In This Issue
Summer Bridge: A Vision for the Future of America’s Infrastructure
June 15, 2018 Volume 48 Issue 2
The articles in this issue, by academic and industry experts, focus on what’s needed to prepare US infrastructure systems for the coming decades.

Are Our Bridges Safe?

Thursday, June 14, 2018

Author: Andrzej S. Nowak and Olga Iatsko

Transportation, including the road network, is a very important part of the national economy, providing necessary connections between people, business, and industries. The US Interstate Highway System was initiated by President Eisenhower in the 1950s to benefit commercial and military transportation. Its nearly 50,000 miles were largely completed in 35 years and are now part of over 4 million miles of roads across the country (FHWA 2015). Bridges increase the efficiency of the road network operation and are a vital component of the nation’s transportation system and economy. Loss of a major bridge can have national impacts.


Half of the country’s bridges are owned and administered by the states, the other half by counties and cities; very few are privately or federally owned. State-owned bridges are generally in better shape and regularly inspected (at least once every two years); counties and cities often lack sufficient funds for inspections, maintenance, and repairs.

Although the percentage of bridges that are in poor condition decreases from year to year, this is mostly because of an increase in the number of newly built structures rather than improved maintenance. The percentages of structurally deficient bridges around the country range from 1.6 percent (Texas and Nevada) to 23.3 percent (Rhode Island) (FHWA 2017). The variation is due to differences in state policies regarding the inspection, maintenance, and repair of bridges.

The National Bridge Inventory (NBI[1]) was created to support the use of data about the country’s 614,387 bridges on public roads, 25 percent (145,104) of which are part of the National Highway System (NHS). With data collected in accordance with the National Bridge Inspection Standards[2] and submitted annually to the Federal Highway Administration (FHWA), the NBI includes basic information about every bridge—its dimensions (span length, roadway width), material, structural type, location, maintenance plan, repair history, and traffic.

The availability of bridge failure data, however, is rather limited. A bridge fails when it cannot perform its function, for example because of excessive vibrations or deflection, cracking of concrete, fatigue cracking of steel, or, in drastic cases, collapse of components or the whole structure. A 25-year review of data found that bridge failures occur with an annual frequency of approximately 1 per 5,000, but, because of incomplete or incorrect information about bridge failures reported by state DOTs, the actual failure rate is significantly higher (Cook et al. 2015).

About half of collapsed bridges were structurally deficient as a result of age, excessive loads, extreme weather, inadequate maintenance, and other aspects. Continuous corrosion and fatigue can lead to loss of the load carrying capacity and a major collapse.

The estimated total cost of US bridge repairs is $123 billion (ASCE 2017). To avoid the high costs of replacement or repair, bridge evaluation must be done at regular intervals and accurately assess load carrying capacity based on predicted loads and expected changes in capacity (deterioration).

What Are the Problems with Bridges?

Existing bridges are subject to aging, deterioration, corrosion, cracking, delamination, material fatigue, and chemical degradation. These may occur naturally over time or as a result of conditions such as traffic and weather events. Failures due to overload or deterioration are strongly age related.

The average age of US bridges is 43 years, and many were designed before 1970 for a service life of 50 years. Those bridges are thus nearing the end of their design life.

About 10 percent of all bridges have one or more deteriorated structural components (ASCE 2017). For example, an inspection of the Benjamin Franklin Bridge between Philadelphia and Camden, New Jersey, built in 1926, revealed that over 10 percent of the wires in the main suspension cables are corroded and broken (Weidlinger Associates 2000). Structural deterioration of materials and components accounts for 9 percent of bridge failures.

Growth in the volume and weight of truck traffic during recent decades is seriously affecting the long-term performance of bridges and increasing the need for maintenance. Bridge damage or failure is often due to extreme truck loading or collisions involving oversized and/or overweight vehicles. Many bridges were designed for loads that were specified years ago, and their design loads are now too small for the current traffic. In addition, with larger trucks on the roads, many bridges do not provide adequate clearance in width and/or height. Vehicle (or vessel) collision accounts for 7 percent of bridge failures.

About 60,000 bridges (10 percent) are posted for a weight or speed limit. Violation of the weight limit may lead to substantial damage or even collapse. But 80 percent of bridges with a posted weight limit are on local roads where truck loads are not properly monitored (if at all). Overweight vehicles account for 12 percent of bridge failures.

According to NBI data, about 60 percent of US -bridges are made of concrete, 30 percent are made of steel, about 3 percent of wood, and the rest from -other materials (masonry, aluminum iron, etc.). Steel beam/girder bridges are more prone to collapse than -other types. The causes are mostly extraordinary or extreme events that produce stress levels significantly exceeding the capacity of the bridge, especially -hydraulic disasters such as a flood or scour (erosion of the soil base under the foundations of piers or abutments), which account for over 50 percent of such events.

What Is the Role of Design Codes?

The main stakeholders in any construction are the -owners/investors and the users/occupants. The former are interested in keeping costs down and maximizing profits; the latter are interested in having a safe and functional structure. The owner/investor hires the designer and contractor, so they represent the owner/investor’s side. There may be a conflict of interest between keeping costs down and ensuring safety and functionality. The role of design codes is to balance these two conflicting interests.

AASHTO design codes (AASHTO 2017) for bridges specify the loads to be considered by the designer. The loads have to be conservative to provide a safety margin by using load factors. The codes also articulate procedures for selecting the type of structure and materials that will be sufficient to resist expected loads, again with a conservative safety factor. The determination of safety factors has evolved from one based on intuition to an advanced reliability-based code calibration (Nowak and Iatsko 2017).

According to AASHTO (2017), a bridge’s expected performance life is 75 years (Kulicki et al. 2007). The expectation can be expressed in terms of the probability of failure that is acceptable to society: if the probability is too high then the bridge may require expensive repairs or replacements, while a very low probability of failure can be prohibitively expensive to achieve. Therefore, the development of a design code depends on the answers to the following three fundamental questions:

  • How is bridge safety measured?
  • How is the level of required safety determined?
  • How is safety implemented?

How Is Bridge Safety Measured?

A bridge’s safety margin is the difference between load and resistance. Failure occurs when the load exceeds a bridge’s load carrying capacity or resistance.

But the loads acting on a bridge usually cannot be accurately predicted; they are random in nature. And the ability of the structure to resist loads depends on mechanical properties of materials (steel, concrete), connections, and dimensions that also cannot be predicted with certainty and are random in nature. Because load and resistance are random variables, the safety margin is also a random variable. The probability of failure, Pf, is the probability of load exceeding resistance. Safety, or reliability, is defined as 1 – Pf.

A structure can be in one of two states: safe performance or failure. The borderline between these two states is called a limit state and a mathematical formulation of the limit state is called a limit state function. Calculation of Pf requires knowledge of the limit state function and statistical parameters of load and resistance. However, it is convenient to measure safety in terms of the reliability index, â, defined as the ratio of the mean value and standard deviation of the safety margin. There are several methods—from simple formulas to Monte Carlo simulations—for calculating â, taking into account the type of distribution function, nonlinearity of the limit state function, and correlation between variances (Nowak and Collins 2013).

How Is the Level of Required Safety Determined?

Safety is a commodity and depends to a certain extent on the availability of resources. Target reliability levels depend on the location of a structure, its components, and costs associated with safety measures. Selection of the target reliability index depends mostly on two factors: the consequences of failure and the cost of safety.

Determining the Reliability Index

The consequences of exceeding a limit state can vary significantly. For example, the single passage of a heavy vehicle that results in a deflection larger than the limit may not create an immediate problem—but for a steel beam it can cause a permanent deformation or even collapse. Beams in flexure when overloaded typically show some signs of distress, such as cracking and large deflection, so the structure can be closed and/or evacuated before a collapse. However, excessive shear can occur without warning, as a brittle fracture. Similarly, an overloaded compression member can buckle without warning.

The target reliability indices (âT) in bridge design codes are different for beams and columns, depending on the expected failure scenario. The âT for the deflection limit state can be as low as 0 (which corresponds to 50 percent probability of failure) if the consequences of exceeding it are negligible.

In the AASHTO (2017) code, for a ductile mode of failure, such as loss of flexural load carrying capacity for steel and concrete beams, the âT is 3.5 and corresponds to a probability of failure of 0.02 percent. For a -brittle mode of failure, such as shear capacity of concrete beams or buckling failure of columns, the âT is 4.0 and corresponds to a probability of failure of 0.003 percent. With prestressed concrete, cracking caused by a very heavy truck can be a problem if it occurs too often; a single passage is generally not a concern and therefore a âT of 1.0 is sufficient and corresponds to a failure probability of 15 percent.

Calculating the Costs of Safety

The other factor to be considered when determining the target reliability is the cost of safety, which is a function of expected additional expenses or savings resulting from changes in the safety margin. How much can be saved by reducing the safety margin? How much does it cost to increase safety? If safety is cheap, a higher target reliability index can easily be justified; if the cost to increase safety is too high, a lower âT may be tolerated.

For example, the âT will be very different for newly designed bridge girders and for existing structures. The cost of increasing the safety margin for a structure that is still on the computer is relatively low: selecting a -larger steel beam from a catalogue may increase the total cost by a negligible amount. In contrast, increasing the load carrying capacity of an existing bridge can be very expensive as it may involve closing the structure to -traffic, bringing in equipment, extensive labor, and so on. The âT for newly designed bridge girders is 3.5 and for existing girders 2.5, corresponding to 0.02 percent and 0.62 percent probability of failure, respectively.

How Is Safety Implemented?

The safety margin is implemented through the design code, which specifies load values, and factors that support safety, as well as the required load carrying capacity (or resistance) and resistance factors. Design code provisions are based on available statistical parameters of load and resistance. The selection criterion for load and resistance factors is the requirement that the reliability index be not less than the target value.

The code assumes that the quality of workmanship is either good or average. However, a review of engineering practice shows that most failures are due to human error (other causes are extreme events such as fires, floods, hurricanes, tornadoes, earthquakes, collisions). Surveys of structural failures (e.g., Nowak 1986) indicate that half of the errors are in the design and the other half in the construction. The errors are associated with lack of understanding, miscommunication, neglected or inappropriate maintenance, and wrong construction procedures.

In addition to efforts to reduce human error, safety can be enhanced through the development of new materials that are durable, long-lasting, and economical; new technologies that allow for faster construction and minimal traffic obstruction; and new design procedures using advanced analytical tools. New sensor technology should be applied for diagnostics and monitoring of structural performance, with warning systems for signs of distress. And bridge traffic loads can be -better controlled by a new generation of weigh-in-motion devices that are accurate and reliable.

What Is the Future for Bridges?

Bridges can serve for over 75 years if they are properly built and maintained. New materials, technologies, design techniques, analytical methods, monitoring equipment and procedures, and sensors can and should be used to make bridges safer throughout their performance life. They must be complemented by efforts to ensure quality in design, construction, maintenance, and operation.

One of the major challenges is growth in vehicle size and weight. Recent traffic measurements indicate that over 20 percent of trucks significantly exceed legal load limits. Repeated passage of heavy vehicles can cause the fatigue of structural materials, resulting in more frequent repair/replacement—or collapse. Effective law enforcement and weigh-in-motion monitoring can prevent illegally overloaded vehicles from damaging roads and bridges.

For construction, there are significant new developments in materials, especially composites, but they require more research to assess long-term performance. New technologies allow for significant reduction in the time required for construction: structural components are built in the plant, delivered to the site, and put in place in a short time (e.g., overnight). However, time pressure can result in a higher probability of errors, resulting in tragic failures.

The threat of terrorist attacks points to the need to safeguard national infrastructure facilities. A -better approach is needed for risk assessment of highway -bridges, involving identification and sensitivity analysis of various failure modes and scenarios associated with terrorist acts. Structures should be assessed for vulnerability to terrorist attack, with recommended prevention procedures and damage control measures.


In general, US bridges are safe, but their continued -safety depends on rigorous quality control and adherence to design codes and guides at all stages:

  • planning and design—careful review, especially of new procedures;
  • construction—examination of materials and technologies, on-site inspections;
  • service—regular inspections and the performance of required maintenance;
  • operation—effective enforcement of laws and control of traffic loads to prevent illegally overloaded vehicles from damaging roads and bridges; and,
  • as needed, repairs, rehabilitations, and replacements.


The authors thank NAE editor Cameron Fletcher, who generously dedicated her time to review this paper. Her conscientiousness and valuable comments are appreciated.


AASHTO [American Association of State Highway and Transportation Officials]. 2017. AASHTO LRFD Bridge Design Specifications. Washington.

ASCE [American Society of Civil Engineers]. 2017. ASCE’s 2017 Infrastructure Report Card. Reston VA. Online at

Cook W, Barr PJ, Halling MW. 2015. Bridge failure rate. Journal of Performance of Constructed Facilities 29(3):04014080.

FHWA [Federal Highway Administration]. 2015. Highway Statistics 2015. Washington. Online at 2015/.

FHWA. 2017. Deficient bridges by highway system. National Bridge Inventory. Online at

Kulicki JM, Prucz Z, Clancy C, Mertz D, Nowak AS. 2007. Updating the Calibration Report for AASHTO LRFD Code. Project no. NCHRP 20-7/186. Washington: Transportation Research Board. Online at 20-07(186)_FR.pdf.

Nowak AS. 1986. Modeling Human Error in Structural Design and Construction. Proceedings of a Workshop Sponsored by the National Science Foundation. Reston VA: American Society of Civil Engineers.

Nowak AS, Collins KR. 2013 Reliability of Structures, 2nd ed. Boca Raton: CRC Press.

Nowak AS, Iatsko O. 2017. Revised load and resistance factors for the AASHTO LRFD Bridge Design Specifications. PCI Journal 62(3):46–58.

Weidlinger Associates. 2000. Reliability of the North Cable of the Benjamin Franklin Bridge. New York.




About the Author:Andrzej Nowak is a professor and the Elton and Lois G. Huff Eminent Scholar Chair, and Olga Iatsko is a graduate research assistant, both in the Department of Civil Engineering at Auburn University.