In This Issue
Transportation Infrastructure
June 1, 2008 Volume 38 Issue 2
Transportation Infrastructure issue of The Bridge, Volume 38, Number 2, Summer 2008

The Safety of Bridges

Sunday, June 1, 2008

Author: Theodore V. Galambos

Bridges are part of a city’s transportation system, but also part of its distinctive architectural and aesthetic landscape.

At approximately 6:00 p.m. on August 1, 2007, the bridge carrying Interstate Highway I35W over the Mississippi River in Minneapolis collapsed, plunging rush-hour traffic some 35 meters down with the bridge or into the river. Many of the people in the cars, buses, and trucks survived the unexpected drop; however, 13 were killed and about 80 were injured. Rescue response was almost immediate, and many heroic acts were performed by the surviving passengers, police, and rescue workers, demonstrating that the community was well prepared for an emergency.

The collapse of this bridge was sudden, unexpected, and complete. The structure essentially disintegrated within seconds. Because the demise of the bridge was recorded by traffic management cameras, the whole world witnessed the catastrophe within minutes. The questions on the minds of everybody were how something like this could happen, whether it is safe to drive over any bridge, and who was responsible. These questions are entirely reasonable and must be answered by the bridge-engineering community to reestablish confidence in this part of our infrastructure. This essay is my personal response to the question of why bridges collapse and what we as engineers should do, beyond what is already being done.

Bridges define the character of a city. When I think of New York, San Francisco, London, Paris, Lisbon, or Budapest, it’s the bridges that come to mind. City residents are proud of their bridges, which not only provide transportation links, but are also part of their city’s architectural and literary heritage and the aesthetic landscape of civilization. A Nobel Prize in literature was once awarded for The Bridge on the Drina, a novel of the life of a bridge (Andric, 1977).

There are many thousands of bridges in the world, and there have been thousands more in the past, and it is hardly surprising that there have been partial failures and total collapses. But most of these occurred in the past or in other countries, and so the general American public has not worried about routinely crossing bridges. However, when a failure happens almost next door (the I35W Bridge is a 20-minute walk from my office at the University of Minnesota), we take note!

This is especially so if the failure comes out of the blue on a clear, warm summer day on a bridge that has done its job for half a century. Drivers on the Minneapolis bridge were often not even aware that they were on a bridge. A minute or so, and they were across. No wonder people were upset when this permanent, solid, dependable bridge suddenly gave up the ghost.

People often take great risks. Driving a car, crossing a street, smoking, overeating, even getting married, might have dire consequences. However, individuals voluntarily put themselves into these situations, even though they could have bad outcomes. The collapse of a bridge, however, is not a voluntary risk. It is, quite simply, not supposed to happen. Hardly anybody in the world perishes because of the collapse of a bridge.

In a pioneering book on the safety of structures, Pugsley (1966) relates that fighter pilots in WWII went up against the enemy even when there was a high chance they might be shot down. But these same pilots demanded design changes if the structural failure rate of their aircraft was more than 5X105 per flying hour. So to the public, the collapse of the Minneapolis bridge is a serious event that must somehow be explained.

Historical Perspective
The collapse and disintegration in a matter of seconds of a bridge that is judged to be sound is almost unheard of. Yet it does happen from time to time, and the memory of such an event lasts the lifetime of the generation that experienced it, even remotely through reports by the media. One such event, which has many similarities to the Minneapolis bridge failure, is described below.

On December 15, 1967, at Point Pleasant, West Virginia, the Point Pleasant Bridge over the Ohio River collapsed. My knowledge of this collapse comes from the final accident report of the National Transportation Safety Board (NTSB, 1970) and from my brother, Charles Galambos, who worked on the investigation. The bridge was constructed in 1929 and had been in service continuously for 38 years when the structure suddenly collapsed during the evening rush hour. Forty-six people died, nine were injured, and scores of vehicles fell into the river. This unfortunate collapse was very like the event in Minneapolis on August 1, 2007.

Sketch of the Point Pleasant Bridge
FIGURE 1 - Sketch of the Point Pleasant Bridge. Source: NTSB, 1970.

The Point Pleasant Bridge was a novelty in its time; few such bridges were constructed during its lifetime, and none after its collapse. The sketch in Figure 1 shows the general features of this hybrid structure, somewhere between a truss bridge and a suspension bridge (NTSB, 1970). In the middle of the central span, the tension in the suspension bars is counteracted by the compression in the top chord of the truss. The economic advantage of this type of bridge was that the two forces somewhat cancelled each other out.

Sketch of the eye-bar arrangement of the Point Pleasant Bridge.
FIGURE 2 - Sketch of the eye-bar arrangement of the Point Pleasant Bridge. Source: NTSB, 1970.

The suspension system consisted of sets of two eye-bars connected by pins at the junction between adjacent segments (Figure 2). The eye-bars were made of a new, tempered, high-strength steel. Failure was initiated by the brittle fracture in the eye-bar material at the first joint to the north of the tower on the Ohio side in the suspension structure, as shown in Figure 1. The causes of the failure were described in the NTSB report:


    . . . the cause of the bridge collapse was the cleavage fracture of the lower limb of the eye of eyebar 330 at joint C13N of the north eyebar suspension chain in the Ohio sidespan. The fracture was caused by the development of a critical sized flaw over the 40 year life of the structure as the result of the joint action of stress corrosion and corrosion fatigue.

Contributing causes are:

  1. In 1927, when the bridge was designed, the phenomena of stress corrosion and corrosion fatigue were not known to occur in the classes of bridge material used under conditions of exposure normally encountered in rural areas.
  2. The location of the flaw was inaccessible to visual inspection.
  3. The flaw could not have been detected by any inspection method known in the state of the art today [1970, emphasis added] without disassembly of the eyebar joint.

This disaster was a wake-up call for the bridge-engineering community, and the design, construction, inspection, and maintenance of bridges changed radically as a result. Biannual inspections and material fracture toughness requirements were mandated, as well as other changes over the years, especially related to fatigue and brittle fracture. The change of most interest for this essay was the requirement that a bridge be robust, that is, that it should not totally collapse when a local joint or member fails. In the jargon of bridge-design standards, the system must not be prone to “progressive collapse.” The forces that a failed part is designed to carry must be able to be rerouted to another path. In other words, the structure must be “redundant.”

Eye-bars of the Budapest Chain Bridge
FIGURE 3 - Eye-bars of the Budapest Chain Bridge.

If the Point Pleasant Bridge had been built with many eye-bars in each chain link, like the Budapest Chain Bridge (Figure 3), it would still be in service today, with fractured eye-bars replaced as needed. Thus the cause of failure was not the fracture of the eye-bar per se, but an error in judgment during design.

Causes of Bridge Failures
A bridge is all structure. Unlike a building, which has walls, stairs, elevator shafts, slabs, and other features that usually add uncounted strength and stiffness to the structural system, bridges have no hidden sources of strength. Every part of a bridge is there for structural purposes of strength and stiffness.

Most bridge failures occur during construction, when the structure is most vulnerable. Also during construction, many problems of undesirable bridge behavior are ironed out. Once a bridge is completed, one can confidently expect that the structure will last for its intended life span and perform its intended job. Although there are rare collapses during the service life of bridges, most of these are partial collapses, and most are discovered in good time and repaired.

Cable-stayed bridge over the Savannah River
FIGURE 4 - Cable-stayed bridge over the Savannah River.

The list that follows describes a few of the many possible causes of bridge collapse:

  • Previously unimagined effects of natural forces, such as the force of the Loma Prieta earthquake that caused the partial collapse of one segment of the traffic lane on the San Francisco-Oakland Bay Bridge. The span collapsed as a result of out-of-phase movements of supports as the seismic wave traveled along the bridge. Design standards now require specific measures to ensure structural continuity between the bridge superstructure and its supports.
  • Deliberate destruction in war. Most of the bridges in Central Europe were destroyed during the Second World War. In the long run, this was beneficial for the future of bridge engineering, because the reconstruction of those bridges was a veritable renaissance in the art of bridge building. Out of economic necessity, new architectural forms evolved, such as plate-girder bridges, cable-stayed bridges (Figure 4), and prestressed concrete bridges. Since then, this rejuvenation has moved from Europe to America to Japan and presently to China.
  • Carelessness or accidents. An example of a failure caused by carelessness is the destruction of the original Sunshine Skyway in Tampa, Florida, which was caused by the collision of a ship with one of its piers. Since then, piers in waterways have been surrounded by strong protective barriers.
  • Fatigue and brittle fracture of structural members and connections. This type of failure is frequently caused by a misunderstanding of the details of connections or by the use of incorrect materials.
  • Decay from cracking or corrosion, often indicating improper maintenance and inspection. One of the major responsibilities of bridge inspectors is to watch for such deterioration. Although we know how to design to prevent fatigue and brittle fracture, cracking and corrosion are normal consequences of aging and should be constantly watched for and repaired when discovered.
  • Unfamiliarity with new materials, details, and structural systems at the time of design. This was one of the problems with the Point Pleasant Bridge discussed above.
  • Unseen hazards. The most recent concern is about scouring at the bases of bridge piers where they interface with the riverbed. This problem is currently thought to be the most common cause of bridge failure and is an area of ongoing research.
  • Unknown hazards, the rarest and most dangerous causes of collapse. One of the most complex issues is applying the principles of structural engineering to situations for which designers have no previous experience or are not aware of research done elsewhere. For example, the cause of the failure of the Firth of Tay Bridge in Scotland on December 28, 1979, during a great windstorm was that the designers had not considered the effects of wind forces on the structure (Prebble, 1956). In the case of the collapse during construction of the Quebec Bridge over the St. Lawrence River on August 29, 1907, the designers had neglected one term in a differential equation of column strength; the effect was negligible for columns of solid cross section but significant for the type of latticed, open-section columns used on this bridge (Government Board of Engineers, 1919). The West Gate Bridge in Melbourne, Australia, collapsed on October 15, 1970, for a variety of reasons (Royal Commission, 1971); research by Professor Noel Murray of Monash University in Melbourne showed that the designers had applied the prevailing theory of the behavior of stiffened panels welded from thin steel plates beyond its area of applicability.
Cartoon of the evolution of bridge design
FIGURE 5 - Cartoon of the evolution of bridge design.

Consequences of Bridge Disasters
Every bridge failure has provided builders and engineers with an opportunity to do better next time, to apply the lessons learned to the next design, and to retrofit existing structures. This process is illustrated in the cartoon in Figure 5. When a new configuration of bridge geometry or a new combination of materials is used for a bridge, designers proceed conservatively. If they are successful, they are encouraged to be less and less conservative over time, until they run out of the elbow room provided by the safety factor, which can result in failure of the system.

Society reacts to these failures either by being more conservative in the next design or by judging the type of bridge to be unsuitable. As confidence is regained, the cycle then starts up again, this time buttressed by research, and eventually a good balance is reached between safety and economy.

The classic example is the stiffened roadway on suspension bridges, which became less and less stiff with each new bridge, until the Tacoma Narrows disaster. Research and an understanding of aerodynamic phenomena have now progressed to the stage at which designers all over the world are designing suspension bridges with longer and longer spans.

This scenario has been repeated many times over for other types of bridge structures. Current bridge art is based on the cumulative experience of designers, planners, fabricators, and builders, and on collaboration with researchers in academic, governmental, and industrial organizations and laboratories. Although much has been learned from studying the causes of bridge failures and from research conducted after catastrophes, it would be far better to conduct the research first and avoid the bridge collapses altogether.

Back to Minneapolis
If we have the experience with science, theory, design, and construction techniques, as well as thoroughly researched studies of past failures, why did the bridge in Minneapolis fail? This particular bridge was watched over by inspectors and engineers from the Minnesota Department of Transportation, was thoroughly examined frequently by a consulting firm, and was studied several times in the last decade by graduate students under the guidance of structural engineering professors at the University of Minnesota. To use a medical analogy, the patient died of unknown causes in the hospital with medical specialists standing around the bed studying the symptoms. The precise causes of this collapse are still under investigation by NTSB, and no definitive conclusions have been published at this time (April 2008).

The following is my guess about what happened. It seems now (in April 2008) that the investigators and inspectors somehow missed the fatal points of weakness. A few days after the disaster it was clear from preliminary examinations of the wreckage that one particular node, repeated symmetrically in four places in the trusses of the center span, had been constructed with insufficiently thick “gusset” plates, which are riveted to the members that enter the joint, thus holding them together to form a node in the two trusses. The gusset plates were reported by NTSB to be half as thick as they should have been.

The deck of the bridge was being replaced at the time of collapse, and construction material and equipment were located in the lanes where the replacement work was being done. The resulting extra force on the four presumably half-strength joints could have caused one or more of the gusset plates to fracture, yield, and/or buckle. Although the structure had some redundancy, there was probably no time for the redistribution of forces, or else the critical joints were weakened enough to destabilize the center span. The suddenness of the collapse shows how quickly gravity takes over (think how quickly you find yourself on the ground after slipping on an icy sidewalk). Thus the collapse can be simply explained by the loss of stability caused by overload and under-strength. Gravity took care of the rest! The NTSB report will surely provide specific details and scenarios of other possible contributors to the failure.

Lessons from Minneapolis
As long as there are humans in the world, there will surely be bridge failures in the future, although they will be rarer because of the lessons we have learned. Instructions from government agencies to bridge owners are now being formulated, and the NTSB’s recommendations will reduce this tiny probability even more. Nevertheless, whatever we do, there is always a remote chance that we will miss something. Recommendations for the future will certainly include the following measures:

  • Older bridges and larger bridges should be under observation by experienced bridge engineers who have years of knowledge with the bridges under their supervision. The original designs of these bridges should be periodically examined and reanalyzed using state-of-the-art computer methods and current design codes.
  • The design calculations, drawings, and contract documents for new bridges should be reviewed carefully by experienced bridge engineers who were not involved in the original design. The bridge documents should also be “peer reviewed” by an agency not associated with the original design or the bridge authority that owns the structure.
  • Robustness and redundancy should be incorporated into the bridge design as a matter of course to counter-act unexpected and unimagined hazards, especially the possibility of progressive collapse.
  • The design and construction teams should conduct a “hazard scenario” exercise to reveal possible problems prior to the start of construction.

Many new devices and instruments for monitoring critical parts of bridges, such as accelerators that can reveal changes in the dynamic signature, strain gauges and deflection gauges, and many types of cameras are now available or are being developed. Other, even exotic, high-technology devices may also become available. Data from these instruments can be electronically transmitted, stored, and finally evaluated by inspectors and experienced engineers. No doubt more monitoring schemes will be installed as a consequence of the Minneapolis bridge collapse. However, it will require sustained discipline to store and evaluate data for the life of bridges, for the large majority of which no noteworthy event will ever be observed.

Andric, I. 1977. The Bridge on the Drina. Chicago, Ill.: University of Chicago Press.
Government Board of Engineers. 1919. The Quebec Bridge over the St. Lawrence River. Printed by order of the Governor-General in Council, Ottawa, Canada.
NTSB (National Transportation Safety Board). 1970. Collapse of U.S. 35 Highway Bridge, Point Pleasant, West Virginia, December 15, 1967. NTSB-HAR-71-1. Washington, D.C.: NTSB. Available online at
Prebble, J. 1956. The High Girders. London, U.K.: Secker and Warburg.
Pugsley, A. 1966. The Safety of Structures. London, U.K.: Edward Arnold Publishers.
Royal Commission. 1971. Report of the Royal Commission into the Failure of Westgate Bridge. Melbourne, Australia: C.H. Rixon, Government Printer.

About the Author:Theodore V. Galambos is Emeritus Professor, Department of Civil Engineering, University of Minnesota, and an NAE member.