infrastructure, October 2010, Deft, The Netherlands.

Mller, Ch., Severins, K. (2009), Durability of concretes made with cements containing fly ash.
Concrete Technology Reports 2007-2009.
Mller, Ch., Severins, K., Hauer, B. (2009), New finding concerning the performance of cements
containing limestone, granulated blast-furnace slag and fly ash as main constituents.
Concrete Technology Reports 2007-2009.
NEN 6700 (2005), Technical principles for building structures. The Netherlands.
NEN 8005 (2008), Dutch supplement to NEN-EN 206-1: Concrete - part 1: specification,
performance, production and conformity. The Netherlands.
Nokken, M., Boddy, A., Hooton, R.D., Thomas, M.D.A. (2006), Time dependent diffusion in
concrete three laboratory studies. Cement and Concrete Research, Vol. 36, pp. 200-207.
NP EN 206-1 (2007), Beto Parte 1: Especificao, desempenho, produo e conformidade. IPQ,
Lisboa, Portugal.
NS EN 13670 (2010), Utfrelse av betongkonstruksjoner. Norway.
NS-EN 1992-1-1 (2008), National annex to EN 1992-1-1. Norway.
NS-EN 206-1 (2007), National annex to EN 206-1. Norway.
NT BUILD 443 (1995), Nordtest method - concrete, hardened: accelerated chloride penetration.
NT BUILD 492 (1999), Nordtest method - concrete, mortar and cement-based repair materials:
chloride migration coefficient from non-steady-state migration experiment.
Osterminski, K., Gehlen, Ch. (2009), Zuverlssigkeit Wasserbauwerke Chlorideindringwiderstand. Forschungsbericht 30-F-0019, TU Muenchen, Germany.
Polder, R.B., Wegen, G. van der, Breugel, K. van (2010), Guideline for service life design of
structural concrete with regard to chloride induced corrosion The approach in the
Netherlands. 2nd International Symposium on Service Life Design for Infrastructure,
October 2010, Deft, The Netherlands.
Sarja, A. (2005), Generic limit state design of structures. 10DBMC International Conference on
Durability of Building Materials and Components, 17-20 April 2005, Lyon, France.
Stanish, K., Thomas, M., (2003), The use of bulk diffusion tests to establish time-dependent
concrete. Cement and Concrete Research, Vol. 33, pp. 5562.
Tang, L., Utgenannt, P., Lindvall, A., Boubitsas, D. (2010), Validation of models and test methods
for assessment of durability of concrete structures in road environment. Uppdragsrapport No.
P802606, Lund, Sweden.
Tang, L. (1996), Chloride transport in concrete measurement and prediction. Ph.D. thesis,
Chalmers University of Technology, Gothenburg, Denmark.
Teply, B., Novak, D. (2012), Limit states of concrete structures subjected to environmental actions.
18th International Conference on Engineering Mechanics, 14-17 May 2012, Prague, Czech
Republic.
Visser, J.H.M., Nijland, T.G. (2009), Performance of tenary OPC-FGGBS-FA binder in meeting
conflicting concrete durability demands. IBAUSIL, Weimar, Germany.
Wall, H. (2007), Chloride profiling in marine concrete methods and tools for sampling. Ph.D.
thesis, Lund Institute of Technology, Sweden.
Wiens, U. (2005), Zur Wirkung von Steinkohleflugasche auf die chloridinduzierte Korrosion von
Stahl in Beton. Doctoral thesis, RWTH Aachen, Germany.
Zilch, K., Schiel, A. (2001), The determination of the nominal concrete cover by durability
design. Background document to Chapter 4 in EN 1992-1-1, February 2001.