Vulnerability, Uncertainty, and Risk ©ASCE 2014 2702

Reliability Analysis of a Circular Bridge Pier Subject to Intentional Vehicular

Impact

Kylie Steel1 and Andrew D. Sorensen2

Downloaded from ascelibrary.org by Brigham Young University on 11/18/14. Copyright ASCE. For personal use only; all rights reserved.

1
Civil and Environmental Engineering, Idaho State University, 921 South 8th
Avenue, 83209, Pocatello, ID; PH (208) 705-1640; email: smitkyli@isu.edu
2
Civil and Environmental Engineering, Idaho State University, 921 South 8th
Avenue, 83209, Pocatello, ID; PH (806) 401-4845; email: soreandr@isu.edu

ABSTRACT

Bridges are vital to any transportation infrastructure and have the potential to highly
impact an economic network if damaged and/or inaccessible. One of the major
components in a bridge system is the bridge pier, which has to support several
loading scenarios. This study focuses on the reliability of a reinforced concrete bridge
pier subject to vehicular impact loading, which is designated as a hazardous load.
Hazardous loads are typically high-intensity loads that are short in duration. These
types of loads are receiving increased attention due to recent occurrences and their
ability to initiate structural failure. To study the reliability of the bridge piers under
this loading, Monte Carlo simulation is used in a first-order, second-moment
reliability analysis on several column resistance and impact event combinations. The
reliability analysis is carried out for five different vehicle classes to represent the
likely vehicles to participate in the impact event. Additionally, sensitivity analysis is
carried out to identify the factors that most contribute to high probabilities of failure:
reinforcement area, vehicle speed, and vehicle mass. Fragility curves are generated to
provide a graphical representation of the sensitivity analysis.

INTRODUCTION

Bridges are a vital part of any transportation infrastructure, especially in the United
States’ federal interstate system. Bridges have the potential for hazardous loadings,
are frequently utilized, and can result in negative economical consequences if
removed from service. Removal of a bridge, even if only temporarily, from a regional
transportation system can cause disruption to the network resulting in potentially
large economic losses. Bridge piers are important load-bearing components of a
bridge system, and local failure of these individual components could lead to
progressive system failure. The necessity to provide economic designs with
increased safety capacities against limited-duration, high-intensity loads, (e.g.
seismic, hurricane, impact, etc.) is becoming a serious focus in structural design due
to recent occurrences.

Vulnerability, Uncertainty, and Risk

 Vulnerability, Uncertainty, and Risk ©ASCE 2014 2703

To assess the safety of a structure, engineers attempt to quantify the risk of damage or
failure associated with the application of various loads to the structure through
reliability analysis. One type of hazardous loading that bridges are susceptible to is
vehicle impact. Bridges are vulnerable to vehicular impact attacks due to the
accessibility by the general public and limited surveillance. Hazardous loads are not
typical or predictable and impose a force of great magnitude on the structure in a
short duration of time. Vehicle impact can be categorized as accidental (i.e. the result
Downloaded from ascelibrary.org by Brigham Young University on 11/18/14. Copyright ASCE. For personal use only; all rights reserved.

of traveler or vehicle error) or intentional (terroristic with intent to disrupt a

transportation network). Intentional vehicle impact is of special interest because the
accelerations and masses of the vehicles are typically higher. The objective of
terrorism is to cause chaos or disruption, and as such, the financial impact of a failed
or damaged bridge can be significant.

This study determines the reliability of a reinforced concrete bridge pier that
experiences intentional vehicular impact. By better understanding the reliability of
reinforced concrete bridge columns subject to this loading, future bridge designs can
be improved upon and reduce the risk of disruption to transportation networks.

According to the U.S. Federal Highway Administration, vehicle impact is the third
leading cause of bridge damage and/or failure (Agrawal & Chen, 2008). In a
vehicular impact event with a bridge pier, damage simultaneously occurs in both the
vehicle and the reinforced concrete bridge pier (El-Tawil, 2004). El-Tawil defines
this type of impact event as a “soft” impact, which allows for the impact loading to be
treated as a static loading due to similar failure mechanisms.

In 2008, Cizmar and others studied the reliability of rectangular concrete columns
under vehicle impact loading (Cizmar et. Al., 2008). This study provides useful
information about probabilistic and deterministic modeling of reinforced concrete
columns subject to vehicle impact. However, their study only investigates square
columns, whereas in the United States, round or circular columns are more frequently
used. Additionally, their study considers only accidental impact or cases where the
vehicles were decelerating to avoid crashing. This study builds upon the previous
work by considering: 1) circular instead of square columns, 2) terroristic intention
(i.e. acceleration vs. deceleration), 3) sensitivity analysis to determine those design
factors that most contribute to high probabilities of failure and as such can be
modified to decrease probabilities of failure.

METHODOLOGY

Quantifying the risk of damage or failure associated with the application of various
loads to a structure allows an engineer to assess the safety and ability of a structural
element. A structure’s ability to resist structural damage and/or failure depends on
both the strength of the structure and the different load combinations experienced by
the structural component (Ayyub & McCuen, 2011). In general, the resistance of the
component must exceed the loading to prevent failure. In this study, a reliability

Vulnerability, Uncertainty, and Risk

 Vulnerability, Uncertainty, and Risk ©ASCE 2014 2704

analysis is performed to determine the probability of failure of bridge piers exposed

to vehicular impact under varying magnitudes.

To achieve the objectives of this study, a numerical model using Monte Carlo
simulation is created in the software package MATLAB to perform the reliability
analysis utilizing the first-order, second-moment method. A simulation method,
rather than physical testing, is used because it allows for the manipulation of a
Downloaded from ascelibrary.org by Brigham Young University on 11/18/14. Copyright ASCE. For personal use only; all rights reserved.

representative system rather than a real system, which is ideal for sensitivity analysis.
The Monte Carlo method is a technique used to simulate random numbers and
formulate results using randomly generated values without any or limited physical
testing (Collins & Nowak, 2000). The steps involved in the Monte Carlo simulation
are: 1) system definition, 2) random variable identification, 3) random number
simulation, and 4) data analysis.

Using the simulation, fragility curves are developed to provide a graphical

representation of the probability of failure of the column due to column impact as a
function of impact load correlated to different vehicle types and velocities. A general
limit state equation, also known as a performance function, is the resistance minus the
load effect (Ayyub & McCuen, 2011). The general limit state equation for the bridge
column is defined in Equation (1):

M rd − M e = 0 (1)

where: the resistance, Mrd, is identified as the bending strength of the concrete bridge
column, and the load effect, Me, is identified as the bending moment due to vehicle
impact.

The analysis is limited to bridge piers of circular cross-sections with spiral

reinforcement, which is assumed to be adequate to support against shear. For
simplicity, eight evenly spaced vertical reinforcing bars are specified for the vertical
reinforcement in the bridge pier. The MATLAB program considers these conventions
and allows for the user to input the column length, column radius, vertical bar size,
and impact velocity.

The resistance bending moment equation of columns is given by Equation (2)

(Cizmar et. Al., 2008):

" y y %
M rd = λM $ As2 ⋅ f y ⋅ y2 + 2As1 ⋅ f y ⋅ 2 + N Ed ⋅ 2 ' (2)
# 2 2&
where: λM is the resistance uncertainty and is taken as one in the resistance equation,
As1 and As2 are the steel reinforcement areas in the column, fy is the steel yield
strength, y2 is the moment arm, and NEd is the non-factored axial load. Both resistance
and loading portions of the limit state equation have at least one random variable for

Vulnerability, Uncertainty, and Risk

 Vulnerability, Uncertainty, and Risk ©ASCE 2014 2705

the Monte Carlo simulation. The resistance random variable is the yield strength of
the reinforcing steel, which has a lognormal distribution.

The bending moment equation due to an impact load is given by Equation (3)
(Cizmar et. Al., 2008):
Downloaded from ascelibrary.org by Brigham Young University on 11/18/14. Copyright ASCE. For personal use only; all rights reserved.

M E = h ⋅ k ⋅ m ⋅ (v2 − 2 ⋅ a ⋅ r ) (3)

where: h is the height of the impact force, k is the vehicle stiffness, m is the vehicle
mass, v is the vehicle speed, a is the vehicle acceleration, and r is the distance from
the original vehicle path to the point of impact. As previously discussed, the study by
Cizmar and others (2008) consider an accidental impact event on the bridge pier,
while the main focus of this study is a terroristic, intentional impact event. For this
reason, the variable a is taken as the acceleration, rather than a deceleration, and
inserted into the equation as a negative value to make the adjustment.
With the resistance and loading moment equations defined, the performance function
for this analysis is developed and shown in Equation (4). The loading random
variables are the vehicle mass and velocity, which have normal and lognormal
distributions, respectively.
" y y %
λM $ As2 ⋅ f y ⋅ y2 + 2As1 ⋅ f y ⋅ 2 + N Ed ⋅ 2 ' − h ⋅ k ⋅ m ⋅ ( v 2 − 2 ⋅ a ⋅ r ) = 0 (4)
# 2 2&

The reliability analysis is carried out for five classes of vehicles, shown in Table 1, in
order to consider the significance of vehicle type on the failure probability of the
bridge pier subject to the impact event. For simplicity, the vehicle stiffness and
acceleration and the distance from original vehicle path to the bridge pier are held
constant amongst all five classes of vehicles. The vehicle masses and impact heights,
however, are adjusted according to the vehicle class and are also shown in Table 1
(US Department of Energy, 2012). The velocity is also taken as constant amongst the
five classes and the MATLAB program allows for the user to enter in the desired
velocity for the analysis.

Table 1. Vehicle Class, Mass, and Impact Height

Vehicle Mass (kg) Impact Height (m)

Car 2,270 0.61
SUV/Van 4,540 0.91
Small Moving Van/Delivery Truck 8,170 1.22
Moving Van/Water Truck 11,800 1.5
Semi-Truck/Trailer 19,100 1.8

Table 2 defines the variables used in the analysis, along with their means and
standard deviations.

Vulnerability, Uncertainty, and Risk

 Vulnerability, Uncertainty, and Risk ©ASCE 2014 2706

Table 2: Variable Definition and Parameters

Standard Deviation
Variable Units Mean
(% of mean)
λM - 1 10%
As1 cm2 User Input 5%
Downloaded from ascelibrary.org by Brigham Young University on 11/18/14. Copyright ASCE. For personal use only; all rights reserved.

As2 cm2 User Input 5%

fy kN/cm2 50 5%
y2 cm User Input 5%
N kN User Input 10%
k kN/m 300 20%
m kg See Table 1 33%
v kmh/h User Input 15%
a m/s 5 25%
r m 23 10%
h m See Table 1 10%

Using the equations and parameters defined previously, a Monte Carlo simulation is
completed for the mass of each vehicle class, vehicle speed, and the steel yield
strength. The reliability index (β) is the inverse of the coefficient of variation and is
determined using Equation (5). The reliability model computes β for various
scenarios and the different vehicle classes. The corresponding probability of failure
(Pf) is defined in Equation (6). Applying Equations (5) and (6), the probabilities of
failure corresponding to each vehicle type are calculated for the specified column and
impact event.

β= c
σz
(5) Pf = 1− φ (β ) (6)

where: µz is the mean of the performance equation, σz is the standard deviation of the
performance equation, and ϕ is the standard normal cumulative distribution function.

Using the simulation procedure and equations defined previously, several column
resistance and impact event scenarios are considered and sensitivity analysis is
carried out for the purpose of identifying the most influential parameters of the
impact event causing failure as defined by the limit state equation. Collins and
Nowak (2000) identify a general procedure for performing sensitivity analysis as: 1)
system definition and possible scenario identification, 2) scenario reliability
determination, and 3) most sensitive parameter identification.

RESULTS

To study the effects of the size of reinforcement and vehicle speed on the reliability,
bar sizes between 3 and 11 are input into the model holding the column radius (0.5

Vulnerability, Uncertainty, and Risk

 Vulnerability, Uncertainty, and Risk ©ASCE 2014 2707

meters) and height (5 meters) constant for various vehicle speeds. Figures 1 and 2
show the fragility curves of impact events with the SUV and semi-truck vehicle
classes with increasing vehicle speed versus reinforcement ratio (ρ) (where ρ is the
ratio of steel reinforcement area to concrete cross sectional area).
Downloaded from ascelibrary.org by Brigham Young University on 11/18/14. Copyright ASCE. For personal use only; all rights reserved.

Figure 1: Probability of Failure with Reinforcement Ratio and Different Vehicle

(SUV) Velocities

Evaluation of these curves demonstrates that the probability of failure of the bridge
pier decreases as ρ increases; which is expected. By increasing the reinforcement
ratio, the probability of failure can be reduced by up to 60% for the SUV class and
36% for the semi-truck class. The probability of failure is much higher with the larger
mass vehicle than the smaller.

Figure 2: Probability of Failure with Reinforcement Ratio and Different Vehicle

(Semi-Truck) Velocities

Vulnerability, Uncertainty, and Risk

 Vulnerability, Uncertainty, and Risk ©ASCE 2014 2708

In addition to the parameters held constant in this analysis, the column diameter was
held constant (2 meters) to determine a probability of failure increase by up to 171%
from a smaller vehicle class to the next larger vehicle class (e.g. from Car to
SUV/Van or from Delivery Truck to Moving Van). This should be considered for
bridge piers accessible to high mass vehicles. Another conclusion drawn from the
curves is vehicles travelling at higher speeds result in up to a 9% increase in failure
Downloaded from ascelibrary.org by Brigham Young University on 11/18/14. Copyright ASCE. For personal use only; all rights reserved.

probability. Again, this is a significant increase in the probability of failure, which

should be considered on bridges that usually contain traffic travelling at high speeds.

To study the effects of the column diameter and axial load on the reliability, various
column diameter values were input into the model holding the column radius height
(5 meters), reinforcement ratio (0.0029), and vehicle speed (130 kilometers/second)
constant for various axial loads. Figure 3 shows the fragility curves of an impact
event with the SUV vehicle class with increasing axial load, as a percentage of the
allowable axial load, and column diameter.

Figure 3: Probability of Failure with Changing Diameter for Different Axial

Loads (SUV)

From these curves, it can be concluded that the probability of failure is not
dramatically affected by the change in axial load. Though it can be reduced by up to
90% with an increase in axial load, the failure probabilities are so small it is not
relatively significant compared to the other parameters considered. The same is true
for the increase in column diameter. The highest decrease in probability of failure
(about 97%) occurs when increasing the diameter from one meter to 1.5 meters, but
these values are so small they are not significant.

Vulnerability, Uncertainty, and Risk

 Vulnerability, Uncertainty, and Risk ©ASCE 2014 2709

CONCLUSIONS

The reliability of reinforced concrete bridge columns subject to vehicular impact is

considered in this research study. Monte Carlo simulation is used in the reliability
analysis, which provides data for a sensitivity analysis to identify the factors that
most contribute to high probabilities of failure. The investigation led to the
identification of the following column resistance factors, in descending order, that
Downloaded from ascelibrary.org by Brigham Young University on 11/18/14. Copyright ASCE. For personal use only; all rights reserved.

most likely contribute to failure: 1) reinforcement ratio 2) column diameter and 3)

axial load in the column. The investigation also led to the identification of the
following impact loading factors, in descending order, that most likely contribute to
failure: 1) vehicle mass and 2) vehicle speed.

Current countermeasures are typically intended to slow the vehicle or increase the
distance between the roadway and column. According to the results, this method of
reducing the accessibility and vulnerability of the bridge pier lowers the probability
of failure by a substantial amount. Design countermeasures to reduce the probability
of failure include increasing the reinforcement ratio by either adding more vertical
reinforcing bars or increasing the bar size and increasing the diameter of the column
radius to provide a larger gross area of concrete. Another countermeasure is to
provide protection surrounding the bridge pier, such as barriers, to reduce the
accessibility and vulnerability of the bridge piers. A potential continuation of this
study would be to determine the effectiveness of design countermeasures versus
determent countermeasures.

REFERENCES

Agrawal, A.K. & Chen, C. (2008). Bridge vehicle impact assessment. University
Transportation Research Consortium New York State Department of
Transportation, 1-28.
Ayyub, B.M., Kaminskiy, M., & McGill, W.L. (2007). Critical asset and portfolio
risk analysis: an all-hazards framework. Risk Analysis, 27, (4) 789-801.
Ayyub, B.M. & McCuen, R.H. (2011). In Probability, statistics, and reliability for
engineers and scientists (3 ed.). Florida: Taylor & Francis Group.
Cizmar, D., Mestrovic, D., & Miculinic, L. (2008). Reliability of concrete columns
under vehicle impact. Structures Under Shock and Impact X, 98, 157-165.
Collins, K.R. & Nowak, A.S. (2000). Reliability of Structures. Boston: McGraw-Hill
Higher Education.
El-Tawil, S. (2004). Vulnerability of bridge piers to impact by heavy vehicles.
Simulation Technology, 2, (7)13-20.
Faber, M.H. (2001). Probabilistic model code. JCSS Publication.
Notteboom, T. & Rodrigue, J.-P. (2012). Transportation and economic development.
The Geography of Transportation Systems.
US Department of Energy (2012). Maps and Data. Alternative Fuels Data Center.

Vulnerability, Uncertainty, and Risk