ANAIS DO 55 CONGRESSO BRASILEIRO DO CONCRETO - CBC2013 55CBC

1

Life-Cycle Reliability Assessment of Concrete Bridges
Exposed to Corrosion

Fabio Biondini

Department of Civil and Environmental Engineering, Politecnico di Milano
Piazza Leonardo da Vinci 32 20133 Milan, Italy
fabio.biondini@polimi.it



Abstract

For concrete bridges exposed to damaging environments the structural performance must be considered as
time-dependent. Therefore, a life-cycle approach to design of concrete bridges should lead to structures
which are able to comply with the desired performance not only at the initial stage when the system is intact,
but also during the expected lifetime. At present, design for durability with respect to chemical-physical
damage phenomena is based on threshold values for concrete cover, water-cement ratio, amount and type
of cement, among other prescriptive requirements, to limit the effects of structural damage induced by
carbonation of concrete and corrosion of reinforcement. However, a durable design cannot be based only on
such indirect evaluations of the effects of structural damage, but also needs to consider the global effects of
the local damage phenomena on the overall performance of the structure. To this aim, a life-cycle
probabilistic approach to structural assessment and design of concrete structures exposed to the diffusive
attack from external aggressive agents has been proposed in previous works. This approach allows to
reproduce the diffusion process of aggressive agents, such as chlorides, and to describe the mechanical
damage coupled to diffusion, including corrosion of steel reinforcement and deterioration of concrete. The
global effects of local damage are evaluated by means of nonlinear and limit analysis procedures and the
time-variant structural reliability is evaluated at the system level by Monte Carlo simulation. An overview of
the proposed methodology is presented in this paper, with application to time-variant reliability analysis and
lifetime assessment of a concrete arch bridge exposed to corrosion.

Keywords: Concrete Bridges; Diffusion Processes; Corrosion; Life-Cycle Reliability; Structural Lifetime.

ANAIS DO 55 CONGRESSO BRASILEIRO DO CONCRETO - CBC2013 55CBC

2
1 Introduction
The structural performance of concrete bridges is time-variant due to deterioration effects of
aging and damage processes of materials and components (Ellingwood 2005). The main
sources of damage include chemical processes associated to sulfate and chloride attacks
and alkali-silica reactions, physical processes due to freeze/thaw cycles and thermal cycles,
and mechanical processes such as cracking, abrasion, erosion, and fatigue (Kilareski
1980, CEB 1992, Bertolini et al. 2004). The detrimental effects of these phenomena can
lead over time to unsatisfactory levels of structural reliability. This problem presents a
major challenge to bridge engineering, since the classical time-invariant structural design
criteria and methods need to be revised to account for a proper modeling of the structural
system over its entire life-cycle by taking the effects of deterioration processes under
uncertainty into account (Frangopol & Ellingwood 2010, Frangopol 2011).
In recent years, a considerable amount of research has been conducted in the area of life-
cycle performance of structural systems under uncertainty (Frangopol & Furuta 2001,
Frangopol et al. 2004a, 2004b, 2007, 2012, Estes & Frangopol 2005, Nowak & Frangopol
2005, Cho et al. 2007, Biondini & Frangopol 2008a, Biondini 2009, Chen et al. 2010,
Frangopol & Ellingwood 2010, Biondini & Frangopol 2011, Strauss et al. 2012), and
relevant advances have been accomplished in the fields of modeling, analysis, design,
maintenance, monitoring, and management of deteriorating bridges (Frangopol et al.
1997a, 1997b, Frangopol 1999, Frangopol et al. 2002, Casas et al. 2002, Watanabe et al.
2004, Cruz et al. 2006, Furuta et al. 2006, Kho & Frangopol 2008, Frangopol et al. 2010,
Biondini & Frangopol 2012, Malerba 2013, Zhu & Frangopol 2013). Despite this research
trend, life-cycle concepts are not yet explicitly addressed in design codes and the checking
of system performance requirements is referred to the initial time of construction when the
system is intact. In this approach, design for durability of concrete structures with respect
to chemical-physical damage phenomena is based on simplified criteria associated with
classes of environmental conditions. Such criteria introduce threshold values for concrete
cover, water-cement ratio, amount and type of cement, among others, to limit the effects of
local damage due to carbonation of concrete and corrosion of reinforcement. However, a
durable design cannot be based only on such indirect evaluations of the effects of
structural damage, but also needs to take into account the global effects of the local
damage phenomena on the overall performance of the structure.
To this purpose, a life-cycle probabilistic approach to structural assessment and design of
concrete structures exposed to the diffusive attack from external aggressive agents has
been proposed in previous works (Biondini et al. 2004b, 2006a, 2006b, 2008, 2011, 2013,
Biondini & Frangopol 2008b, 2009, Biondini 2011, Biondini & Vergani 2012). This approach
allows to reproduce the diffusion process of aggressive agents, such as chlorides, and to
describe the mechanical damage coupled to diffusion, including corrosion of reinforcement
and deterioration of concrete. The global effects of local damage are evaluated by means
of nonlinear and limit analysis procedures and the time-variant structural reliability is
evaluated at the system level by Monte Carlo simulation. An overview of the proposed
methodology is presented in this paper, with application to time-variant reliability analysis
and lifetime assessment of a concrete arch bridge exposed to corrosion.

ANAIS DO 55 CONGRESSO BRASILEIRO DO CONCRETO - CBC2013 55CBC

3
2 Simulation of Diffusion Processes

The diffusion process of chemical components in solids can be described by the Ficks laws
which, in case of single component diffusion in homogeneous, isotropic and time-invariant
media, lead to the following second order partial differential linear equation (Glicksman 2000):


t
C
C D

=
2
(1)

where D is the diffusivity coefficient of the medium, C=C(z, t) is the concentration of the
chemical component at point z=(x,y,z) and time t, C = grad C(z, t) and
2
=.

For one-dimensional diffusion (1D) the Ficks equation can be solved analytically. The 1D
diffusion model is frequently used to simulate the chloride diffusion process in concrete
structures (fib 2006). However, the actual diffusion process is generally characterized by
two- or three-dimensional patterns of concentration gradients and 1D diffusion models can
lead to a loss of accuracy depending on the exposure conditions, geometrical shape ratio
of the cross-section, and location of points where concentration is evaluated (Titi &
Biondini 2012). For this reason, a numerical solution of the Ficks diffusion laws in two or
three dimensions may be necessary for accurate life-cycle evaluations.

The diffusion differential equation can be effectively solved numerically by means of
cellular automata which, in their basic form, consists of regular uniform grids of cells with a
discrete variable in each cell which can take on a finite number of states (Wolfram 1994). It
can be shown that the Ficks laws in d-dimensions can be accurately reproduced by
adopting the following evolutionary rule (Biondini et al. 2004b):

=
+
+
+

+ =
d
j
k
j i
k
j i
k
i
k
i
C C
d
C C
1
, 1 , 1
0
0
1
) (
2
1
(2)

where the discrete variable C
i
k
= C(z
i
,t
k
) represents the concentration of the component at
time t
k
in the cell i located at point z
i
=(x
i
,y
i
,z
i
),
k
j i
C
, 1
is the concentration in the adjacent cells
i1 in the direction j=1,..,d, and
0
is a suitable evolutionary coefficient related to the rate
of mass diffusion. In order to regulate the process according to a given diffusivity D, a
proper discretization in space and time should be chosen in such a way that the grid
dimension x=y=z and the time step t satisfy the following relationship:


t
x
d
D

=
2
0
2
1
(3)

A proof is given in Biondini et al. (2008). The value
0
= 1/2 usually ensures a good
accuracy of the automaton. A validation of this approach can be found in Biondini (2011).


ANAIS DO 55 CONGRESSO BRASILEIRO DO CONCRETO - CBC2013 55CBC

4
3 Modeling of Corrosion Damage

3.1 Reduction of the Cross-section of Reinforcing Steel Bars

The most relevant effect of corrosion is the reduction of the cross-section of the reinforcing
steel bars. The time evolution of the area A
s
of a corroded bar is represented as follows:


0
)] ( 1 [ ) (
s s s
A t t A = (4)

where A
s0
is the area of the undamaged steel bar and
s
=
s
(t) is a dimensionless damage
index which provides a measure of cross-section reduction in the range [0; 1]. The damage
function
s
=
s
(t) depends on both corrosion rate and corrosion mechanism. The relationship
between the damage index
s
and the corrosion penetration depth associated to uniform
corrosion, pitting corrosion, and mixed type of corrosion with components of uniform and
pitting corrosion, can be found in Biondini & Vergani (2012).

3.2 Reduction of Ductility of Reinforcing Steel

Pitting corrosion may involve a significant reduction of steel ductility. Tensile tests on
corroded bars show that for a relatively small mass loss (about 13%) steel behavior may
become brittle (Almusallam 2001). The results of experimental tests reported in
Apostolopoulos & Papadakis (2008) indicate that ductility reduction of corroded steel bars
depends on the loss of resistant area. Based on these results, the steel ultimate strain
su

can be related to the damage index
s
as follows (Biondini & Vergani 2012):

<
<
=

1 0.016 , 1521 . 0
016 . 0 0 ,
0
4583 . 0
0
s su s
s su
su

(5)

where
su0
is the steel ultimate strain of the undamaged bar.

3.3 Effects of Corrosion on Concrete

Effects of corrosion are not limited to damage of reinforcing steel bars. In fact, in case of
uniform corrosion with low penetration rate, the formation of oxidation products may lead
to propagation of longitudinal cracks and concrete cover spalling. This local deterioration
of concrete can effectively be modeled by means of a degradation law of the effective
resistant area of concrete matrix A
c
(Biondini et al. 2004b):


0
)] ( 1 [
c c c
A t A = (6)

where A
c0
is the area of undamaged concrete and
c
=
c
(t) is a dimensionless damage
function which provides a measure of concrete damage in the range [0; 1].

ANAIS DO 55 CONGRESSO BRASILEIRO DO CONCRETO - CBC2013 55CBC

5
Alternatively, concrete degradation can be taken into account by means of a reduction of
concrete compression strength f
c
due to cover cracking (Biondini & Vergani 2012):


0
)] ( 1 [
c c c
f t f = (7)

where f
c0
is the strength of undamaged concrete. In this approach, an effective empirical
relationship between the damage index
c
and the amount of steel damage
s
can be
established based on experimental results (Vidal et al. 2004, Zhang et al. 2009).

3.4 Damage Rate

The corrosion rate of steel depends on the concentration of the aggressive agent. This
dependence is, in general, complex and the available information about environmental
agents and material characteristics is usually not sufficient for a detailed modeling of the
damage process. However, despite such difficulties, simple models can be often
successfully adopted for an overall evaluation of the lifetime structural performance. Based
on available data for sulfate and chloride attacks and correlations between chloride
content and corrosion current density in concrete (Bertolini et al. 2004, Liu & Weyers 1998,
Pastore & Pedeferri 1994) the approach presented in this paper assumes, for both
materials, a linear relationship between the rate of damage and the concentration C=C(z,t)
of the aggressive agent (Biondini et al. 2004b):


c c
c
t C
t C
t
t

) , ( ) , ( z z
(8)

s s
m m sm
t C
t C
t
t

) , ( ) , ( z z
(9)

where
c
=
c
(z,t) and
sm
=
sm
(z
m
,t) are the time-variant damage functions of concrete and of
the m
th
steel bar, respectively, C
c
and C
s
are the values of constant concentration leading
to complete damage of concrete and steel over the time intervals t
c
and t
s
. The initial
conditions
c
(t
cr
)=
s
(t
cr
)=0, with t
cr
= max{t | C(t)C
cr
}, are assumed.


4 Structural Analysis of Concrete Structures under Damage

4.1 Time-variant Nonlinear Analysis

The formulation of a reinforced concrete beam finite element for time-variant nonlinear
analysis of concrete frame structures is considered. Mechanical nonlinearity, associated to
the constitutive properties of the materials, and geometrical nonlinearity, due to the second
order effects, are taken into account (Bontempi et al. 1995, Malerba 1998, Biondini et al.
2004a, Biondini 2004).

ANAIS DO 55 CONGRESSO BRASILEIRO DO CONCRETO - CBC2013 55CBC

6
For the cross-section at abscissa x, the time-variant vectors of stress resultants
r(x,t)=[N M
z
M
y
]
T
(axial force N and bending moments M
z
and M
y
) and global strains
e(x,t)=[
0

z

y
]
T
(axial strain
0
and bending curvatures
z
and
y
) are related as follows:
) , ( ) , ( ) , ( t x t x t x e H r = (10)
where H=H(x,t) is the time-variant stiffness matrix of the cross-section, that is computed by
integration over the area of the composite element, or by assembling the contributions of
both concrete H
c
=H
c
(x,t) and steel H
s
=H
s
(x,t):
) , ( ) , ( ) , ( t x t x t x
s c
H H H + = (11)

=
) (
)] , , , ( 1 )[ , ( ) , ( ) , , , ( ) , (
x A
c
T
c c
c
dA t z y x z y z y t z y x E t x b b H (12)

=
m
m s sm m
T
m m s s
A t x t x E t x
0
)] , ( 1 [ ) , ( ) , ( b b H (13)
where E
c
=E
c
(x,y,z) and E
sm
=E
sm
(x) are the generalized moduli of the damaged materials,
the symbol m refers to the m
th
steel bar located at point ) , (
m m
z y , and b(y, z) = [1 y z ]
T
.

The mechanical K
M
=K
M
(t) and geometrical K
G
=K
G
(t) contributions to the stiffness matrix
K
e
=K
e
(t) of the beam element are obtained by integration over the beam length l:
) ( ) ( ) ( t t t
G M e
K K K + = (14)

=
l
T
M
dx x t x x t

0
) ( ) , ( ) ( ) ( B S B K (15)

=
l
T
G
dx x x t x N t

0
) ( ) ( ) , ( ) ( G G K (16)
where N=N(x) is the axial force, and B=B(x) and G=G(x) are compatibility matrices
associated to Hermitian shape functions.

The stiffness matrices H
c
and K
M
are evaluated at each time instant by numerical
integration. The integration can be effectively performed based on the grid of the cellular
automaton adopted to describe the diffusion process (Biondini et al. 2004b), as well as by
using Gauss-Legendre or Gauss-Lobatto integration rules (Malerba 1998, Bontempi et al.
1995). Finally, by assembling the stiffness matrices K
e
of all elements, the equilibrium of
the structure can be expressed by the following system of nonlinear equations:
) ( ) ( ) , ( t t t f s s K = (17)
where K=K(s,t) is the stiffness matrix, s=s(t) is the vector of nodal displacements, and
f=f(t) is the vector of applied forces. It is worth noting that the vectors f and s have to be
considered as total or incremental quantities depending on the nature of the element
stiffness matrices H and K
e
, which depends on the adopted formulation (i.e., secant or
tangent) for the generalized moduli of the materials.

ANAIS DO 55 CONGRESSO BRASILEIRO DO CONCRETO - CBC2013 55CBC

7
The deteriorating reinforced concrete beam finite element has been successfully validated
(see Biondini & Vergani 2012) with reference to the results experimental tests carried out
on beams with corroded reinforcement (Rodriguez et al. 1997, Castel et al. 2000).

4.2 Time-variant Limit Analysis

The numerical solution process of the time-variant nonlinear structural analysis problem is in
general computationally demanding, particularly within probabilistic frameworks. For
problems where only the limit state of structural collapse is of interest, the theory of limit
analysis can represent a more effective approach to time-variant structural analysis.

Let 0 be a scalar multiplier of the live loads. By assuming that the structure is safe for
= 0, the collapse multiplier
c
associated with structural collapse can be obtained from
the static (lower bound) and kinematic (upper bound) theorems of limit analysis under the
hypotheses of perfectly plastic behavior and negligible second order effects. In spite of
such idealizations, limit analysis can be successfully applied to concrete structures by
assuming a suitable effective value of the concrete compression strength (Nielsen 1999).
Moreover, it can be shown that the limit analysis problem can be conveniently formulated
as a linear programming problem if a stepwise linear approximation of the resistance
domains is adopted. For frame structures, resistance domains are defined at cross-
sectional level in terms of internal stress resultants r=r(x), i.e. axial force N and bending
moments M
z
and M
y
.

Due to the deterioration of materials, the cross-sectional resistance domains, and
consequently the corresponding collapse multiplier
c
, vary over time. A time-variant limit
analysis leading to assess the time evolution of the collapse multiplier
c
=
c
(t) can be
performed by solving the linear programming problem at several time instants. Details on
this approach can be found in Biondini & Frangopol (2008b).


5 Time-variant Structural Reliability and Lifetime Assessment

5.1 Probability of Failure and Reliability index

Let R=R(t) and S=S(t) be time-variant measures of structural resistance and demand,
respectively. Because of the uncertainties involved in the problem, both functions R=R(t)
and S=S(t) have to be considered as random variables or processes. By denoting r
k
and s
k

the outcomes of the random variables R
k
=R(t
k
) and S
k
=S(t
k
), respectively, the probability of
failure at time t = t
k
can be evaluated by the integration of the joint density function
f
k
R,S
(r,s) within the failure domain D
k
= { r, s | r
k
<s
k
}:

= < =
k
D
k
S R k k k F
s r s r f S R P t P

,
d d ) , ( ] [ ) ( (18)

ANAIS DO 55 CONGRESSO BRASILEIRO DO CONCRETO - CBC2013 55CBC

8
Alternatively, the structural safety can also be measured by means of the time-variant
reliability index =(t):

k S k R k k S k R
k S k R
k
t
, ,
2
,
2
,
, ,
2
) (

+

= (19)
where
k
=(t
k
) and
k
=(t
k
) denote the mean and standard deviation of random variables,
respectively, and
k
=(t
k
) is the correlation coefficient relating R
k
and S
k
.

In practice the joint density distribution of R and S is not known, and at most some
information is available only about a set of N basic random variables x=[ x
1
x
2
x
N
]
T

which defines the problem at the initial time t = t
0
(Biondini et al. 2004a). By denoting
0 x g ) , ( t the corresponding set of time-variant limit state functions, the probability of
failure at time instant t = t
k
can be evaluated by the integration of the joint density function
f
k
x
(x) within the failure domain D
k
= { x | 0 x g x g < = ) , ( ) (
k k
t }:
) ( d ) ( ] ) ( [ ) (

k
D
k
k k F
k
f P t P = = < =

x x 0 x g
x
(20)
where = () is the standard normal cumulative probability function.

Finally, it is worth noting that in structural design the levels of verification are usually
formulated in terms of functions of random variables y=y(x) which describe the structural
response, and such derivation is generally only available in an implicit form. For this
reason, in most cases a numerical approach, based for example on Monte Carlo
simulation, is required to perform a life-cycle reliability analysis (Ang & Tang 2007).

5.2 Structural Lifetime Assessment

The probabilistic assessment of the time-variant structural reliability allows to evaluate the
lifetime T of the structure (Biondini et al. 2006a):
{ } 0 ) , ( ) ( min
0
< = t t t T x g (21)
where t
0
is the time instant at the end of the construction phase. In particular, the limit
threshold T* of the random variable T associated to a given target reliability level
expressed in terms of acceptable values of probability of failure P
F
* or reliability index *,
can be evaluated as follows:
{ } { }
*
0
*
0
*
) ( min ) ( min = = t t P P t t T
F F
(22)
This formulation can be used to select proper design strategies for new structures, or to
plan effective repair and maintenance strategies for existing structures, in order to achieve
a given design value of the structural lifetime T
d
T
*
(Biondini et al. 2006a).

ANAIS DO 55 CONGRESSO BRASILEIRO DO CONCRETO - CBC2013 55CBC

9
6 Application to an Arch Bridge

The presented methodology is applied to a concrete arch bridge exposed to corrosion
(Biondini & Frangopol 2008b, Biondini 2011). The structural scheme and dimensions of the
bridge are shown in Figure 1 (Galli & Franciosi 1955). The beam has a two-cellular cross-
section with main dimensions 2.006.00 m (Figure 2.a). The other dimensions are: web
thickness = 0.20 m; top slab thickness = 0.18 m; bottom slab thickness = 0.16 m. The
distribution of the reinforcement along the beam is given in Table 1 with reference to the
subdivision shown in Figure 3.a. The arch has a rectangular cross-section with dimensions
0.576.00 m and it is reinforced with 45+45=90 steel bars with diameter 28 mm (Figure
2.b). The nominal values of the material strengths are f
c
=30 MPa and f
sy
=300 MPa. The
structure is subjected to a set of dead loads g and to a live load q, as shown in Figure 1.


Figure 1 Geometrical dimensions, structural scheme, and loading of the arch bridge.

(a)
(b)
Figure 2 Details of the cross-sections and reinforcement layout: (a) beam; (b) arch.

g
g
p
0
1
l
h f
x
y

Arch Shape
x [m] y [m]
0.00 0.00
4.44 5.90
8.89 11.43
13.35 15.39
17.80 19.04
22.20 21.53
26.70 23.76
31.20 24.99
35.60 25.98
40.00 26.10


kN/m 3 . 53
kN/m 0 . 85
kN/m 9 . 102
m 00 . 27
m 10 . 26
m 00 . 80
1
0
=
=
=
=
=
=
q
g
g
h
f
l

q

ANAIS DO 55 CONGRESSO BRASILEIRO DO CONCRETO - CBC2013 55CBC

10
Table 1. Reinforcement of the beam (see Figure 3.a).
Segment 1 2 3 4 5 6 7 8 9
A
s
2128
1308
4828
1308
4228
1308
3028
1308
2428
1308
4828
1308
4828
1308
4528
1308
3328
1308
A
s
2128 3028 4228 2428 2428 2128 3628 2728 2428



(a) (b)
Figure 3 Strength capacity of the undamaged cross-sections (t=t
0
=0): (a) Beam subdivision (see Table 1) and
resistant bending moments. (b) Axial force vs bending moment resistant domain of the arch cross-section.

The arch bridge is subjected to a diffusive attack from an environmental aggressive agent
located with concentration C(t)=C
0
along the free edges of both beam and arch. The diffusion
process is described by assuming a nominal diffusivity coefficient D = 10
11
m
2
/sec. The
damage rates are defined by assuming the nominal values C
c
=C
s
=C
0
, t
c
=25 years,
t
s
=50 years, and C
cr
= 0. Structural damage is modeled by assuming uniform corrosion of
the steel bars and a reduction of the effective resistant area of the concrete matrix.

Figure 3 shows the resistant bending moment diagram of the axially unloaded beam
(Figure 3.a) and the axial force-bending moment resistance diagram of the arch (Figure 3.b)
for the undamaged structure. At the cross-sectional level the damage effects induced by
diffusion lead to a reduction over time of these strength capacities, as reported in Biondini
& Frangopol (2008b). Consequently, at the system level the local damage causes a
deterioration over time of the carrying load capacity of the bridge. In particular, the time-
variant multiplier
c
=
c
(t) of the live load associated with structural collapse is obtained by
limit analysis. Three limiting scenarios with (I) simultaneous deterioration of both beam and
arch, (II) damage of the beam only, and (III) damage of the arch only, are investigated to
highlight the effects of different exposures. The time evolution of the nominal value of the
collapse multiplier for the three investigated cases is shown in Figure 4.
-24.0
-20.0
-16.0
-12.0
-8.0
-4.0
0.0
4.0
8.0
12.0
16.0
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0
B
e
n
d
i
n
g

M
o
m
e
n
t


M
[
M
N
m
]
Abscissa x [m]
1 2 3 4
5
9 8 7 6
3/4 a a/2 a/2 a
-12.0
-9.0
-6.0
-3.0
0.0
3.0
6.0
9.0
12.0
-120.0 -100.0 -80.0 -60.0 -40.0 -20.0 0.0 20.0
B
e
n
d
i
n
g

M
o
m
e
n
t


M
[
M
N
m
]
Axial Force N [MN]

ANAIS DO 55 CONGRESSO BRASILEIRO DO CONCRETO - CBC2013 55CBC

11

Figure 4 Time evolution of the collapse multiplier
c
for three damage scenarios: (I) damage of both the
beam and arch; (II) damage of the beam only; (III) damage of the arch only.


The comparison of the results indicates that case (I) is the worst scenario, as expected, and
that the effects of damage in case (II) are larger than in case (III). In all cases the damage
process leads to a significant reduction of the collapse multiplier. It is worth noting that such
reduction involves a redistribution of the internal stresses, with a consequent modification of
the collapse mechanism over time, as reported in Biondini & Frangopol (2008b).

The uncertainties involved in the problem are taken into account based on a probabilistic
modeling of both diffusion and damage processes. The probabilistic model assumes as
random variables the strengths of concrete f
c
and steel f
sy
, the coordinates (y
p
, z
p
) of each
nodal point p of the member cross-section, the coordinates (y
m
, z
m
) and diameter
m
of
each steel bar m, the diffusivity coefficient D, and the damage rates q
c
=(C
c
t
c
)
1
and
q
s
=(C
s
t
s
)
1
. These variables are assumed uncorrelated with probability density functions
and standard deviation from nominal values listed in Table 2 (Biondini et al. 2006a).

Table 2. Probability density functions and standard deviation from nominal (mean) values.
Random Variable (t = 0) Distribution Type Standard Deviation
Concrete strength, f
c
Lognormal 5 MPa
Steel strength, f
sy
Lognormal 30 MPa
Coordinates of the nodal points, (y
p
, z
p
) Normal 5 mm
Coordinates of the steel bars, (y
m
, z
m
) Normal 5 mm
Diameter of the steel bars,
m
Normal
(*)
0.10
m,nom

Diffusivity, D Normal
(*)
0.10 D
nom

Concrete damage rate, q
c
=(C
c
t
c
)
1
Normal
(*)
0.30 q
c,nom

Steel damage rate, q
s
=(C
s
t
s
)
1
Normal
(*)
0.30 q
s,nom

(*)
Truncated distributions with non negative outcomes. nom = nominal values
1.00
1.50
2.00
2.50
3.00
3.50
4.00
4.50
0 10 20 30 40 50
C
o
l
l
a
p
s
e

M
u
l
t
i
p
l
i
e
r

c
Time t [years]
Beam & Arch
Beam only
Arch only

ANAIS DO 55 CONGRESSO BRASILEIRO DO CONCRETO - CBC2013 55CBC

12
With reference to the exposure scenario (I), Figure 5.a shows a set of probability curves
which provide at each point in time the probability of failure P
F
for deterministic target
levels of the live load multiplier. Based on these life-cycle probability curves, the lifetime T*
associated to given target reliability levels * can be computed as a function of the live
load multiplier, as shown in Figure 5.b. These results allow to assess the remaining
structural lifetime under prescribed reliability levels without maintenance.



(a) (b)
Figure 5 (a) Time-variant probability of failure P
F
(t = 1 year) and (b) residual structural lifetime T*
associated to prescribed values of the reliability index (*=1.0, 2.0, 3.0) versus deterministic target values of
the live load multiplier for damage scenario (I).

7 Conclusions

Damage processes in concrete structures are usually investigated based on simplified
models of diffusion processes and through the study of the local deterioration of the
materials, concrete and steel, with limited attention paid to the global effects of these local
phenomena on the overall performance of the structural system. This is clearly not
consistent with the actual nature of the problem, since the simulation of diffusion processes
should be able to account for complex geometrical and mechanical boundary conditions,
which generally characterize engineering applications. Moreover, the local deterioration
mechanisms interact with the global structural response. As a consequence, the structural
scheme plays a fundamental role in the assessment of deteriorating structures, particularly
for redundant systems, where damage may lead to time-variant redistributions of the
internal actions. These aspects can be consistently taken into account by means of the
general methodology presented in this paper for time-variant reliability analysis and
lifetime assessment of concrete structures exposed to the diffusive attack from
environmental aggressive agents, with emphasis on concrete bridges under corrosion.

0.00
0.20
0.40
0.60
0.80
1.00
0.00 1.00 2.00 3.00 4.00 5.00
P
r
o
b
a
b
i
l
i
t
y

o
f

F
a
i
l
u
r
e


P
F
Load Multiplier
t =50 years
t = 0
0
10
20
30
40
50
0.00 1.00 2.00 3.00 4.00 5.00
L
i
f
e
t
i
m
e

T
*

[
y
e
a
r
s
]
Load Multiplier
* = 1.0
* = 2.0
* = 3.0

ANAIS DO 55 CONGRESSO BRASILEIRO DO CONCRETO - CBC2013 55CBC

13
8 References

ALMUSALLAM, A. 2001. Effect of Degree of Corrosion on the Properties of Reinforcing Steel
Bars, Construction and Building Materials, 15, 361-368.
ANG, A.H.-S. & TANG, W. H. 2007. Probability Concepts in Engineering. 2
nd
Edition, John
Wiley & Sons, Hoboken, NJ, USA.
APOSTOLOPOULOS, C.A., PAPADAKIS, V.G. 2008. Consequences of Steel Corrosion on the
Ductility Properties of Reinforcement Bar, Construction and Building Materials, 22, 2316-2324.
BERTOLINI, L., ELSENER, B., PEDEFERRI, P. & POLDER, R. 2004. Corrosion of Steel in
Concrete. Wiley-VCH, Weinheim, Germany.
BIONDINI, F. 2004. A Three-dimensional Finite Beam Element for Multiscale Damage
Measure and Seismic Analysis of Concrete Structures. 13th World Conference on Earthquake
Engineering, Vancouver, B.C., Canada, August 1-6, Paper No. 2963.
BIONDINI, F., BONTEMPI, F., FRANGOPOL, D.M., MALERBA, P.G. 2004a. Reliability of
Material and Geometrically Nonlinear Reinforced and Prestressed Concrete Structures.
Computers & Structures, 82(13-14), 1021-1031.
BIONDINI, F., BONTEMPI, F., FRANGOPOL, D.M., MALERBA, P.G. 2004b. Cellular Automata
Approach to Durability Analysis of Concrete Structures in Aggressive Environments,
Journal of Structural Engineering, ASCE, 130(11), 1724-1737.
BIONDINI, F., BONTEMPI, F., FRANGOPOL, D.M. & MALERBA, P.G. 2006a. Probabilistic
Service Life Assessment and Maintenance Planning of Concrete Structures, Journal of
Structural Engineering, ASCE, 132(5), 810-825.
BIONDINI, F., FRANGOPOL, D.M. & MALERBA, P.G. 2006b. Time-variant Performance of the
Certosa Cable-stayed Bridge. Structural Engineering International, IABSE, 16(3), 2006, 235-244.
BIONDINI, F., FRANGOPOL, D.M. & MALERBA, P.G. 2008. Uncertainty Effects on Lifetime
Structural Performance of Cable-Stayed Bridges, Probabilistic Engineering Mechanics, 23(4),
509-522.
BIONDINI, F. & FRANGOPOL, D.M. 2008a. Life-Cycle Civil Engineering. CRC Press, Taylor &
Francis Group plc., A.A. Balkema, Boca Raton, London, New York, Leiden.
BIONDINI, F. & FRANGOPOL, D.M. 2008b. Probabilistic Limit Analysis and Lifetime
Prediction of Concrete Structures, Taylor & Francis, Structure and Infrastructure Engineering,
4(5), 399-412.
BIONDINI, F. (Ed.) 2009. Long-term Performance of Structural Systems, Special Issue of
Structure and Infrastructure Engineering, Taylor & Francis, London, 4(2), 2008, 75-176.
BIONDINI, F., FRANGOPOL, D.M. 2009. Lifetime Reliability-Based Optimization of Reinforced
Concrete Cross-Sections under Corrosion, Structural Safety, 31, 483-489.
BIONDINI, F. 2011. Cellular Automata Simulation of Damage Processes in Concrete
Structures, In: Soft Computing Methods for Civil and Structural Engineering, Y. Tsompanakis and
B.H.V. Topping, (Eds.), Saxe-Coburg Publications, Stirlingshire, Scotland, Chapter 10, 229-264.

ANAIS DO 55 CONGRESSO BRASILEIRO DO CONCRETO - CBC2013 55CBC

14
BIONDINI, F. & FRANGOPOL, D.M. (Eds.) 2011. Life-Cycle of Civil Engineering Systems,
Special Issue of Structure and Infrastructure Engineering, Taylor & Francis, London, Vol. 7, Nos.1-
2, 2011, 1-196.
BIONDINI, F., PALERMO, A. & TONIOLO, G. 2011. Seismic performance of concrete
structures exposed to corrosion: Case studies of low-rise precast buildings. Structure and
Infrastructure Engineering, 7(12), 109119.
BIONDINI, F. & FRANGOPOL, D.M. 2012. Bridge Maintenance, Safety, Management,
Resilience and Sustainability. CRC Press/Balkema, Taylor and Francis Group, London.
BIONDINI, F. & VERGANI, M. 2012. Damage modeling and nonlinear analysis of concrete
bridges under corrosion, Sixth International Conference on Bridge Maintenance, Safety and
Management (IABMAS 2012), Stresa, Italy, July 8-12, 2012. In: Bridge Maintenance, Safety,
Management, Resilience and Sustainability, F. Biondini, D.M. Frangopol (Eds.), CRC Press.
BIONDINI, F., CAMNASIO, E. & PALERMO, A. 2013. Lifetime seismic performance of concrete
bridges exposed to corrosion. Structure and Infrastructure Engineering, DOI:
10.1080/15732479.2012.761248 (In press).
BONTEMPI, F., MALERBA, P. G., ROMANO, L. 1995. Formulazione diretta secante dellanalisi
non lineare di telai in C.A./C.A.P., Studi e Ricerche, Graduate School for Concrete Structures
F.lli Pesenti, Politecnico di Milano, Italy, 16, 351-386 (in Italian).
CASAS, J.R., FRANGOPOL, D.M. & NOWAK, A.S. 2002. Bridge Maintenance, Safety and
Management. CIMNE, Barcelona, Spain.
CASTEL, A., FRANOIS, R., ARLIGUIE, G. 2000. Mechanical Behaviour of Corroded
Reinforced Concrete Beams - Part 1: Experimental Study of Corroded Beams. Materials and
Structures, 33, 539-544.
CEB 1992. Durable Concrete Structures Design Guide, Thomas Telford.
CHEN, S-S, FRANGOPOL, D. M. & ANG, A-H. S. 2010. Life-Cycle of Civil Engineering
Systems, Taiwan Building Technology Center, DnE Information Service Net, Taipei, Taiwan, 2010.
CHO, H-N., FRANGOPOL, D.M. & ANG, A-H.S. (Eds.) 2007. Life-Cycle Cost and Performance
of Civil Infrastructure Systems, Taylor & Francis Group plc., A. A. Balkema, London.
CRUZ, P.J.S., FRANGOPOL D.M., & NEVES, L.C. (Eds.) 2006. Bridge Maintenance, Safety,
Management, Life-Cycle Performance and Cost. CRC Press/Balkema, Taylor and Francis
Group, London.
ELLINGWOOD, B. R. 2005. Risk-informed Condition Assessment of Civil Infrastructure:
State of Practice and Research Issues. Structure and Infrastructure Engineering, Taylor &
Francis, 1(1), 7-18.
ESTES, A.C. & FRANGOPOL D.M. 2005. Life-Cycle Evaluation and Condition Assessment of
Structures, Chapter 36 in Structural Engineering Handbook, Second Edition, W-F. Chen & E. M.
Lui (Eds.), CRC Press, 36-1 to 36-51.
fib 2006. Model Code for Service Life Design. Bulletin, 34.

ANAIS DO 55 CONGRESSO BRASILEIRO DO CONCRETO - CBC2013 55CBC

15
FRANGOPOL, D.M., LIN, K-Y. & ESTES, A.C. 1997a. Reliability of reinforced concrete girders
under corrosion attack, Journal of Structural Engineering, ASCE, 123(3), 286-297.
FRANGOPOL, D.M., LIN. K-Y. & ESTES, A.C. 1997b. Life-Cycle Cost Design of Deteriorating
Structures, Journal of Structural engineering, ASCE, 123(10), 1390-1401.
FRANGOPOL, D.M. 1999. Life-Cycle Cost Analysis for Bridges, Chapter 9 in Bridge Safety and
Reliability, Frangopol, D.M., ed., ASCE, Reston, Virginia, 210-236.
FRANGOPOL, D.M. & FURUTA, H. (Eds.) 2001. Life-Cycle Cost Analysis and Design of Civil
Infrastructure Systems, ASCE , Reston, Virginia.
FRANGOPOL, D.M., MIYAKE, M., KONG, J.S., GHARAIBEH, E.S. & ESTES A.C. 2002.
Reliability and CostOriented Optimal Bridge Maintenance Planning, Chapter 10 in Recent
Advances in Optimal Structural Design, S. Burns, ed., ASCE, Reston, Virginia, 257-270.
FRANGOPOL, D. M., KALLEN, M-J., VAN NOORTWIJK, J. 2004a. Probabilistic Models for Life-
Cycle Performance of Deteriorating Structures: Review and Future Directions. Progress in
Structural Engineering and Mechanics, 6(4), 197-212
FRANGOPOL, D.M., BRUHWILER, E., FABER, M.H. & ADEY, B. (Eds.) 2004b. Life-Cycle
Performance of Deteriorating Structures: Assessment, Design and Management, ASCE ,
Reston, Virginia.
FRANGOPOL, D.M. & LIU, M. 2007. Maintenance and Management of Civil Infrastructure
based on Condition, Safety, Optimization, and Life-Cycle Cost, Structure and Infrastructure
Engineering, Taylor & Francis, 3(1), 29-41.
FRANGOPOL, D.M., KAWATANI, M. & KIM, C-W. (Eds.) 2007. Reliability and Optimization of
Structural Systems: Assessment, Design, and Life-Cycle Performance, Taylor & Francis
Group plc., A. A. Balkema, London.
FRANGOPOL, D.M., SAUSE, R. & KUSKO, C.S. 2010. Bridge Maintenance, Safety,
Management and Life-Cycle Optimization, CRC Press, Taylor & Francis Group plc., A.A.
Balkema, Boca Raton, London, New York, Leiden.
FRANGOPOL, D.M. & ELLINGWOOD, B.R. 2010. Life-Cycle Performance, Safety, Reliability
and Risk of Structural Systems, Editorial, Structure Magazine, Joint Publication of NCSEA,
CASE, SEI.
FRANGOPOL, D.M. 2011. Life-Cycle Performance, Management, and Optimization of
Structural Systems under Uncertainty: Accomplishments and Challenges, Structure and
Infrastructure Engineering, Taylor & Francis, 7(6), 389-413.
FRANGOPOL, D.M., SAYDAM, D. & KIM, S. 2012. Maintenance, Management, Life-Cycle
Design and Performance of Structures and Infrastructures: A Brief Review, Structure and
Infrastructure Engineering, Taylor & Francis, 8(1), 1-25.
FURUTA, H., KAMEDA, T., NAKAHARA, K., TAKAHASHI, Y. & FRANGOPOL, D.M. 2006.
Optimal Bridge Maintenance Planning using Improved Multi-Objective Genetic Algorithm,
Structure and Infrastructure Engineering, Taylor & Francis, 2(1), 33-41.
GALLI, A. & FRANCIOSI, V. 1955. Il calcolo a rottura dei ponti a volta sottile ed impalcato
irrigidente. Giornale del Genio Civile, 11, 686-700, 1955 (in Italian).

ANAIS DO 55 CONGRESSO BRASILEIRO DO CONCRETO - CBC2013 55CBC

16
GLICKSMAN, M. E. 2000. Diffusion in Solids, John Wiley and Sons.
KILARESKI, W.P. 1980. Corrosion Induced Deterioration of Reinforced Concrete An
Overview, Materials Performance, NACE, 19(3), 48-50.
KOH, H-M. & FRANGOPOL, D.M. 2008. Bridge Maintenance, Safety, Management, Health
Monitoring and Informatics. CRC Press/Balkema, Taylor and Francis Group, London.
LIU, T., & WEYERS, R.W. 1998. Modeling the Dynamic Corrosion Process in Chloride
Contaminated Structures. Cement and Concrete Research, 28(3), 365379.
MALERBA, P.G. (Ed.) 1998. Limit and Nonlinear Analysis of Reinforced Concrete Structures.
International Centre for Mechanical Sciences (CISM), Udine, Italy (in Italian).
MALERBA, P.G. 2013. Inspecting and Repairing Old Bridges: Experiences and Lessons.
Structure and Infrastructure Engineering, Taylor & Francis, DOI:10.1080/15732479.2013.769010
(In press).
NIELSEN, M.P. 1999. Limit Analysis and Concrete Plasticity, CRC Press.
NOWAK, A.S. & FRANGOPOL, D.M. (Eds.) 2005. Advances in Life-Cycle Analysis and Design
of Civil Infrastructure Systems, Lincoln, Nebraska.
PASTORE, T., PEDEFERRI, P. 1994. La corrosione e la protezione delle opere metalliche
esposte allatmosfera, Ledilizia, December, 1994, 75-92 (In Italian).
RODRIGUEZ, J., ORTEGA, L.M., CASAL, J. 1997. Load Carrying Capacity of Concrete
Structures with Corroded Reinforcement, Construction and Building Materials, 11(4), 239-248.
STRAUSS, A., FRANGOPOL, D.M. & BERGMEISTER K. 2012. Life-Cycle and Sustainability of
Civil Infrastructure Systems, CRC Press, Taylor & Francis Group plc., A. A. Balkema, Boca
Raton, London, New York, Leiden.
TITI, A. & BIONDINI, F. 2012. Validation of Diffusion Models for Life-Cycle Assessment of
Concrete Structures. Third International Symposium on Life-Cycle Civil Engineering (IALCCE
2012), Vienna, Austria, October 3-6, 2012, 2382-2389.
VIDAL, T., CASTEL, A., FRANCOIS, R. 2004. Analyzing Crack Width to Predict Corrosion in
Reinforced Concrete, Cement and Concrete Research, 34, 165-174.
WATANABE, E., FRANGOPOL, D.M. & UTSUNOMIYA, T. (Eds.) 2004. Bridge Maintenance,
Safety, Management and Cost. A. A. Balkema, Leiden, The Netherlands.
WOLFRAM, S. 1994. Cellular Automata and Complexity Collected Papers, Addison-Wesley.
ZHANG, R., CASTEL, A., FRANOIS, R. 2009. Concrete Cover Cracking with Reinforcement
Corrosion of RC Beams During Chloride-Induced Corrosion Process, Cement and Concrete
Research, 39(11), 1077-1086.
ZHU, B., FRANGOPOL, D.M. 2013. Risk-Based Approach for Optimum Maintenance of
Bridges under Traffic and Earthquake Loads, Journal of Structural Engineering, ASCE, 139(3),
422-434.