Next Page
Go Back
Print
Prediction Of Chloride Penetration Into
Concrete Exposed To Various Exposure
Environments
L Tang1,2) & L-O Nilsson1)
1)
Chalmers University of Technology Göteborg, Sweden 2) SP Swedish National
Testing and Research Institute
Summary: This paper presents the results from a study of prediction model for chloride
penetration into concrete exposed to various exposure environments including alternative wet-anddry environment. A few years ago, a scientific model called ClinConc was developed from our
previous work. The model is essentially based on the current knowledge of physical and chemical
processes involved in the chloride transport and binding in concrete and has been verified by using
the field data from one to five years exposure under seawater. In this study, the model is further
developed for the application to alternative wet-and-dry environment, such as splash zone and road
environment.
The actual chloride profiles measured from the field exposure stations are used to verify the
modified model. The predicted results are in general fairly well in agreement with the field data,
especially the shapes of chloride profiles from alternative wet-and-dry environments. The
limitations and needs for further improvement of the latest version of the model are discussed.
Keywords. Chlorides, concrete, modelling, prediction.
1
INTRODUCTION
The numerical model ClinConc (Cl in Concrete) for prediction of chloride penetration into concrete was first presented in the
middle of 1990’s (Tang & Nilsson 1994; Tang 1995). The model consists of two main procedures:
1.
Simulation of free chloride penetration through the pore solution in concrete using a genuine flux equation
based on the principle of Fick’s law with the free chloride concentration as the driving potential, and
2.
Calculation of the distribution of the total chloride content in concrete using the mass balance equation
combined with non-linear chloride binding.
Not like other models, a unique character of the model ClinConc is that the chloride diffusivity, which can be determined by,
e.g. the Nordtest method NT BUILD 492 (Nordtest 1999), is considered as a material property. It changes only when concrete
is young, like many other material properties, such as porosity and strength. After an age of a half of year, this diffusivity
becomes more or less constant according to the experiments (Tang & Nilsson 1992; Tang 1996). Another unique character of
the model ClinConc is that the climatic parameters, such as chloride concentration and temperature, are used in both the flux
and the mass balance equations. Therefore, the model can well describe the effects of exposure conditions on chloride
penetration.
The original version of ClinConc was developed based on the field data up to two years exposure under seawater. Due to the
difficulties in combining moisture transport, the application of the original ClinConc was limited to submerged zone only.
When five-years field exposure data were available (Andersen et al 1998), it was found that the original ClinConc
underestimated the chloride content in the zone closer to the exposure surface, even though it predicted the penetration depth
fairly well. In other words, the surface chloride content tends to increase with exposure time even under submerged conditions.
This increased chloride content cannot be explained by drying-and-wetting effect, like in the splash zone. Time-dependent
chloride binding might be a potential reason, since the chloride binding isotherms used in the original ClinConc were those
obtained in the laboratory after about two weeks equilibrium (Tang & Nillson 1993). The effect of alkalinity on chloride
binding was also based on a limited investigation (Sandberg & Larsson 1993). In reality, the pore solution compositions may
change due to leaching and penetration of different substances, resulting in different characteristics of chloride binding.
Another possible reason is an increased saturation degree of the air voids near the surface. The saturation degree of the air
voids will increase after such a long period of immersion, especially in contact with a salt solution. It is difficult, however, to
9DBMC-2002 Paper 238 Page 1
Next Page
Go Back
model the saturation degree of the air voids. Therefore, the time-dependent chloride binding was assumed as a dominant reason
for the increased chloride contents in the surface zone (Tang & Nillson 2000). After this modification, the agreement between
modelled and measured chloride profiles becomes better (Tang & Nillson 2000b).
Very recently, the model ClinConc was modified again in order to make it applicable to various exposure
environments including alternative wet-and-dry ones. In fact, nothing except for the exposure conditions has been
modified in the latest modification. This paper present this latest modification and the verification of the model
using chloride profiles measured from the fields under various exposure environments.
2
MODELLING FOR VARIOUS EXPOSURE ENVIRONMENTS
2.1
Exposure Conditions for Marine Environment under Submerged Zone
Under submerged zone, concrete is constantly in contact with seawater. This might be the easiest case when compared with
other exposure environments. However, since both temperature and chloride concentration in seawater change with time, even
in this easiest case it is still difficult to use constant boundary conditions for chloride transport in concrete. In the previous
versions, both the exposure temperature and chloride concentration were assumed as a sine function. The sine function of
annual average temperature has been well verified, but not the chloride concentration in seawater. Therefore, it was suggested
in the latest version that average chloride concentration in seawater should be used unless the actual function of chloride
concentration is known. An example of the exposure conditions for submerged zone is shown in Fig. 1.
2.2
Exposure Conditions for Marine Environment above Seawater
In the marine environment above seawater, such as splash zone or atmospheric zone, concrete is subjected to alternative
wetting-and-drying. The wetting includes both salt water and rain. Owing to the complicated mechanisms involved in both the
moisture and chloride transport, it is not an easy task to combine both moisture and chloride transports into a single model,
even though some attempts have been done (Nilsson 2000; Francy et al 1996). On the other hand, it could be reasonable to
assume that, under such a wet-and-dry environment, the chloride concentration in contact with the concrete surface alters
between zero and a specified level. The wick effect due to drying is compensated by the effect of capillary suction due to rewetting. Therefore, the chloride transport could be assumed dominated by diffusion in a saturated pore system, despite of
wetting-and-drying processes. In this way, the difficulties in modelling of moisture transport could be skipped and the question
becomes how to
Chloride Concentration
Annual minimum
14
Temperature
gCl/l
Annual minimum
2
°C
Annual maximum
20
°C
Frequency
1
cycles/year
120
days of year
Annual maximum
14
gCl/l
Annual Cl period
12
months
(number)
First mean at
25
16
14
12
10
8
6
4
2
0
Temperature, °C
Cl concentration, g/l
Max. in the month
0 1 2 3 4 5 6 7 8 9 10 11 12
Months
20
15
10
5
0
0 1 2 3 4 5 6 7 8 9 10 11 12
Months
0
0
Figure 1. Example of exposure conditions for submerged zone (Swedish west coast).
define the chloride concentration curve. In the latest modification, a statistic normal distribution function was proposed to
describe the annual chloride concentration, that is,
 τ2
c0s = c 0 exp − 2
 2σ



(1)
where c0s is the chloride concentration in contact with the concrete surface, c0 is the average annual chloride concentration in
seawater, σ is the standard deviation that will be explained later, and τ is the time difference, which is a periodic function and
expressed as
9DBMC-2002 Paper 238 Page 2
Next Page
Go Back
 t − tm
−n

L


τ = f (t ) = f (t + nL ) = 

 t − t m − (n + 1)
 L
nL ≤ t − t m ≤
(2n + 1) L
2
(2)
(2n + 1) L < t − t
2
m
≤ (n + 1)L
n = 0, 1, 2, 3…
where t is the actual time, L is the period, tm is the time when the chloride concentration reaches maximum during the period,
and n is the integral of t/L.
The standard deviation σ is a decisive parameter to the width of a statistic normal distribution curve and can be expressed as
σ = στ
LCl
L
(3)
where LCl is the chloride duration during the period (LCl ≤ L), and στ (= 0.15) is the standard deviation of τ. It should be noticed
that the time difference τ is a dimensionless parameter, implying that t, L, tm, and LCl must have the consistent dimension,
which could be hours, days or months. Since the actual repetition of chloride concentration in splash zone is unknown, L was
simply assumed to be 12 months, that is, annually repeated in order to simplify the calculation. In this case the sine function of
temperature is inapplicable, thus an average annual temperature should be used. Some examples of exposure conditions for the
marine environment above seawater are given in Figs. 2 and 3.
Chloride Concentration
Annual minimum
0
Temperature
gCl/l
Annual minimum
11
°C
Annual maximum
11
°C
Frequency
1
cycles/year
14
gCl/l
Annual Cl period
12
months
Max. in the month
10
(number)
16
14
12
10
8
6
4
2
0
First mean at
10
8
6
4
2
0
0 1 2 3 4 5 6 7 8 9 10 11 12
0
0 1 2 3 4 5 6 7 8 9 10 11 12
Months
Months
Figure 2.
days of year
12
Temperature, °C
Cl concentration, g/l
Annual maximum
0
Example of exposure conditions for splash zone (0∼30 cm above seawater in Swedish west coast).
Chloride Concentration
Temperature
0
gCl/l
Annual minimum
11
°C
14
gCl/l
Annual maximum
11
°C
Annual Cl period
6
months
Frequency
1
cycles/year
Max. in the month
7
(number)
16
14
12
10
8
6
4
2
0
First mean at
days of year
12
Temperature, °C
Cl concentration, g/l
Annual minimum
Annual maximum
0 1 2 3 4 5 6 7 8 9 10 11 12
Months
10
8
6
4
2
0
0 1 2 3 4 5 6 7 8 9 10 11 12
Months
0
0
Figure 3. Example of exposure conditions for atmospheric zone (30∼60 cm above seawater in Swedish west coast).
9DBMC-2002 Paper 238 Page 3
Next Page
Go Back
2.3
Exposure Conditions for Road Environment Using De-Icing Salt
The same principles as described above can be used for the road environment. Thus the chloride concentration in contact with
concrete surface is
 τ2
c0s = c max exp − 2
 2σ



(4)
where cmax is the maximum chloride concentration during the period. The difference from the marine environment is that the
chloride period (application of de-icing salt) is a more or less known parameter under the road environment, for instance, from
November to March in the winter. Thus the sine function of temperature can be applied as in reality. However, the maximum
chloride concentration cmax in this case becomes unknown. From the field data obtained from the two-winters exposure along
the highway Rv 40 between Borås and Göteborg it was found that, when assuming a maximum concentration of 50 g Cl per
litre, the predicted profiles correspond fairly well with the field data, which will be presented later. An example of exposure
conditions for the road environment is shown in Fig. 4.
Chloride Concentration
Temperature
0
gCl/l
Annual minimum
-5
°C
Annual maximum
50
gCl/l
Annual maximum
18
°C
Annual Cl period
5
months
Frequency
1
cycles/year
Max. in the month
1
(number)
120
days of year
First mean at
60
20
50
15
40
Temperature, °C
Cl concentration, g/l
Annual minimum
30
20
10
0
10
5
0
-5
0
1
2
3
4
5
6
7
8
0
1
2
3
4
5
6
7
8
9 10 11 12
9 10 11 12
-10
Months
Months
Figure 4.Example of exposure conditions for a road environment (Highway Rv 40 between Borås & Göteborg, Sweden)
3
VERIFICATIONS OF THE LATEST MODIFIED MODEL
From two Swedish national projects, many chloride profiles obtained after five years exposure under the marine environment
and after two winters exposure under the road environment are available (Andersen et al 1998; Lindvall et al 2000). The
chloride profiles from two types of binder and two water-binder ratios, which are commonly used in Sweden for
infrastructures, were utilised to verify the latest modified ClinConc. The mixture proportions of concrete and relevant
properties are given in Table 1, and the common parameters used in the calculation are listed in Table 2. The exposure
conditions are as shown in Figs. 1 to 4. The results are shown in Figs. 5-8. Considering the very complicated mechanisms of
chloride transport in concrete, we can conclude that the predicted results are in general fairly well in agreement with the field
data, especially the shapes of chloride profiles from alternative wet-and-dry environments. This implies that the assumptions
made in the latest modification for various exposure environments are reasonable and close to the reality.
Table 1. Mixture proportions and diffusivity of concrete used for verification
Binder type
Waterbinder
ratio
Cement
content
Aggregate
Air
content
kg/m3
kg/m3
kg/m3
Diffusivity*
DCTH
m2/s
SRPC
0.40
420
1692
6.0
8.1×10-12
SRPC
0.50
370
1689
6.4
19.9×10-12
95%SRPC + 5%CSF
0.40
420
1685
5.9
2.7×10-12
95%SRPC + 5%CSF
0.50
370
1683
6.0
13.4×10-12
* Determined by the CTH method (NT BUILD 492) at an age of about 180 days.
9DBMC-2002 Paper 238 Page 4
Next Page
Go Back
Table 2. Common parameters used in the calculation
Chloride Binding
Variable Diffusivity
Isotherm slope
Activation energy
fb = 3.57
Non-linear exponent
B = 0.38
Activation energy
Eb = 40000 J/mol
Time-dependent factor
Eb = 40000 J/mol
Age dependent
βt = 0.152 (w/b)-0.6
t Da = 180 days
Depth dependent
None (steel form)
ft = 0.36ln(tCl + 0.5) + 1,
where tCl is the local chloride
contamination time in years.
4
LIMITATIONS AND NEEDS FOR FURTHER IMPROVEMENT
Although the verification results show a good agreement with the field data, there still exist the following limitations:
• Limited concrete type In the above verification, the concrete type is limited to two water-binder ratios (0.4 and 0.5) and
two types of binder (SRPC and 5% silica fume). Although these two types of concrete are very commonly used in
Sweden for infrastructures, more types of concrete, especially HPC with low water-binder ratios and different types of
binder, such as fly ash, slag, etc., should be used for verification.
• Limited exposure time So far the available data from the field exposure stations are limited to 5 years for marine
environment and 2 years for road environment. This exposure time is relatively short when compared with the whole
service life of concrete structures. More data from the long term exposure fields, especially with traceable exposure
environments, are needed for a better verification.
• Characterising exposure environment In the latest modifications of the model, the alternative wet-and-dry
environment is described using statistic normal distribution functions. The question is how to determine the
key parameters LCl – chloride duration for marine environment in equation (3) and cmax – maximum chloride
concentration for road environment in equation (4). Some simple methods are needed for characterising
different exposure environments.
9DBMC-2002 Paper 238 Page 5
Next Page
Go Back
6
Calc. 1 y
Calc. 2 y
Cl% w t of binder
60 cm above seawater
Calc. 5 y
4
Meas. 1 y
Meas. 2 y
Meas. 5 y
2
0
0
10
20
30
50
Depth, mm
6
Calc. 1 y
30 cm above seawater
Cl% w t of binder
40
Calc. 2 y
Calc. 5 y
4
Meas. 1 y
Meas. 2 y
Meas. 5 y
2
0
0
10
20
30
40
50
Depth, mm
6
Calc. 1 y
Cl% w t of binder
Under seawater
Calc. 2 y
Calc. 5 y
4
Meas. 1 y
Meas. 2 y
Meas. 5 y
2
0
0
10
20
6
Cl% w t of binder
30
40
50
Depth, mm
Calc. 20 months
Highway Rv40
4
Meas. 20 months
2
0
0
10
20
30
40
50
Depth, mm
Figure 5. Example of the predicted chloride profiles. SRPC w/b 0.40.
9DBMC-2002 Paper 238 Page 6
Next Page
Go Back
6
Calc. 1 y
Calc. 2 y
Cl% w t of binder
60 cm above seawater
Calc. 5 y
4
Meas. 1 y
Meas. 2 y
Meas. 5 y
2
0
0
10
20
30
40
50
Depth, mm
6
Calc. 1 y
Cl% w t of binder
30 cm above seawater
Calc. 2 y
Calc. 5 y
4
Meas. 1 y
Meas. 2 y
Meas. 5 y
2
0
0
10
20
30
40
50
Depth, mm
6
Calc. 1 y
Cl% w t of binder
Under seawater
Calc. 2 y
Calc. 5 y
4
Meas. 1 y
Meas. 2 y
Meas. 5 y
2
0
0
10
20
40
50
Depth, mm
6
Cl% w t of binder
30
Calc. 20 months
Highway Rv40
4
Meas. 20 months
2
0
0
10
20
30
40
50
Depth, mm
Figure 6. Example of the predicted chloride profiles. SRPC, w/b 0.50.
9DBMC-2002 Paper 238 Page 7
Next Page
Go Back
6
Calc. 1 y
Calc. 2 y
Cl% w t of binder
60 cm above seawater
Calc. 5 y
4
Meas. 1 y
Meas. 2 y
Meas. 5 y
2
0
0
10
20
30
50
Depth, mm
6
Calc. 1 y
30 cm above seawater
Cl% w t of binder
40
Calc. 2 y
Calc. 5 y
4
Meas. 1 y
Meas. 2 y
Meas. 5 y
2
0
0
10
20
30
40
50
Depth, mm
6
Calc. 1 y
Cl% w t of binder
Under seawater
Calc. 2 y
Calc. 5 y
4
Meas. 1 y
Meas. 2 y
Meas. 5 y
2
0
0
10
20
6
Cl% w t of binder
30
40
50
Depth, mm
Calc. 20 months
Highway Rv40
4
Meas. 20 months
2
0
0
10
20
30
40
50
Depth, mm
Figure 7. Example of the predicted chloride profiles. SRPC + 5%CSF, w/b 0.40.
9DBMC-2002 Paper 238 Page 8
Next Page
Go Back
6
Calc. 1 y
Calc. 2 y
Cl% w t of binder
60 cm above seawater
Calc. 5 y
4
Meas. 1 y
Meas. 2 y
Meas. 5 y
2
0
0
10
20
30
40
50
Depth, mm
6
Calc. 1 y
Cl% w t of binder
30 cm above seawater
Calc. 2 y
Calc. 5 y
4
Meas. 1 y
Meas. 2 y
Meas. 5 y
2
0
0
10
20
30
40
50
Depth, mm
6
Calc. 1 y
Cl% w t of binder
Under seawater
Calc. 2 y
Calc. 5 y
4
Meas. 1 y
Meas. 2 y
Meas. 5 y
2
0
0
10
20
6
Cl% w t of binder
30
40
50
Depth, mm
Calc. 20 months
Highway Rv40
4
Meas. 20 months
2
0
0
10
20
30
40
50
Depth, mm
Figure 8. Example of the predicted chloride profiles. SRPC + 5%CSF, w/b 0.50.
9DBMC-2002 Paper 238 Page 9
Go Back
5
CONCLUSIONS
The latest modification has made the model ClinConc applicable to both the marine environment, including submerged, splash
and atmospheric zones, and the road environment using de-icing salt in the winter. The verifications up to five-years marine
exposure data and two-winters road exposure data show that the predictions of chloride penetration into concrete structures are,
in general, fairly well in agreement with the measured chloride profiles.
The exposure environment can be described by the combination of temperature and concentration functions. The former can be
expressed as a sine function, while the latter expressed by a statistic normal distribution function. With such a combination, the
chloride ingress into concrete under various environments could be approximated by diffusion in a saturated pore system, thus
the actual wetting-and-drying processes could be skipped.
There is an urgent need to develop some simple methods for characterising different exposure environments.
6
1.
REFERENCES
Andersen, A., Hjelm, S., Janz, M., Johannesson, B., Pettersson, K., Sandberg, P., Sørensen, H., Tang, L. And Woltze, K.
1998, ‘Total chloride profiles in uncracked concrete exposed at Träslövsläge marine field station - Raw data from 1992 to
1997’, Report TVBM-7126, Division of Building Materials, Lund Institute of Technology, Lund, Sweden, 1998.
2.
Francy, O., Bonnet, S., Francois, R. & Perrin, B. 1996, ‘Modelling of chloride ingress into cement-based materials due to
capillary suction’, Proceedings of the 10th ICCC. 4iv078.
3.
Lindvall, A., Adersen, A. & Nilsson, L.-O. 2000, ‘Chloride ingress data from Danish and Swedish road bridges exposed to
splash from de-icing salts’. Proceedings of the 2nd International RILEM Workshop on Testing and Modelling the Chloride
Ingress into Concrete, Paris, 11-12 September 2000, pp. 85-103.
4.
Nilsson, L.-O. 2000, ‘A numerical model for combined diffusion and convection of chloride in non-saturated concrete’.
Proceedings of the 2nd International RILEM Workshop on Testing and Modelling the Chloride Ingress into Concrete,
Paris, 11-12 September 2000, pp. 261-275.
5.
Nordtest 1999, ‘Concrete, Mortar and Cement Based Repair Materials: Chloride Migration Coefficient from Non-steady
State Migration Experiments’, NT BUILD 492, Esbo, Finland.
6.
Sandberg, P. & Larsson, J. 1993, ‘Chloride binding in cement pastes in equilibrium with synthetic pore solutions as a
function of [Cl] and [OH]’, in Chloride Penetration into Concrete Structures - Nordic Miniseminar, ed. by L.-O. Nilsson,
Publication P-93:1, Division of Building Materials, Chalmers University of Technology, pp. 98-107, Gothenburg, Sweden.
7.
Tang, L. 1995, ‘A Windows program for the prediction of chloride penetration into submerged concrete’, Proceedings of
the RILEM International Workshop on Chloride Penetration into Concrete, Oct. 15-18, 1995, St. Rémy-lès-Chevreuse,
France, ed. by L.-O. Nilsson and J.P. Ollivier, pp. 206-215.
8.
Tang, L. 1996, ‘Electrically accelerated methods for determining chloride diffusivity in concrete’, Magazine of Concrete
Research, 48(176), 173-179.
9.
Tang L. & Andersen, A. 2000, ‘Chloride ingress data from five years field exposure in a Swedish marine environment’,
Proceedings of the 2nd International RILEM Workshop on Testing and Modelling the Chloride Ingress into Concrete,
Paris, 11-12 September 2000, pp. 105-119.
10. Tang, L. & Nilsson, L.-O. 1992, ‘Chloride diffusivity in high strength concrete at different ages’, Nordic Concrete
Research, Publication No. 11, pp. 162-171.
11. Tang, L. & Nilsson, L.-O. 1993, ‘Chloride binding capacity and binding isotherms of OPC pastes and mortars’, Cement
and Concrete Research, 23(2), 347-353.
12. Tang, L. & Nilsson, L.-O. 1994, ‘A numerical method for prediction of chloride penetration into concrete structures’, in
The Modelling of Microstructure and its Potential for Studying Transport Properties and Durability, ed. H. Jennings et al,
Kluwer Academic Publisher, pp. 539-552.
13. Tang, L. & Nilsson, L.-O. 2000, ‘Modeling of chloride penetration into concrete - Tracing five years field exposure’,
Concrete Science and Engineering, 2(8), 170-175.
14. Tang L. & Nilsson, L.-O. 2000b, ‘Current development and verification of the numerical model ClinConc for predicting
chloride penetration into concrete’, Proceedings of the 2nd International RILEM Workshop on Testing and Modelling the
Chloride Ingress into Concrete, Paris, 11-12 September 2000, pp. 305-316.
9DBMC-2002 Paper 238 Page 10