Corrosion Science 51 (2009) 22182230

Contents lists available at ScienceDirect

Corrosion Science
journal homepage: www.elsevier.com/locate/corsci

Throwing power of cathodic prevention applied by means of sacricial

anodes to partially submerged marine reinforced concrete piles:
Results of numerical simulations
Luca Bertolini, Elena Redaelli *
Politecnico di Milano, Dipartimento di Chimica, Materiali e Ingegneria Chimica G. Natta, via Mancinelli 7, 20131 Milan, Italy

a r t i c l e i n f o a b s t r a c t

Article history: The paper deals with the determination of current and potential distribution in reinforced concrete ele-
Received 27 January 2009 ments partially submerged in seawater aimed at predicting the throwing power of cathodic prevention
Accepted 7 June 2009 applied by means of sacricial anodes. Experimental results from previous laboratory tests showed that
Available online 14 June 2009
the throwing power of cathodic prevention is higher compared to that of cathodic protection [1]. In order
to extend the results obtained on small-scale specimens to elements of higher dimensions, FEM numer-
Keywords: ical simulations of potential distribution were carried out. Several cases were considered, representative
A. Steel reinforced concrete
of conditions differing in electrochemical behaviour of steel bars, geometry of the pile and of sacricial
B. Modelling studies
C. Cathodic protection
anodes, concrete resistivity. The results allowed to discuss the role of different factors on the throwing
C. Pitting corrosion power that can be reached by using sacricial anodes immersed in the seawater to protect reinforcing
steel bars in the emerged part of a pile.
2009 Elsevier Ltd. All rights reserved.

1. Introduction risk of overprotection, it can be safely applied to prestressed struc-

Chloride induced corrosion is one of the main deterioration
phenomena affecting the durability of reinforced concrete struc-
tures exposed to seawater. Chlorides penetrate the concrete cover
and cause the onset of pitting corrosion when their concentration
at the steel surface exceeds a critical threshold value. This thresh-
old depends on several factors, such as pH of concrete, steel poten-
tial and presence of microvoids in the cement paste [2,3].
Cathodic prevention is an electrochemical technique aimed at
delaying the onset of corrosion in reinforced concrete structures
subject to chloride penetration [2,46]. It is applied in the same
way as cathodic protection, i.e. applying a cathodic current to the
reinforcement through an external anode, but, unlike cathodic pro-
tection, it is applied when chlorides have not reached the critical
threshold at the reinforcement and the steel is still passive. Cathodic
prevention relies on the increase in the critical chloride threshold for
the initiation of corrosion due to a cathodic polarisation of the steel.
As a consequence, the time necessary for chlorides to penetrate the
concrete cover and reach the critical content at the steel surface is
increased and the service life of the structure is extended [5]. Catho-
dic prevention operates at much lower current densities (about
12 mA/m2 of steel surface) compared to cathodic protection (10
20 mA/m2). Owing to its higher throwing power and thus the lower
* Corresponding author. Tel.: +39 02 2399 3115; fax: +39 02 2399 3180.
E-mail address: elena.redaelli@polimi.it (E. Redaelli).
tures without any risk of hydrogen embrittlement [6].
Marine structures which are partially submerged in seawater
show some peculiarities. Here the higher risk of corrosion usually
occurs in the emerged part of the structure (in particular the tidal
and splash zones), where both oxygen and chlorides are available.
Conversely, the steel bars in the permanently immersed concrete
are usually free from corrosion, owing to the lack of oxygen and
consequent very negative potential values (typical of saturated
conditions) that promote high chloride threshold values for corro-
sion initiation.
The possibility of protecting steel bars in the emerged part of
these structures through the use of sacricial anodes (either
immersed in seawater or embedded in the structure) has been
investigated by several authors [1,710]. In particular, the possibil-
ity of protecting the reinforcement by means of sacricial anodes
which are simply submerged and electrically connected to steel
bars would be very attractive since it would allow to reduce the
costs related to the application. However, a possible limitation is
the throwing power of the protection, i.e. the height at which
protection can be achieved: it appears that this height is limited
to a few tens of centimetres above the water level, depending on
the concrete resistivity [7,8]. The use of cathodic prevention,
instead, can avoid this limitation owing to the higher polarisability
of passive steel compared to that of active steel [1,11].
Indeed, previous experimental tests carried out on reinforced
concrete columns 0.15  0.15  1.20 m showed that, if sacricial
anodes were applied when the concrete was free of chloride and

0010-938X/$ - see front matter 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.corsci.2009.06.012
 L. Bertolini, E. Redaelli / Corrosion Science 51 (2009) 22182230
steel bars were still passive, even the highest bar, placed at 1 m
from the level of water, was under polarisation conditions that pre-
vented corrosion initiation (as indicated by 4-h potential decay
values higher than 100 mV [6]). Conversely, if steel bars were
already corroding, polarisation conditions sufcient for protection
(4-h decay >100 mV) could be reached only up to about 0.60.8 m
from the water level. This result is shown in Fig. 1, that plots the
4-h decay of bars placed at different heights in the two cases of
cathodic protection and prevention.
In order to extend these results to more realistic elements of
higher dimension and different geometry, numerical simulations
of current and potential distribution were carried out on models
representative of reinforced concrete piles containing steel bars
at different heights, protected with sacricial anodes placed in
the water and connected to steel bars. Boundary conditions
describing the electrochemical behaviour of bars and values of con-
crete resistivity were obtained from the results of previously men-
tioned experimental tests.
This paper presents the results of the numerical simulations: rst
the numerical model was calibrated with the results obtained from
experimental tests presented in [1]; then the models were used to
study the effect of different parameters on the throwing power of
the techniques. Compared to existing literature on simulation of cur-
rent distribution in partially submerged concrete structures, this
work aimed at making a systematic comparison between the cases
of protection and prevention. Several cases were considered, repre-
sentative of conditions differing in geometry and dimension of ele-
ments, position and dimension of sacricial anodes, position of the
water level, use of a blended cement that increases concrete resistiv-
ity, with the aim of highlighting and quantifying the possible bene-
ts that can be achieved with cathodic prevention.
2. General description of the models
The numerical models were aimed at determining the distribu-
tion of potential and current in a system describing a concrete
element containing steel reinforcement partially immersed in sea-
water and protected with submerged sacricial anodes. In general,
this problem requires to solve Laplaces Eq. (1) together with
Ohms law (2):
Fig. 1. Comparison of experimental data of 4-h decay along the height of the
columns subjected to cathodic protection and cathodic prevention [1].
2219
r  jrV 0 1
and
i jrV 2
where j indicates the conductivity (X1 m1), V is the potential (V)
and i the current density (mA/m2) [12]. In the present work a nite
element method was used to make calculations of potential distri-
bution with a commercial software [13].
Owing to the higher electrical conductivity of metals compared
to that of concrete or seawater, the rebars and the anodes were as-
sumed to be equipotential regions, and were not considered in the
domain where Laplaces equation was solved; their electrochemi-
cal behaviour was introduced through proper boundary conditions
at the boundary representing the interface with the electrolyte.
Different boundary conditions were introduced. Polarisation
curves describing the kinetics of electrochemical reactions were
imposed at the steel/concrete interface, either in the form of But-
lerVolmer equations:
i i0  fexp2:303  V  V 0 =ba  exp2:303  V  V 0 =bc g
3
or by introducing the limiting current of oxygen diffusion:
1  exp2:303  V  V 0 =bc
i 4
1 exp2:303  V  V 0 =bc
i0 ilim
In the previous equations, i0 is the corrosion current density
(mA/m2), V0 the free corrosion potential (V), ilim the limiting cur-
rent density of oxygen (mA/m2), ba and bc the anodic and cathodic
slopes of the polarisation curves (mV/dec).
For rebars embedded in the submerged concrete, simple condi-
tions of constant current density were applied, taking into account
the limiting conditions due to lack of oxygen under both free
corrosion and cathodic polarisation:
i iconst 5
On the surface of the anodes, a condition of constant potential
was imposed (i.e. the anode was assumed to be non-polarisable):
V V an 6
All the other boundaries, such as border of the geometry or
rebars not connected to the anode, were characterised by insula-
tion conditions:
i0 7
The resistivity of the concrete was selected taking into account
differences in the water content and chloride content in the concrete
along the height of the column, as it will be explained later. The
resistivity of seawater was assumed equal to 0.2 X m in all cases.
3. Calibration of the model with experimental data
A rst series of simulations was carried out to calibrate the
model and to check its capability to describe the considered phe-
nomena. The same geometry of the laboratory specimens was con-
sidered and parameters measured on the specimens, such as V0, i0,
q, were introduced in the model. The specimens consisted of con-
crete columns 0.15  0.15  1.20 m containing 15 horizontal steel
bars at different heights and placed in a 3.5% NaCl solution, in a
way that only the lower rebar was permanently immersed. The
rebars were electrically connected to each other and with AlZn
In alloy discs (anodes) immersed in the solution. This simple geom-
etry allowed the measurement of the protection current picked-up
by each single bar. In one case the concrete was initially chloride-
free and steel bars were passive, thus representing conditions of
application of cathodic prevention; in the other case, chlorides
2220 L. Bertolini, E. Redaelli / Corrosion Science 51 (2009) 22182230

The geometry of the model is shown in Fig. 2. Assuming a trans-

Fig. 2. Geometry of the two-dimensional domain used in the numerical models
(dimensions in m).
were added to the concrete mix in an amount of 3% and the steel
bars were corroding when the anodes were connected, thus repre-
senting conditions of application of cathodic protection. Details of
the experimental set-up are described elsewhere [1].
lational symmetry of the specimens, the domain in which Laplaces
equation was solved was reduced to two dimensions. Similarly as
in the experimental tests, both the condition where all the 15 bars
were protected and the condition where only 8 alternate bars were
protected were considered. In the latter, only odd rebars (1, 3, 5. . .)
were connected to the anode, and the boundary of even rebars (2,
4, 6. . .) was considered as insulating (Eq. (7)).
The parameters used in the boundary conditions in the two
cases (i.e. concrete with chlorides representative of cathodic pro-
tection and chloride-free concrete representative of cathodic pre-
vention) are summarised in Table 1. On the rebar in the
permanently submerged concrete, a condition of constant current
density was chosen. This relies on the fact that this rebar received
a very small current from the anode, because of the low limiting
current of oxygen in saturated conditions. The current density
was imposed equal to the value absorbed by the immersed rebar
in the experimental tests, which was constant in time and equal
to 4 mA/m2 in the specimen with chlorides and 1 mA/m2 in the
specimen without chlorides.
The bars within a height h above the water level, equal to
0.24 m in chloride-bearing concrete (i.e. two bars) and 0.16 m in
chloride-free concrete (i.e. one bar), were characterised by polari-
sation curves with limiting current of oxygen, i.e. Eq. (4). This
choice stemmed from the high degree of water saturation that
characterised the concrete immediately above the water level, as
indicated by experimental values of concrete resistivity measured
on the laboratory specimens: concrete resistivity was similar to
that in immersed concrete over a height h. A higher value of h in
chloride-contaminated concrete should be ascribed to its higher
hygroscopy compared to chloride-free concrete. Values of limiting
current density, deriving from experimental tests, were 140 mA/
m2 in the specimen with chlorides and 40 mA/m2 in chloride-free
concrete.
The upper bars were described with ButlerVolmer expressions,
i.e. Eq. (3). The slope of the cathodic polarisation curve was chosen
according to the typical relationship found between the corrosion
potential and the corrosion rate of steel. It was assumed equal to
300 mV/dec; such a slope is in agreement with experimental
results reported in reference [14] and shown in Fig. 3 (where also
the results from experimental tests reported in [1] were plotted).
The slope of the anodic curve was assumed as 75 mV/dec for active
steel, and 10 V/dec (i.e. virtually innite) for passive steel. The val-
ues of corrosion potential V0 and corrosion current density i0 were
chosen according to measurements on steel in free corrosion con-
ditions using the tting line reported in Fig. 3: they were 200 mV/
SCE and 3 mA/m2 for steel in the specimen with 3% chlorides and
100 mV/SCE and 0.1 mA/m2 for steel in chloride-free concrete.
The potential imposed at the anode was 1.05 V/SCE.

Table 1
Boundary conditions and parameters used in the models describing the experimental tests.

Element Boundary condition Equation Parameters 3% chloride 0% chloride

(cathodic protection) (cathodic prevention)
Immersed bar Constant current density (5) iconst (mA/m2) 4 1
Bars within a height h from the sea level Limiting current of oxygen (4) h (m) 0.24 0.16
V0 (mV/SCE) 300 440
i0 (mA/m2) 10 8
ilim (mA/m2) 140 40
bc (mV/dec) 300 300
Upper bars ButlerVolmer (3) V0 (mV/SCE) 200 100
i0 (mA/m2) 3 0.1
ba (mV/dec) 75 10000
bc (mV/dec) 300 300
Anode Constant potential (6) Van (mV/SCE) 1050 1050
Others Insulation (7)
 L. Bertolini, E. Redaelli / Corrosion Science 51 (2009) 22182230
Fig. 3. Experimental data on free corrosion potential and corrosion rate of steel in
concrete in different conditions of exposure [14,1].
As already mentioned, measurements on the specimens had
shown that, in both cases, the concrete in the lower part of the
column was characterised by low values of resistivity, owing to
the wet conditions to which it was subjected, while in the upper
part concrete resistivity was much higher. To reproduce the
changes in the electrical resistivity of concrete as a function of
height, that were measured on each laboratory specimen due to
the differences in water content, different values of resistivity
were imposed in the different parts of the concrete domain in
the model. The concrete in permanently immersed conditions
and in the rst emerged part was assumed to have a low resistiv-
ity: 30 X m in the immersed part plus the rst emerged 0.24 m
for concrete with 3% chlorides and 100 X m in the immersed plus
the rst emerged 0.16 m for concrete without chlorides. The resis-
tivity in the upper remaining part of the pile was 200 and
1500 X m, respectively. The difference between the two cases ac-
counts for the different amount of chlorides in the concrete (3%
and none). Each model was solved with a mesh of 10,361 triangu-
lar elements.
The solutions of the numerical models were used to compare
the calculated distribution of potential and current density with
the experimental results. Fig. 4 shows, as an example, the distri-
Fig. 4. Comparison between distribution of potential obtained from experimental tests and from the numerical models for the specimen with 3% chlorides (a, cathodic
protection) and the specimen without chlorides (b, cathodic prevention). Also free corrosion potential values (V0) are indicated.
2221
butions of potential of rebars along the columns obtained with
the numerical simulation and compares them with the experi-
mental distributions (in the case shown in the gure only 8 alter-
nate bars were connected to the anode). It can be observed that
the results of the numerical model are in good agreement with
the experimental values both in the case of corroding steel in con-
crete contaminated with 3% chlorides by mass of cement (i.e.
cathodic protection, Fig. 4a) and in the case of passive steel in
chloride-free concrete (i.e. cathodic prevention, Fig. 4b). The po-
tential prole in Fig. 4a started from a value of 1.02 V/SCE for
the immersed bar (this value is very close to the potential of
the anode) and increased up to 0.22 V/SCE for the last bar, at
the top of the specimen. The coefcient of correlation between re-
sults of the model and experimental data was 0.973. Also the cur-
rent density obtained from the model showed a good agreement:
it was equal to 4 mA/m2 for the immersed bar (value imposed by
the boundary condition), 130.6 mA/m2 for rebar #3 and then pro-
gressively decreased to 1.9 mA/m2 for the rebar at the height of
1.12 m (#15); the coefcient of correlation was 0.984. In the case
of chloride-free concrete, Fig. 4b shows that the potential moved
from 1.03 V/SCE for the immersed rebar to 0.28 V/SCE for the
last bar in the emerged part; the coefcient of correlation was
equal to 0.960. The current density was 1 mA/m2 for the im-
mersed bar, 33.5 mA/m2 for rebar #3 and decreased to 0.3 mA/
m2 for the last bar (#15) (R2 of 0.995). A comparison of Fig. 4a
and b shows that the cathodic polarisation of the last rebar was
higher when steel was passive, although the current received
was lower than in the case of active steel (0.3 mA/m2 versus
1.9 mA/m2).
The throwing power was evaluated through the cathodic
polarisation, i.e. the difference between the potential of each re-
bar and its free corrosion potential V0 (also indicated in Fig. 4a
and b): this difference can be considered as an overestimation
of the 4-h or 24-h decay that is expected to be measured during
a depolarisation test, which is usually compared to the threshold
value of 100 mV to determine the effectiveness of cathodic pro-
tection. The solution of the model, similarly to the experimental
results, shows that the cathodic polarisation of steel, i.e. the low-
ering of the steel potential with respect to the free corrosion po-
tential, is higher than 100 mV up to a height of 0.65 m in the case
of cathodic protection and on the entire height of the column in
the case of cathodic prevention.
2222 L. Bertolini, E. Redaelli / Corrosion Science 51 (2009) 22182230

Therefore, it can be concluded that the numerical model is suit-

able to reproduce the current and potential distributions obtained
from experimental tests and, so, the advantages of cathodic pre-
vention compared to cathodic protection.

3.1. Analysis of sensitivity

An analysis of sensitivity was also carried out in order to high-

light the contribution of the different parameters in determining
the results obtained from the model: the corrosion current den-
sity of the rebars in the emerged part of the specimen, the slope
of the anodic and cathodic curves and the electrical resistivity of
the concrete were changed within different intervals of variation.
The values chosen are listed in Table 2. Fig. 5 shows, as an exam-
ple, the potential distributions obtained by varying the concrete
resistivity, in the case with and without chlorides. It can be seen
that increasing concrete resistivity from 1500 to 2000 X m, in the
chloride-free specimen, reduced the cathodic polarisation of the
Fig. 6. Height of protection (Hp) evaluated with 100-mV criterion as a function of
uppest rebar of 30 mV, while reducing concrete resistivity to
the concrete resistivity in the emerged part of the column (q), in the case of
500 X m increased the polarisation of 110 mV. Figs. 68 show cathodic protection and cathodic prevention.
the effect of the parameters considered in the analysis of sensitiv-
ity in terms of throwing power of the protection, i.e. the height at
which a cathodic polarisation of at least 100 mV could be
achieved. Arrows indicate that even the last rebar had a polarisa-
tion higher than 100 mV. These charts show that, as expected, an
increase in concrete resistivity and in corrosion current density
(both took place only in the emerged part of the pile) induced a
decrease in the height of protection. This can be attributed either
to ohmic effects or to the higher current density necessary to pro-
tect the steel. Conversely, an increase in the cathodic slope bc pro-

Table 2
Values of parameters considered in the analysis of sensitivity in the models describing
the experimental tests (only the indicated parameter was varied with respect to the
value of the initial case, indicated in brackets).

Parameters 3% chloride 0% chloride

(cathodic protection) (cathodic prevention)
i0 (mA/m2) (3) 1, 5 (0.1) 0.2, 0.3, 0.5
ba (mV/dec) (75) 50, 100, 125
bc (mV/dec) (300) 150, 200, 400 (300) 200, 250, 400 Fig. 7. Height of protection (Hp) evaluated with 100-mV criterion as a function of
qconcrete (X m) (200) 350, 400 (1500) 500, 2000, 5000 the free corrosion current density (i0), in the case of cathodic protection and
cathodic prevention.

Fig. 5. Comparison between distribution of potential obtained from the numerical models with different values of resistivity of concrete in the emerged part (values in X m)
for the specimen with 3% chlorides (a, cathodic protection) and the specimen without chlorides (b, cathodic prevention). Also experimental values are indicated.
 L. Bertolini, E. Redaelli / Corrosion Science 51 (2009) 22182230
Fig. 8. Height of protection (Hp) evaluated with 100-mV criterion as a function of
the cathodic slope (bc), in the case of cathodic protection and cathodic prevention.
Fig. 9. Ratio between the cathodic polarisation of the last bar (DElast) and the
cathodic polarisation of the last bar in the reference case (DElast,ref) as a function of
the ratio between parameter P and parameter P in the reference case (Pref). White
symbols refer to cathodic protection, black symbols to cathodic prevention.
Fig. 10. Comparison between distributions of potential obtained from the numerical models for the column 1.20 m high and for the column 2.40 m high, in the case with 3%
chlorides (a, cathodic protection) and without chlorides (b, cathodic prevention). Also free corrosion potential values (V0) are indicated.
2223
moted cathodic polarisation and so resulted in an increase of the
throwing power.
Fig. 9 summarises the effects of all the parameters considered
(including the anodic slope ba) in terms of the cathodic polarisation
achieved on the last bar (DElast), that allows to better quantify the
throwing power, independently of the criterion chosen to assess it.
The ratio DElast/DElast,ref was plotted as a function of the ratio P/Pref,
where P is the value of the parameter considered (the subscript
ref indicates the values in the reference case). Variations of the
anodic and cathodic polarisation slopes (ba and bc) produced vari-
ations of polarisation with a linear correlation, in the range consid-
ered here, with a higher sensitivity on bc compared to ba. The
corrosion current density i0 and the concrete resistivity (rho)
showed similar decreasing trends, i.e. an increase in one of these
parameters produced a decrease in the polarisation of the last
bar: the effect was higher in the case of cathodic protection com-
pared to the case of cathodic prevention.
Although potential distribution and, therefore, throwing power,
are mainly governed by concrete resistivity, and so are under oh-
mic control (as indicated by the decrease of the cathodic polarisa-
tion of the rebars increasing their distance from the anode), Fig. 9
shows that this effect is mitigated in the case of passive steel: for
instance, if the resistivity of concrete is doubled, from Fig. 9 it
can be estimated that the polarisation of the last bar is only 20%
of that in the reference case if steel is active, while it is about
70% if steel is passive. So, not only cathodic prevention has a higher
throwing power compared to cathodic protection (in spite of the
higher resistivity of chloride-free concrete, that would hinder the
throwing power), it is also less sensitive, in terms of reduction of
the throwing power, to an increase in concrete resistivity.
Although these quantitative comments apply only to the case con-
sidered here, they allow to compare the variations in the throwing
power that can be expected by varying some of the parameters
that characterise the system.
3.2. Model of specimens with double height (2.40 m)
Finally, also the case in which the column had a height of 2.4 m
instead of 1.2 m was considered, leaving unchanged all parameters
with respect to the initial simulation (obviously in the added part
of the pile all conditions were equal to those in the emerged part).
The results of potential distribution are shown in Fig. 10. In the
column with chlorides the potential of the immersed bar was
2224 L. Bertolini, E. Redaelli / Corrosion Science 51 (2009) 22182230
1.03 V/SCE; moving towards the top of the column the potential
increased up to a value of 0.2 V/SCE (i.e. the free corrosion poten-
tial) which characterised rebars from 1.28 to 2.24 m: these rebars
did not receive any appreciable current from the anode. Consider-
ing the criterion of a cathodic polarisation of 100 mV, rebars were
protected up to about 0.56 m from the level of the water. In the
column without chlorides the rebars were protected up to 1.2 m,
as for the shorter column. However, in this case the potential
had a value of 0.15 V/SCE at 2.24 m, so even on the last rebar
the cathodic polarisation (DElast) was 50 mV. Considering the cur-
rent density, the lowest rebars received roughly the same current
with respect to the column 1.20 m high, then the current
decreased: for instance, in the case without chlorides, the rebar
at 1.12 m received a current of 0.290 mA/m2 in the column
1.20 m high, and 0.164 mA/m2 in the column 2.40 m high. The total
current fed by the anode did not change (0.18 mA in this case), so
the current received by the bars at higher heights was caught from
those at lower heights, and this explains the lower polarisation of
bars at the same height.
4. Simulations of marine piles
In order to extend the results of the experimental tests to
geometries and dimensions of the columns more likely to occur
in reality, the numerical model was applied to a partially sub-
merged concrete pile containing reinforcing bars on its entire
height. A reference case was dened where the pile had a radius
R equal to 0.25 m, a total height of 3 m and the depth of the sea-
water in which it was permanently submerged (D) was equal to
1 m (so the height above the water level was 2 m). The pile is
sketched in Fig. 11, where details of geometric parameters are
also given. The pile contained eight vertical bars (12 mm in diam-
eter) and spiral reinforcement (10 mm in diameter); the spacing
of the spiral reinforcement p was 0.15 m. The concrete cover c
was equal to 30 mm. All the rebars were short-circuited as a con-
sequence of direct contact and they were electrically connected to
Fig. 11. Geometry of the marine pile in the reference case.
a sacricial anode, placed on the surface of the concrete. The an-
ode was also circular, with a height b equal to 0.2 m and a thick-
ness of 0.05 m, and was placed around the concrete pile 0.20 m
below the average level of seawater. The position of the anode a
was dened as the distance between the upper edge of the anode
and the bottom of the sea, and was equal to 0.8 m in the reference
case. The width of the domain representing seawater was equal to
1 m in all cases.
Initially, a three-dimensional model was considered. Numerical
simulations showed that the results of these 3-D models were
practically the same if the presence of the vertical bar was
neglected and only the spiral bar was considered. This allowed,
because of the axial-symmetric geometry, to reduce the problem
to two dimensions, and, therefore, to reduce the complexity of
the model and the computational time required to solve it and to
analyse the results [15].
In the two-dimensional model, boundary conditions followed
those described for the model of the experimental tests: the bars
in the immersed part of the pile had a constant current density;
bars in the rst emerged part with height h (0.24 m in chloride-
contaminated concrete and 0.16 m in chloride-free concrete) had
an expression of the polarisation curve limited by oxygen diffu-
sion (Eq. (4)), with changes of some parameters with respect to
values in Table 1 to account for the different heights from the
water level: in the pile with chlorides values of V0, i0 and ilim
were 420 mV/SCE, 20 and 160 m/m2, while in chloride-free
concrete they were 480 mV/SCE, 12.5 and 33 mA/m2. The
remaining upper bars were described with ButlerVolmer
expressions (Eq. (3)) with the same values of parameters as
those in Table 1. The potential imposed to the anode was
1.05 V/SCE (this condition was applied only on the boundaries
in contact with seawater, the boundary in contact with concrete
was assumed isolated).
Also concrete resistivity values were chosen in the same way as
for the previous models, i.e. values representative of saturated
concrete up to a height H + h and values representative of aerated
concrete for heights higher than H + h. This led to resistivity values
of 30 X m up to 1.24 m for concrete with chlorides and 100 X m up
to 1.16 m for concrete without chlorides; the resistivity in the
remaining part of the pile was 200 and 1500 X m, respectively.
These values can reasonably be considered representative of a calm
sea with low tides. This model was solved with a mesh of approx-
imately 15,500 triangular elements.
Figs. 12 and 13 show the potential and current distributions in
the pile. Both in the case of cathodic protection and in the case of
cathodic prevention, the potential of rebars in the immersed part
was close to 1.05 V/SCE, i.e. the potential of the anode. In the
emerged part of the pile the potential progressively increased: in
chloride-contaminated concrete a value of 0.2 V/SCE (i.e. equal
to V0) was reached at a height of 1.2 m, indicating that above this
height the cathodic polarisation was nil, while in chloride-free con-
crete the potential of the last bar was 0.16 V/SCE (lower than the
V0 value of 0.1 V/SCE).
Applying the 100-mV criterion to the difference between the
calculated potential and the free corrosion potential V01, if rebars
had an active behaviour, the cathodic polarisation was lower than
100 mV above 0.6 m from the level of seawater; if rebars were
passive, even the polarisation of the rebar at 1.2 m from the level
of seawater was higher than 100 mV, and it was 60 mV at the top
of the pile. The current that rebars received in chloride-free con-
crete was lower than the current that was received by the rebars
1
For the immersed rebars, that were characterised by a boundary condition of
constant current density, a free corrosion potential of 0.8 V/SCE was considered to
evaluate the cathodic polarisation.
 L. Bertolini, E. Redaelli / Corrosion Science 51 (2009) 22182230
Fig. 12. Distribution of potential obtained from the numerical models in the reference case for the pile with chlorides (a, cathodic protection) and without chlorides (b,
cathodic prevention). Also free corrosion potential values (V0) are indicated.
Fig. 13. Comparison of distributions of current density obtained from the numer-
ical models in the reference case for cathodic protection and cathodic prevention.
at the same height in chloride-contaminated concrete (Fig. 13). So,
in spite of the higher resistivity of chloride-free concrete, the
throwing power of cathodic prevention was higher than that of
cathodic protection, owing to the improved polarisability of pas-
sive steel. While the low polarisation of the upper rebars was
not enough to protect them if they were already corroding, it
could be benecial if they were passive, e.g. in increasing the chlo-
ride threshold for pitting corrosion initiation; the results obtained
from the model suggest that this effect can be present even at
heights above 2 m from the average level of the seawater (where
the cathodic polarisation was 60 mV). Fig. 13 also shows that the
total current received by the passive rebars is lower, and so is the
current fed by the anode: the total current erogated by the anode,
calculated from the integration of the current density on the sur-
face of the anode, in fact, was 1.65 mA in the case of passive steel
and 8.22 mA in the case of corroding steel. This means that the
consumption of the anode due to the current fed to the rebars
was about 24 g/year for cathodic protection and only 5 g/year
for cathodic prevention. Therefore, saving on the anode and reduc-
ing the amount of metals dissolved in seawater are further bene-
2225
cial effects of using cathodic prevention instead of cathodic
protection.
4.1. Variations with respect to the reference case
After the reference case, other scenarios were considered, by
varying parameters such as the concrete cover thickness, the
radius of the pile, the distance between rebars (and so the num-
ber of rebars), the position of anode in seawater. All the varia-
tions with respect to the reference case2 are shown in Table 3.
Table 4 summarises the main results obtained from the simula-
tions in terms of height of protection (Hp), cathodic polarisation
of the last rebar on top of the pile (DElast), cathodic current den-
sity received by the last rebar (ilast) and total current fed by the
anode (Ian).
Changing the concrete cover thickness from 30 mm to 40 and
100 mm did not produce signicant variations in current and po-
tential distributions, as it can be seen from data in Table 4: the
height of protection remained equal to 1.25 m from the level of
the sea for passive rebars and 0.62 m for active rebars. The current
density fed by the anode slightly decreased with increasing con-
crete cover.
On the contrary, the pile dimension had a signicant inu-
ence on the results of the model: the pile radius was increased
to 0.375, 0.425, 0.5 and 0.75 m, and this resulted, in both cases
of cathodic protection and cathodic prevention, in a higher
height of protection. A larger radius allowed the current to ow
more easily in the internal part of the pile. Fig. 14 shows
changes in the current distribution, while Fig. 15 plots the height
of protection evaluated with the 100-mV criterion as a function
of the pile radius. The height of protection increased with pile
radius, with a remarkably higher slope in the case of cathodic
prevention.
The dimension of the anode b showed no effect on potential dis-
tribution; the total current fed by the anode remained the same
after doubling the anode height b, so, being the anode surface high-
er, the current density decreased and the duration of the anode
increased of approximately a factor 2 (Table 4). Also the position
2
The geometry of the reference case was designed according to requirements of
Eurocode 2. Variations introduced later may have led to geometries that were not
consistent with design rules. They were merely intended to discuss electrochemical
effects.
2226 L. Bertolini, E. Redaelli / Corrosion Science 51 (2009) 22182230

Table 3
Variations of parameters with respect to the reference case (depicted in Fig. 11) in the models of marine piles (variations are with respect to the reference case whose parameters
are indicated in brackets).

Parameters Cathodic protection Cathodic prevention

Concrete cover c (mm) (30) 40, 100 (30) 40, 100
Pile radius R (m) (0.25) 0.375, 0.5, 0.75 (0.25) 0.375, 0.425, 0.5, 0.75
Anode dimension b (m) (0.2) 0.4 (0.2) 0.4
Anode position a (m) (0.8) 0.6, 0.4 (0.8) 0.6, 0.4
Depth of seawater D (m) (1) 0.5, 1.5 (1) 0.5, 1.5
Depth of seawater D (m) and Height of saturated concrete H + h (m) (1) 1.5 and (1 + 0.24) 1.5 + 0.24 (1) 1.5 and (1 + 0.16) 1.5 + 0.16
Pace of reinforcement p (m) (0.15) 0.07, 0.09, 0.12 (0.15) 0.07, 0.09, 0.12
Concrete resistivity q (X m): immersed splash and zone atmospheric zone (30) 180 and (200) 1000 (100) 600 and (1500) 5000
Presence of anode (Yes) No (Yes) No

Table 4
Summary of results obtained from the numerical simulations for the different scenarios considered: Hp, height of protection from the water level; DElast, cathodic polarisation of
the rebar on top of the pile; ilast, cathodic current density received by the rebar on top of the pile; and Ian, total current fed by the anode.

Cathodic protection Cathodic prevention

Case Hp (m) DElast (mV) ilast (mA/m2) Ian (mA) Hp (m) DElast (mV) ilast (mA/m2) Ian (mA)
Reference 0.575 0 0.03 8.22 1.175 59 0.06 1.65
c = 40 mm 0.575 0 0.03 7.87 1.175 62 0.06 1.56
c = 100 mm 0.575 0 0.04 7.15 1.175 64 0.06 1.44
R = 0.375 m 0.575 1 0.14 13.63 1.475 88 0.10 2.63
R = 0.425 m 1.775 99 0.11 3.06
R = 0.5 m 0.725 4 0.39 19.72 >1.925 114 0.14 3.72
R = 0.75 m 0.875 12 1.24 32.86 >1.925 156 0.23 6.01
b = 0.4 m 0.575 0 0.03 7.96 1.175 59 0.06 1.59
a = 0.6 m 0.575 0 0.03 8.07 1.175 59 0.06 1.64
a = 0.4 m 0.575 0 0.03 8.16 1.175 59 0.06 1.63
D = 0.5 m (low tide) 0.575 + 0.5 0 0.02 5.50 1.175 + 0.5 56 0.06 1.25
D = 1.5 m (high tide) 0.875  0.5 0 0.1 19.06 1.625  0.5 86 0.1 2.83
D = 1.5 m and H = 1.5 m 1.025  0.5 3 0.33 9.22 >1.925  0.5 110 0.13 1.88
p = 0.12 m 0.5 0 0.01 11.62 0.98 44 0.04 2.63
p = 0.09 m 0.41 0 0 12.31 0.86 30 0.03 2.68
p = 0.07 m 0.395 0 0 13.48 0.815 20 0.02 2.90
p = 0.12 m and R = 0.5 m 0.62 2 0.21 27.64 1.58 93 0.11 5.81
p = 0.09 m and R = 0.5 m 0.59 1 0.1 30.56 1.31 72 0.07 6.16
p = 0.07 m and R = 0.5 m 0.535 0 0.05 33.71 1.165 56 0.05 6.78
Blended cement 0.275 0 0 3.14 0.575 9 0.01 0.81
No anode 0.275 0 0 0.875 41 0.04

of the anode a did not show any effect: for a given depth of sea
water, the position of the anode had no inuence on the current
and potential distributions.
Variations of the water level were also considered in order to
simulate conditions of high or low tides (Fig. 16). When the
depth of the water was increased to 1.5 m (Fig. 16a), leaving
the same boundary conditions on the rebars and the same con-
crete resistivity as in the reference case, the current density re-
ceived by the rebars increased with respect to the reference case,
and so did the cathodic polarisation, in particular in the case of
cathodic prevention (Figs. 17 and 18). Rebars up to a height of
1.62 m (i.e. about 1.12 m from the level of the sea) were
protected in the case of cathodic prevention, while only up to
0.9 m (i.e. about 0.4 m from the water level) in the case of
cathodic protection. When the depth of the water was decreased
to 0.5 m (Fig. 16b), leaving the anode above the water level3, as
in the case of a low tide, the cathodic current density received by
3
In this case the surface of the anode that erogated the current was the one in
contact with the concrete surface and the boundary condition of rebars in water-
saturated concrete were characterised by Eq. (4) with V0 = 0.7 V/SCE in both cases,
and i0 = ilim = 4 mA/m2 in chloride-contaminated concrete and i0 = ilim = 1 mA/m2 in
chloride-free concrete.
the rebars slightly decreased at all heights and so the cathodic
polarisation. The rebars were protected up to 1.2 m in chloride-
free concrete and up to 0.6 m in chloride-contaminated concrete,
however, considering the height from the water level the height
of protection was 1.7 m and 1.1 m, respectively. The results of
these models indicated that if the water level decreases
temporarily and the concrete previously in contact with seawater
remains water-saturated, the anode can still supply the same
current as when it was immersed, provided that there is good
electrolytic contact between anode surface and concrete surface
to support current ow. Finally, also the case of depth of seawater
of 1.5 m with a height of saturated concrete of 1.5 m (+0.24 or
0.16 m as usual) was considered (Fig. 16c). The results of potential
distribution are presented in Fig. 19 and compared with the case
of high tide (i.e. with height of saturated concrete equal to 1 m).
This case may be considered representative of a high tide that
lasts enough to induce saturation of concrete. In this case it was
chosen to apply a condition of constant current density (Eq. (5))
on the immersed rebars. It can be observed that the increase in
the water level together with the saturation of immersed con-
crete brings about a considerable increase in the height of protec-
tion (in agreement with the results obtained varying the anode
position, that showed that the height of protection is not
affected by the position of the anode below the water level). In
 L. Bertolini, E. Redaelli / Corrosion Science 51 (2009) 22182230
Fig. 14. Comparison between distributions of current density obtained from the numerical models with different values of radius R for the pile with chlorides (a, cathodic
protection) and without chlorides (b, cathodic prevention).
Fig. 15. Height of protection evaluated with 100-mV criterion as a function of the
pile radius (R), in the case of cathodic protection and cathodic prevention.
this case the height of protection was about 0.5 m above the
water level in the case of cathodic protection and was higher than
1.5 m (i.e. reached the top of the pile) in the case of cathodic
prevention.
The spacing of reinforcement was changed from 0.15 m to 0.12,
0.09 and 0.07 m (the number of rebars changed from 20 in the ref-
erence case to 25, 33 and 42, respectively). A decrease in the dis-
tance between the rebars resulted in an increase in the steel area
to be protected and induced a decrease in the current density
received by the rebars from the anode, as shown in Fig. 20 that
plots the height of protection as function of rebars spacing, both
for R = 0.25 m and R = 0.5 m.
The use of a blended cement was also considered, instead of
an ordinary portland cement: it was assumed that the effect of
the blended cement from the point of view of the potential dis-
tribution was only an increase in the concrete resistivity (possi-
ble effects on the electrochemical behaviour in terms of
polarisation curves were neglected): the effect of the increased
resistivity resulted in a decrease in the height of protection, both
in cathodic protection and in cathodic prevention (Table 4), as it
2227
had already been shown in the analysis of sensitivity previously
presented (Section 3.1). It should be noted that this result only
applies to potential distribution and does not consider possible
advantages connected with the use of blended cements, such
as a delay in chloride penetration and so in the time for corro-
sion initiation.
It can be observed that, in all the cases considered, the cathodic
polarisation was always high (i.e. higher than 100 mV) on the re-
bars in the rst emerged part of the pile (immediately above the
water level) and this is benecial either in preventing the onset
of corrosion or in protecting the steel from corrosion in the tidal
and splash zones that are more affected by the ingress of chlorides
and oxygen.
All the results previously discussed clearly show that, although
many factors inuence the throwing power of the protection,
applying sacricial anodes before corrosion has initiated on the
steel bars (namely of applying cathodic prevention instead of
cathodic protection) is always more advantageous than applying
them after corrosion initiation. This is an example of the cheaper
costs and higher long-term benets of a prevention technique
instead of a repair technique.
Nevertheless, even the application of sacricial anodes some-
times after corrosion has initiated may have a benecial effect.
In fact, numerical analyses were also carried out to study the case
where no anode was applied. Fig. 21 reports potential distribu-
tions. In this case, the boundary conditions of the immersed bars
were characterised by a curve with limiting current (Eq. (4)) in or-
der to consider the low amount of oxygen in the immersed con-
crete and to allow for a possible anodic behaviour of the bars
(applying a constant cathodic current would obviously avoid such
a possibility). It can be observed that the rebars in the emerged
part of the pile were still cathodically polarised, although to a less
extent if compared to the case with anode, while the immersed
rebars were anodically polarised. So, due to the different electro-
chemical conditions of the bars placed at different heights, a mac-
rocell generates inside the pile: emerged rebars behave as
cathode and immersed rebars as anode. This macrocell therefore
stimulates corrosion even in bars in contact with the saturated
concrete under the water level, where availability of oxygen is
low.
Therefore, an electrochemical protection would be benecial
also for the rebars in the permanently submerged concrete, hin-
dering this macrocell. If cathodic protection is applied in the
2228 L. Bertolini, E. Redaelli / Corrosion Science 51 (2009) 22182230

Fig. 16. Geometry of the marine pile after variations of the depth of seawater D (a and b) and of the depth of the seawater D together with the height of the saturated concrete
H (c).

early stages of corrosion, i.e. when only the bars within a
few tens of centimetres above the sea level have been depassi-
vated, then it may be effective in reducing the effects of corro-
sion both in the emerged and submerged bars. This suggests
that, even in the case where corrosion has already initiated,
the earlier the sacricial anodes are applied, the greater their
effectiveness is.
5. Conclusions
A numerical model was developed that was able to simulate
experimental results on reinforced concrete specimens subjected
to cathodic protection and prevention, in order to investigate
the throwing power of protection with immersed sacricial
anodes.
 L. Bertolini, E. Redaelli / Corrosion Science 51 (2009) 22182230 2229

Fig. 17. Comparison between distributions of potential obtained from the numerical models with different values of depth of seawater D for the pile with chlorides
(a, cathodic protection) and without chlorides (b, cathodic prevention).

Fig. 18. Comparison between distributions of current density obtained from the numerical models with different values of depth of seawater D for the pile with chlorides
(a, cathodic protection) and without chlorides (b, cathodic prevention).

Fig. 19. Comparison between distributions of potential obtained from the numerical models with depth of seawater D = 1.5 m and different values of height of saturated
concrete H for the pile with chlorides (a, cathodic protection) and without chlorides (b, cathodic prevention).
2230 L. Bertolini, E. Redaelli / Corrosion Science 51 (2009) 22182230
Fig. 20. Variation of height of protection (Hp) evaluated with 100-mV criterion as a
function of the spacing of reinforcement p (white symbols refer to cathodic
protection, black symbols refer to cathodic prevention).
The application of this model showed that the steel reinforce-
ment of a circular pile, 3 m high and with a diameter of 0.5 m,
was polarised more than 100 mV with respect to the free corrosion
potential even at 1.2 m from the level of the water if the steel was
passive, while polarisation almost vanished above 0.6 m from the
level of the water if the steel was corroding, in spite of the lower
resistivity of the concrete due to the presence of chlorides.
The throwing power of protection was inuenced by the geom-
etry of the pile, in particular its diameter, and by the amount of
steel surface to be protected; other parameters such as concrete
cover, position and dimension of submerged anodes showed no ef-
fects in the ranges of variation considered.
In the absence of the anode, a macrocell generates between im-
mersed and emerged rebars, which stimulates corrosion even in
water-saturated concrete. The application of cathodic protection
with submerged sacricial anodes inhibits this effect.
Fig. 21. Comparison between distributions of potential obtained from the numerical models with and without the anode for the pile with chlorides (a, cathodic protection)
and without chlorides (b, cathodic prevention). Also free corrosion potential values (V0) are indicated.
References
[1] L. Bertolini, M. Gastaldi, M.P. Pedeferri, E. Redaelli, Prevention of steel
corrosion in concrete exposed to seawater with submerged sacricial
anodes, Corros. Sci. 44 (2002) 14971513.
[2] L. Bertolini, B. Elsener, P. Pedeferri, R. Polder, Corrosion of Steel in Concrete.
Prevention, Diagnosis, Repair, Wiley-VCH, Weinheim, 2004.
[3] C.L. Page, Corrosion and protection of reinforcing steel in concrete, in: C.L.
Page, M.M. Page (Eds.), Durability of Concrete and Cement Composites,
Woodhead Publishing Ltd., Cambridge, 2007, pp. 156186.
[4] P. Pedeferri, Cathodic protection and cathodic prevention, Constr. Build. Mater.
10 (1996) 391402.
[5] L. Bertolini, F. Bolzoni, M. Gastaldi, T. Pastore, P. Pedeferri, E. Redaelli, Effects of
cathodic prevention on the chloride threshold for steel corrosion in concrete,
Electrochim. Acta 54 (2009) 14521463.
[6] EN 12696-1 Standard, Cathodic Protection of Steel in Atmospherically Exposed
Concrete, March 2000.
[7] O.T. de Rincn, M.F. de Romero, A.R. de Carruyo, M. Sanchez, J. Bravo,
Performance of sacricial anodes to protect the splash zone of concrete piles,
Mater. Struct. 30 (1997) 556560.
[8] A.A. Sags, S.C. Kranc, F. Presuel-Moreno, Advanced computational model for
sacricial cathodic protection of partially submerged reinforced concrete
marine footers, in: W.F. Silva-Araya, O.T. de Rincn, L.P. ONeill (Eds.), Repair
and Rehabilitation of Reinforced Concrete Structures: the State-of-the-Art,
American Society of Civil Engineering, Reston, 1997, pp. 113.
[9] F.J. Presuel-Moreno, S.C. Kranc, A.A. Sags, Cathodic prevention distribution in
partially submerged reinforced concrete, Corrosion 61 (2005) 548558.
[10] F.J. Presuel-Moreno, A.A. Sags, S.C. Kranc, Steel activation in concrete
following interruption of long-term cathodic polarization, Corrosion 61 (2005)
428436.
[11] T. Pastore, P. Pedeferri, L. Bertolini, F. Bolzoni, Current distribution problems in
the cathodic protection of reinforced concrete structures, in: D.W.S. Ho, F.
Collins (Eds.), Proceedings of the International RILEM/CSIRO/ACRA Conference
on Rehabilitation of Concrete Structures, Melbourne, 31 August2 September,
1992, pp. 189200.
[12] J.S. Newman, Electrochemical Systems, Prentice Hall, Englewood Cliffs, NJ,
1973.
[13] www.comsol.com.
[14] L. Bertolini, R. Polder, Concrete resistivity and reinforcement corrosion
rate as a function of temperature and humidity of the environment, TNO
Building and Construction Research, Report No. 97-BT-R0574, Delft (NL),
1997.
[15] L. Bertolini, E. Redaelli, Numerical simulation of the throwing power of
cathodic prevention applied to marine reinforced concrete piles by means of
sacricial anodes, Proceedings of EUROCORR 2004, Nice (F), 1216 September
2004, pp. 110.