5th International Conference on

Structural Health Monitoring of Intelligent Infrastructure (SHMII-5) 2011

11-15 December 2011, Cancn, Mxico

VEHICULAR BRIDGES PIERS

RELIABILITY: THE CASE OF THE STATE
OF MEXICO
David Joaqun Delgado Hernndez,
Juan Carlos Arteaga Arcos,
David de Len Escobedo.
Full Time Lecturers, Engineering School,
Autonomous University of the State of
Mexico, Cerro de Coatepec, Ciudad
Universitaria, Toluca, Mxico, C.P. 50130

Jair Gonzlez Rojas,

Jos Omar Jimnez Miranda,
Luis Horacio Martnez Martnez,
Jos Emmanuel Rivero Santana.
Engineering School, Autonomous
University of the State of Mexico, Cerro
de Coatepec, Ciudad Universitaria,
Toluca, Mxico, C.P. 50130

ABSTRACT
Bridges are structures whose main purpose is to overcome natural and artificial
obstacles, being the more common kind those aimed at moving vehicles in a road or
way. Because they play an important role in any countrys development, attention
should be paid to preserve these structures in good condition. Note that when a bridge
fails, both economic and human costs are common, these costs might affect nearby
communities. The present research work presents the calculations of the reliability index
(a) on vehicular bridges piers that are subjected to, for instance accidental lateral loads.
To do so, a sample of bridges situated in the State of Mexico has been used as reference.
Various scenarios were appraised and the optimal value of a was determined for
different pier geometries.
INTRODUCTION
Several authors have presented their views about the purpose of designing a structure
(Rosenblueth, E.; Esteva, L., 1972). For example, it has been argued (Agarwal, J., 2003)
that the objective of a design process is to reach an acceptable probability that the
designed structures will behave satisfactorily during their lifetimes. Therefore they are
designed to withstand all possible loads and deformations during construction and
normal use. Meli, (2004) sustains that the structural safety is defined during the design
process, when the designer must verify that the resistance is over the demands that will
act over it during its lifetime. Such descriptions have implicit the concept of structural
reliability.
According to Meli, (2004), the reliability of a structure is associated to a certain cost
which should be minimized to balance safety with cost. Therefore, an optimization
process should be performed where the objective function must include the initial cost
of the work and the cost of the potential damages and other consequences in case a
failure occur. Therefore, if Ct is the total cost of the structure, Ci the initial cost, Cd the
cost of failure consequences and Pf the failure probability:

5th International Conference on

Structural Health Monitoring of Intelligent Infrastructure (SHMII-5) 2011
11-15 December 2011, Cancn, Mxico

C t = C i + C d Pf

(1)

ACCEPTABLE FAILURE PROBABILITY

By recognizing the uncertainties inherent in the design process, especially the seismic
hazard, it has been proposed (Frangopol et al, 2001), to appraise bridge performance by
using the expected life-cycle assessment. In the offshore technology (Stahl, 1986), the
expected life-cycle cost E[Ct], is expressed in terms of the initial cost Ci, and the
expected failure/damage cost E[Cd].
E[Ct ] = Ci + E[C d ]

(2)

Where:
E[C d ] = PVF ( Pf )C d

(3)

And PVF is the present value factor. Given that this formulation includes all possible
ad-verse events, either failure or damage that may occur within the bridge lifetime, the
PVF considers all those potentially damaging events not just the worst scenario. Also,
the aver-age damage cost Cd is composed by the costs of consequences:

Cd = Cr + C f + Ce

(4)

Where Cr is the repair/restitution cost, Cf is the cost related to fatalities and Ce is the
economic loss due to the service interruption, user costs, while the bridge is repaired or
rebuilt. PVF depends on the net annual discount rate r and the bridge lifetime T:

PVF =

1 exp(rT )
r

(5)

If the initial cost Ci is expressed as a function of the failure probability (Rosenblueth,

1986), the expected lifecycle cost becomes a function of the failure probability.
E [Ct ] = C1 C 2 ln( Pf ) + PVF ( Pf )C d

(6)

The acceptable (optimal) failure probability may then be calculated by minimizing the
expected life-cycle cost respect the failure probability.

E[Ct ]
=0
Pf
Pf =

0.434C 2
PVF [C d ]

(7)

(8)

The acceptable failure probability depends inversely of the cost of consequences which
means that, according to the bridge importance, the safety requirement should be stricter
as those consequences increase. Also, the requirement may be expressed in terms of the
bridge reliability index.

a = 1 (1 Pf )

(9)

5th International Conference on

Structural Health Monitoring of Intelligent Infrastructure (SHMII-5) 2011
11-15 December 2011, Cancn, Mxico

According to previous results (De Len et al, 2008) the cost of consequences has been
normalized to the initial cost and C2/Ci = 0.08 for typical bridges. Also, for T = 200
years and r = 0.08. For the bridge considered here, it has been estimated that the costs of
consequences is 800 times the initial cost (because of the high traffic volume) and,
therefore, the acceptable bridge reliability a is approximately 3.84.
BRIDGE RELIABILITY THEORETICALLY
From well known structural reliability theory, the bridge reliability may be calculated
(Ang and Tang, 2007) with expression (10):

E (G )

(10)

Where G is the bridge limit state considering its exposure to seismic loads, E(G) the
expected value of such limit state and G its standard deviation. Although the bridge is a
complex structural system, from previous analyses for typical bridges (De Len et al,
2008), the limit state has been conservatively approximated in terms of the failure of the
most critical structural element. It was found that this element is one of the main piles
and it is subject to a combination of axial load and bending. Therefore, G is calculated
as:

P M
G = 1 A + A
PR M R

(11)

Where PA is the maximum acting axial load, PR the axial resistant force, MA the
maximum acting moment and MR the resistant moment of the critical cross section.
Given that PA and MA are a consequence of the random earthquakes that may occur
during the bridge lifetime, these effects are random variables. Also, from the variability
of materials properties, the resistances PR and MR are also random. The standard
deviation G is:

G
G
i =1 X i
4

xi 2

(12)

In Equation (12), X is the vector of acting and resisting axial loads and moments, such
that, X1 = PA, X2 = PR, X3 = MA and X4 = MR, and the derivatives are evaluated on the
mean values. Therefore:

5th International Conference on

Structural Health Monitoring of Intelligent Infrastructure (SHMII-5) 2011
11-15 December 2011, Cancn, Mxico

PR

MA M A
P
2
M 2
+ A2 PR +
+
2
2
R
MR
PR
MR
2

(13)

Where MR, MA, PR and PA are the standard deviations of the resistant and acting
moments, and the resistant and acting axial loads, respectively.
APPLICATION ON SELECTED BRIDGE
The structure is a vehicular bridge built on federal road Toluca-Ciudad Altamirano on
Temascaltepec Town, located in the south of the State of Mexico. In order to conduct a
conservative analysis and due to the lack of official Construction Codes in such a State,
it was assumed that the bridge was located in the transition seismic zone III, according
to the Mexico Citys Construction Code. The bridge has a 70 m total span divided into 3
segments. The structural modelling was made through a finite element-based
commercial software (RAM Advanse, 2006). Figure 1 shows some pictures of the
bridge piers and caps as well as the lower side of road slabs. Figure 2 shows the finite
element-model studied herein. Essentially, the main structural components of the bridge
are: the transverse cap and one pier which was assumed to be a fixed supported.
The mean reinforced concrete properties are fc = 25.48 MPa and fy = 428 MPa.
BRIDGE RELIABILITY IN PRACTICE
A family of bridge designs were obtained (AASHTO, 2002) by varying the original
design dimensions and steel areas. These designs allowed for a series of alternative
designs to measure the variation of reliability with cost under specified seismic
intensities. The bridge designs were analyzed under given maximum seismic
coefficients c/g, using the typical spectral form for Mexico City, and according to the
range of intensities as reported in Mexican seismic hazard and failure rates studies
(Esteva and Ruiz, 1989).

5th International Conference on

Structural Health Monitoring of Intelligent Infrastructure (SHMII-5) 2011
11-15 December 2011, Cancn, Mxico

Figure 1 Main supports of the bridge

Figure 2 Main supports of the bridge

Table 1 shows a sample of the results obtained by varying the seismic coefficients from
0g to 0.6g at each 0.15g and for specific design alternatives (in this case a pier). The
Table contains the seismic coefficient, the rebars size, number of rebars, mean values of
maximum axial load and moment and axial and moment resistances, reliability index
and the initial costs obtained.
Table 1 Sample of the calculations for cost-reliability (pier)
c/g

0
0.15
0.3
0.45
0.6

Rebar
No PA (kN)
size Rebars
(/8)
22
11
2.246
26
11
2.231
28
11
2.215
30
11
2.215
32
11
2.184

PR (kN)

MA
(kN*m)

MR
(kN*m)

Ci (USD)

8.022
9.294
10.070
10.549
11.496

5.197
5.646
6.094
6.543
6.992

5.310
6.069
6.438
6.761
7.085

2.45
2.60
2.58
2.52
2.48

6338.4
6918.6
7193.6
7468.7
7743.7

Five alternative designs and the five maximum intensities, shown in Table 1, were
considered and the corresponding reliability indices and initial costs calculated. For the
standard deviations, they were calculated as CVA = 0.025 (following the procedure
proposed by De Len et al, 2008) and CVR = 0.026 (estimated from 20 field data
obtained by means of non destructive rebound hammer tests), and the following
simplifications were made (De Len et al, 2008):
CV A =

PA
CVA =

Pr

(14)

MA
=

Mr

(15)

5th International Conference on

Structural Health Monitoring of Intelligent Infrastructure (SHMII-5) 2011
11-15 December 2011, Cancn, Mxico

Note that Beta = 2.48 can be translated into a 7 x 10-3 failure probability, which is not
within the expected range for these type of infrastructure. The fact that this probability
is bigger than the 1 x 10-5 (Joint Comitee of Structural Safety, 2001), recommended for
the most common design situations, reveals that the material resistance collected data
could be improved, maybe with destructive tests.
CONCLUSIONS
As can be seen, the calculated Beta is far below from the optimum expected for this
kind of structures (De Len et al, 2008), which is estimated to be around 3.84. This
result is not surprising since both the CVA and CVR were obtained from a small
sample. As a result, it is strongly recommended to amplify the number of measures
taken in-site. For example, Stahl (1986) estimated such coefficients using a statistical
analysis based on 704 mill tests of ASTM A36 steel plates, that led to a coefficient of
variation of 0.09, i.e. a 4% probability of failure.
In any case, the probability of failure calculated can be associated with moderate
consequences, which imply risk to life and considerable economic costs (e.g. offices,
factories and apartment buildings). To sum up, more research is needed in order to
determine whether the studied bridge will require maintenance in the short or long term.

REFERENCES
Ang, A., Tang, W. (2007); Probability concepts in engineering. John Wiley, US
Agarwal, J. (2003); Protecting existing structures against unforeseen events, In: Procs.
of Applications of Statistics and Probability in Civil Engineering, Der Kiureghian,
Madana & Pestana (eds), Millpres, Netherlands, 775-780.
AASHTO (2002); Standard Specifications for Highway Bridges, American Association
of State Highway and Transportation Officials, US
De Len, D.; Ang, A.; Manjarez, L. (2008); Structural Reliability of the Tampico
Bridge under Wind Loading. Ciencia Ergo Sum. Julio-Octubre, Ao/Vol 15 numero
002. Universdiad Autnoma del Estado de Mxico pp 161 166, Mxico
Esteva, L. and Ruiz, S. (1989); Seismic failure rates of multistory frames. Journal of
Structural Engineering, ASCE, 115, 268-284
Frangopol, D.M.; Kong, J.S.; Gharaibeh, E.S. (2001); Reliability-based life-cycle
management of highway bridges. Journal of Computing in Civil Engineering, ASCE,
15, 27-34
Joint Comitee of Structural Safety (2001) PROBABILISTIC MODEL CODE, Part 1 BASIS OF DESIGN. JCSS-OSTL/DIA/VROU -10-11-2000.

5th International Conference on

Structural Health Monitoring of Intelligent Infrastructure (SHMII-5) 2011
11-15 December 2011, Cancn, Mxico

Meli, R. (1994); Diseo estructural. Limusa, Mxico, 1994

RAM Advanse (2006); Structural Analysis Software. 8th Version, Bentley Systems, US
Rosenblueth, E.; Esteva, L. (1972); Reliability basis for some Mexican codes. In:
Probabilistic design of Reinforced Concrete Building, ACI Public. SP-31, ACI, Detroit,
Mich., 1-42
Rosenblueth E. (1986); Optimum reliabilities and optimum design. Structural Safety, 3,
69-83
Stahl, B. (1986); Reliability Engineering and Risk Analysis. In: Planning and design of
fixed offshore platforms, McClelland., B. and Reifel., M. (Eds), VNR, New York.