Lives and Deaths of Infrastructure

Authored by: Mohammed M. Ettouney , Sreenivas Alampalli

Infrastructure Health in Civil Engineering

Print publication date:  September  2011
Online publication date:  April  2016

Print ISBN: 9780849320408
eBook ISBN: 9781420003758
Adobe ISBN:

10.1201/b11174-4

 

Abstract

We explore in more detail the concept of bridges (or any other civil infrastructure) as an organism that is born, lives and dies (fails). We note that, similar to humans, there are three distinct phases of bridge life cycle: (a) its birth, (b) its normal exis-tence (life), and (c) its death (failure). We propose that ensuring health of bridges applies to all phases of their life cycles. We immediately define the main attributes of healthy bridge during all phases as follows:

A healthy bridge meets or exceeds stated performance goals

At reasonable costs

 Add to shortlist  Cite

Lives and Deaths of Infrastructure

3.1  Overview

We explore in more detail the concept of bridges (or any other civil infrastructure) as an organism that is born, lives and dies (fails). We note that, similar to humans, there are three distinct phases of bridge life cycle: (a) its birth, (b) its normal exis-tence (life), and (c) its death (failure). We propose that ensuring health of bridges applies to all phases of their life cycles. We immediately define the main attributes of healthy bridge during all phases as follows:

  • A healthy bridge meets or exceeds stated performance goals
  • At reasonable costs

Among performance goals we can state further that the bridge should experience no unwarranted service interruptions due to distress, and partial or total failure. A healthy bridge should also perform its stated functions, such as load-carrying specifications, in a safe and secure fashion. Near the end of the life cycle of the bridge, it is essential to control the decommissioning process, as opposed to abrupt failure/distress.

Overall, we aim at discussing categories that would enable decision makers to quantify different phases of life cycle of the bridge; we also explore different roles structural health monitoring (SHM)/structural health in civil engineering (SHCE) dur-ing life cycle of the bridge, and how those roles change as the bridge life span change. Understanding these details can enhance the optimal health of bridges during all the phases of its existence.

3.1.1  Enabling versus Triggering Causes

To simplify examinations of health concepts of bridges, we categorize causes that affect bridge (or any other type of structure) into enabling and triggering causes. The categorization was used by Wardhana and Hadipriono (2003) to explore bridge failures. We extend the definition into other phases of bridge life span (construction, service life, and ultimate failure). Table 3.1 shows a general comparison between the two types of causes.

3.1.2  This Chapter

This chapter explores the health of a bridge at different phases of its life cycle as shown in Figure 3.3. First the health of the bridge during the birth phase is de-scribed. The different issues regarding health during the life span of the bridge is then surveyed. The third phase, bridge failure, or decommissioning, is discussed next. In all life cycle phases the triggering and enabling causes of structural health are defined and explored. The many roles SHM methodologies and techniques play at different phases are discussed in details. Whenever possible, cost–benefit issues of different SHM roles are sited. The chapter ends with an in-depth survey of important bridge failure examples in the United States in the past one hundred years.

Table 3.1   Enabling versus Triggering Causes

Causes

Enabling

Triggering

Temporal behavior

Occurs over a long period of time (see Figure 3.1)

Occurs suddenly (see Figure 3.2)

Severity

Both can have severe effects on bridges

NDT role

Can be useful

Effective only before and after event

SHM role

Long-term monitoring; intermittent monitoring

Before, during and after event monitoring might require different strategies

Example of enabling cause: corrosion of bridge column. (Courtesy of New York State Department of Transportation.)

Figure 3.1   Example of enabling cause: corrosion of bridge column. (Courtesy of New York State Department of Transportation.)

Example of triggering cause: overload on a bridge. (Courtesy of New York State Department of Transportation.)

Figure 3.2   Example of triggering cause: overload on a bridge. (Courtesy of New York State Department of Transportation.)

3.2  Birth of Bridges

3.2.1  Ensuring Healthy Life

Healthy life of structures starts at “birth.” Several phases of birth include (a) preliminary planning, (b) analysis, design, and detailing, (c) construction, and (d) quality assurance/quality control (QA/QC). SHM/SHCE can play many roles during all of those phases. These roles are described next.

Organization of

Figure 3.3   Organization of chapter 3.

3.2.1.1  Preliminary Planning

Preliminary planning of a bridge includes numerous roles of SHM techniques and methods. Perhaps the most obvious are different decision making (DM) processes. Activities such as cost–benefit analysis, finance and economics implica-tions, and social studies must precede any bridge construction phase. Also an ex-haustive risk and life cycle studies must be done. Risk analysis would include dif-ferent uncertainties that the bridge might encounter during its life span. Life cycle analysis of the bridge is performed. Implications of different decisions regarding the bridge should be analyzed. Decisions on optimal attributes of the bridge can then be made. Those DM methodologies are discussed in detail in chapter 8.

Other SHM components also have roles in the preliminary planning phase. For example, soil explorations can affect decisions regarding the bridge, including the very viability of constructing it. STRID activities that make use of available data, such as soil exploration results can help in ensuring healthy life of the bridge, after it is built.

3.2.1.2  Design Phase

Analysis phase of new bridges can include STRID methods for optimal designs. Since STRID methods (chapter 6) need actual test results as an input, simulated test results might be used. Simulated results can be generated using similar pre-vious tests or any other appropriate technique. Another important measure can be to prepare base analytic models for future STRID/DMID activities. Such base analytic models can be for the global bridge structure, or for important local details. As future actual datasets start accumulating, those analytic models can be updated and improved to help in different future decisions and activities such as DMID, inspection, maintenance, and rehabilitation.

Design phase can utilize available test results (such as soil test results) to aid de-sign paradigms, whether it is the capacity/demand or performance-based designs. Also, the design of future maintenance, rehabilitation, retrofits, and decommission-ing activities has to be compatible with the overall bridge design. The designer should ensure that all such designs are optimal. A comprehensive life cycle analysis plan should be intertwined with the initial designs of the bridge. Such life cycle anal-ysis plans should be scalable, flexible, and easy to be adjusted to future events.

Details of the bridge SHM projects must be included with the initial structural de-tails. Such SHM details should be for (a) long-term monitoring, (b) intermittent mon-itoring, (c) one-time specialized in situ monitoring such as bridge load testing, and (d) laboratory testing, such as different NDT techniques. The bridge designer should also try to accommodate conventional visual inspection by using simple detailing as well as easy-to-inspect detailing such that other expensive options can be minimized.

3.2.1.3  Construction Phase

Innovative uses of SHM during construction phase of large civil infrastructure can produce many dividends. An example of such innovations was reported by Ni et al. (2008). The subject of the project is the Guangzhou New Television Tower, which is located in Guangzhou China (see Figure 3.4). It is a 610-m high structure that is constructed of a mix of reinforced concrete and concrete-filled steel tubes. The Hong Kong Polytechnic University designed the SHM project with two major goals: (a) to monitor in-construction performance, and (b) to monitor in-service perfor-mance. Both goals are done in real time. In addition, the SHM project has other in-novative features such as (a) a modular design of the SHM system, (b) efficient wireless-sensing array, (c) advanced fiber optics sensing that is based on Bragg grating technology, and (d) structural identification processing using static and dy-namic monitoring data. The project also has a sophisticated system for data valida-tion and benchmark case studies.

What interests us in this section is the implementation of a during-construction SHM system by Ni et al. (2008) (see Figure 3.5). They used a total of 527 sensors that were placed at 12 cross-sections of the tower to monitor performance during construction. The choices of the sensor locations were obtained by analyzing the structure during construction process and placing them at sections with higher stress levels. Most of the sensors were strain sensors (total of 416) to monitor strains, creep, and shrinkage (vibrating wire strain gauges). In addition, large num-ber of temperature sensors was used (total of 96). Some additional weather, wind speed, inclination, leveling of floors and tower, and displacement sensing were uti-lized. The use of these SHM sensors enabled the project construction engineers to ensure higher quality and safe construction process of the tower.

Guangzhou New Television Tower. (With permission from John Wiley & Sons Ltd.)

Figure 3.4   Guangzhou New Television Tower. (With permission from John Wiley & Sons Ltd.)

In-construction SHM sensor placement for Guangzhou New Television Tower. (With permission from John Wiley & Sons Ltd.)

Figure 3.5   In-construction SHM sensor placement for Guangzhou New Television Tower. (With permission from John Wiley & Sons Ltd.)

We should also note that one of the innovations of the SHM project was to inte-grate the in-construction monitoring with a long-term in-service monitoring. In addi-tion, 280 sensors were used for the in-service SHM task. Of these, arrays of fiber optics Bragg grating sensor network were installed in permanent steel conduits. A total of 120 Bragg grating strain and temperature sensors were utilized for this task.

3.2.1.4  Quality Control/Quality Assurance (QA/QC)

Bridges, as well as all other major civil infrastructure, involve complex and lengthy construction process. The importance of performing construction tasks in accor-dance with the design and detailing specifications is exemplified in a usually rigor-ous and detailed QA and QC strategies. SHM methods are emerging as an accurate and inexpensive construction phase QA/QC device. Simple monitoring of different attributes such as concrete temperature, humidity, ambient temperature, and/or member alignment, can help in QA/QC processes. Consider, for example, the SHM experiment by Hansen and Surlaker (2006). They aimed at coupling the maturity principle of concrete during the curing phase and real time SHM activities in provid-ing real time and accurate QA/QC information that can save costs, and ensure life cycle safety of the system. The maturity principle simply relates the curing time and the curing temperature to the concrete strength. Thus, by knowing the elapsed time, and concrete temperature, it was shown that an accurate estimate of concrete strength is possible. Two important reasons as to why accurate estimation of cured concrete strength is important: (a) it permits the removal of temporary forms when the concrete reaches postulated minimum strength, and (b) it signals the optimal time for post-tensioning in post-tensioned systems. The authors used a set of em-bedded radio frequency ID (RFID) sensors in two pilot structures: a parking garage, and a concrete highway pavement. They reported the following major benefits for their technique.

  • In situ estimation of concrete strength eliminates the need for using laboratory cylinder tests, thus saving time and cost, while ensuring an even more ac-curate strength estimation.
  • When used for a post-tensioned system, an online estimation of concrete strength would reveal the exact time when the concrete strength would per-mit the post-tensioning process. This optimizes the time utilization during construction, while ensuring safe post-tensioning operations.
  • By placing RFID sensors across the depth of the pavement, it was possible to monitor the temperature gradient across the depth. This information can help predict any curling of the pavement; such curling can lead to a prema-ture surface fatigue cracking and reduction of the service life of pavements.

Several other benefits of the experiment were reported by Hansen and Surlaker (2006). The authors provided estimates of cost savings that the use of their tech-nique can provide the owners during the construction phase. We observe that there are additional savings that can be had during the life span of the system. Those additional savings result from factors such as improvements to fatigue strength limits, and reduced maintenance costs (by reducing or eliminating early tensile pavement cracking). It is clear that such innovative, yet simple and inexpensive use of SHM technologies can help in QA/QC of bridges during their construction phase.

One of the applications where NDT is very useful is quality control and assurance of bridges and pavements during construction. Garg et al. (2008) described a porta-ble seismic property analyzer (PSPA) used to determine the modulus of concrete pavement slabs at FAA’s National Airport Pavement Test Facility (Figure 3.6).

PSPA is a portable device (see Figure 3.7) and consists of a receiver transducers and a source transducer; and is based on generating and detecting stress waves in a layered medium. The data collected is processed by spectral analysis to determine the modulus of the layer. Advantages of using PSPA include nondestructive in nature, rapid to perform, immediate availability of results, repeatable, in situ testing of pavement in its natural state, and easy and convenient to operate.

FAA National Airport Pavement Test Facility. (Reprinted from ASNT Publication.)

Figure 3.6   FAA National Airport Pavement Test Facility. (Reprinted from ASNT Publication.)

Portable seismic property analyzer (PSPA). (Reprinted from ASNT Publication.)

Figure 3.7   Portable seismic property analyzer (PSPA). (Reprinted from ASNT Publication.)

PSPA used estimate modulus values from tests on three concrete slabs under three different support conditions at different times after the concrete pours. The results compared well with the modulus values measured by free-free resonance tests performed on the concrete beam and cylindrical specimens prepared at the same time as the concrete slabs. The same PSPA device was also used to deter-mine the modulus of hot mix asphalt (HMA) layer during CC2-OL (HMA overlay over rubblized concrete) project except that the active length of the seismic path was shortened. These two studies demonstrated the application potential of a NDT de-vice such as PSPA for assessing the early-age strength of concrete and the mod-ulus of HMA layer.

3.2.2  Role of SHCE

Structural health monitoring/structural health in civil engineering tools and methods can play major role during early phases of bridge life span. Table 3.2 shows some examples of such a role.

3.3  Why Bridges Live?

3.3.1  Definitions of Healthy Life

Healthy life of bridges can be expressed in three facets: (a) meeting functional goals, (b) adequate performance, and (c) at reasonable costs. Functional goals in-clude traffic, pedestrian, and commercial needs. Performance metrics include en-suring safety and security of users. This can be accomplished, in technical terms, by meeting adequate engineering criteria such as deflection, vibration, and stress/strain limits. Reasonable costs, or adequate cost benefits, are a major issue that is discussed in numerous parts of this volume.

This section addresses healthy life of bridges, its causes, and how to quantify a healthy bridge life. We offer some specific examples as to how SHM/SHCE field can improve healthy bridge life.

3.3.2  Classification of Healthy Life Causes

3.3.2.1  Enabling Causes of Healthy Life

By definition, enabling causes are continuous: they occur nearly all the time and their effects are felt over the long term. There are many enabling causes that would ensure healthy life of a bridge. To continue the framework of quantification of healthy life, we assign formal variables to these causes as discussed in Table 3.3.

Table 3.2   Roles of SHM in Early Phases in Bridge Life Span

Ensuring Healthy Life for New Bridges

Sensing/Measurements

STRID

DMID

Decision Making

Preliminary planning

Soil conditions need to be measured. In some situations, special sensing might be needed, such as propensity for liquefaction in areas with high seismic activities

Conventional structural analysis methods can be used for preliminary planning

Literature searches for damages and methods of DMID for bridges and sites that are comparable to the bridge and the site that are under considerations. The results of these searches can have two benefits:

  • Affects decisions for the current project
  • Affects decisions on any DMID method to be used in an SHM setup for the current project, if any

  • Cost–benefit analysis
  • Risk analysis
  • Life cycle analysis

Design/detailing

Placement of future sensors. Types, numbers, and locations of the sensors

Optimal NDT methods

Base analytical STRID models for updating at different stages of the bridge life cycle

Optimal integration of SHM activities and structural design of the bridge

Optimal decisions that accommodate

  • Cost–benefit
  • Life cycle analysis

Construction

Safe construction process can be monitored by strain, displacement or vibration sensors that are placed in optimal locations. Those locations are predetermined using structural analysis tools

QA/QC

Adequate curing of concrete can be monitored using embedded RFID temperature sensors

Quality of welds can be inspected using different NDT techniques

Table 3.3   Enabling Causes for Healthy Life

Enabling Cause

Objective Rating

Sound management

0 ≤ EL1 ≤ 10

Design

0 ≤ EL2 ≤ 10

Detailing

0 ≤ EL3 ≤ 10

Construction

0 ≤ EL4 ≤ 10

Scheduled maintenance

0 ≤ EL5 ≤ 10

Inspection/monitoring

0 ≤ EL6 ≤ 10

Other

To be added as needed

We thus establish the enabling measure of a bridge healthy life as

3.1 H L T H E = W E i = 1 i = N E ( W E i · E L i )
where
  • WEi = Appropriate weight for the ith parameter
  • WE = Appropriate weight for the whole enabling causes
  • NE = Total number of enabling causes

Note that in Table 3.3 NE = 6. The number of enabling causes can change, depending on the situation on hand. The different weights in Equation 3.1 are assigned based on experience and research results. The measure HLTHE is thus a relative measure that can be used in DM situations, such as allocation of budgeting, and so on.

3.3.2.2  Triggering Causes of Healthy Life

Triggering healthy life causes are relatively sudden: they occur suddenly on the time scale of the bridge life span; their effects are felt suddenly in the form of improved bridge health. There are many triggering causes that would ensure healthy life of a bridge. Similar to the enabling causes, we assign formal variables to triggering causes as discussed in Table 3.4.

We thus establish the triggering measure of a bridge healthy life as

3.2 H L T H T = W T i = 1 i = N T ( W T i · T L i )
where
  • WTi = Appropriate weight for the ith parameter
  • WT = Appropriate weight for the whole triggering causes
  • NT = Total number of triggering causes

Note that in Table 3.4 NT = 4. The number of trig-gering causes can change, depending on the situation on hand. The different weights in Equation 3.2 are assigned based on experience and research results. The measure HLTHT is thus a relative measure that can be used in DM situations, such as allocation of budgeting, and so on.

3.3.3  Quantification of Healthy Life

3.3.3.1  Conventional Metrics

There are numerous methods to quantify the health of a bridge; see, for example, NYSDOT (2008). Among some of those methods:

  • Bridge Condition Rating: A simple rating system from 1 (totally deteriorated condition) to N (new condition). The value of N vary. For example, New York State uses N = 7. The value of the condition rating indicates the health condition of the bridge. Table 3.5 shows qualitative description of the condition ratings as used in New York State.

Table 3.4   Triggering Causes for Healthy Life

Triggering Cause

Objective Rating

Retrofits

0 ≤ TL1 ≤ 10

Rehabilitation

0 ≤ TL2 ≤ 10

Special maintenance

0 ≤ TL3 ≤ 10

Replacement

0 ≤ TL4 ≤ 10

Other

To be added as needed

Table 3.5   Definitions of Bridge Condition Ratings

Rating

Definition

1

Totally deteriorated or failed condition

2

Used for condition between 1 and 3

3

Serious deterioration or not functioning as originally designed

4

Used for conditions between 3 and 5

5

Minor deterioration and is functioning as originally designed

6

Used for conditions between 5 and 7

7

New condition

Table 3.6   Categories of Bridge Deficiencies

Condition Rating

Deficiency Level

Description

Potential Remedial Action

<3

Severe

Comprehensive serious deterioration of the bridge structural elements

Priority to remediate and repair

3.0 and ≤ 3.999

Moderate

Serious deterioration to some of the bridge main structural elements

Comprehensive structural work is likely. Rehabilitation and replacement options are generally available

4.0 and ≤ 4.999

Marginal

Moderate structural deterioration. Minor deterioration to primary support elements Efficient structural performance is rarely compromised

Minor rehabilitation/major maintenance activities might be needed

One common way to quantify the bridge condition rating is described in detail in NYSDOT (2008). The procedure is summarized as follows: during each general in-spection of the bridge, various components, or elements of each bridge span are rated by the inspector according to the extent of deterioration and the ability of the component to function structurally, relative to when it was newly designed and con-structed. These element rating values are then combined using a weighted average formula to compute an overall bridge condition rating value for each bridge. This formula assigns greater weights to the ratings of the bridge elements having the greatest structural importance and uses lesser weights for minor structural and non-structural elements. If a bridge has multiple spans, each element common to mul-tiple spans is rated on a span-by-span basis; the lowest individual span element rating is used in the overall condition rating formula.

Deficiency: A deficient bridge can be defined using its condition rating value. For example, NYSDOT defines deficient bridges as those with a condition rating less than 5 (on a scale from 1 to 7). A deficient condition rating indicates the presence of sufficient deterioration and/or loss of original function to require corrective maintenance or rehabilitation to restore the bridge to its fully functional, nondeficient condition. It does not mean that the bridge is unsafe. In most cases, bridges have enough excess or reserve structural capacity to accommodate some deterioration or degradation of structural function as indicated by a deficient condition rating. Figure 3.13 shows the number of deficient bridges in New York State between 1992 and 2006. The trend of the deficient bridge numbers is decreasing. To correlate state of deficiency with DM, NYSDOT (2001) in a report on bridge overloading identified further different states of bridge deficiency, their condition rating and potential remedial actions in Table 3.6.

Posting: When the rating of a bridge is reduced even further to a rating of 3 or less, some protective measures might be needed. One example is the posting of weight limits that are allowed on the bridge. Also, the bridge might be closed until it can be repaired, rehabilitated, replaced or permanently closed. Another qualitative indication of bridge health is red flags. These flags identify potentially or imminently unsafe structural conditions and require the owner to take prompt, certified corrective, or protective actions to resolve the flag. These include repair, posting, or closure. Figure 3.14 shows the number of red flags issued in New York State between 1989 and 2006.

Alampalli and McCowan (2008) described in detail the flagging procedure, which is a practical measure to describe state of bridge health, in New York State. It is part of the comprehensive bridge inspection program implemented in New York State to identify serious deficiencies, both structural and nonstructural, affecting public safety so that owners can take appropriate action in a timely fashion. The procedure is very robust, safety oriented, and establishes requirements for certifying that appropriate corrective or protective measures are taken within an appropriate time frame. The critical inspection findings (flags) can be either structural or safety related. The structural flags are further subdivided into two categories.

Red structural flags are used to report the failure of a critical primary structural component or a failure that is likely before the next scheduled inspection (see Figures 3.8 and 3.9).

Yellow structural flags are used to report a potentially hazardous condition which, if left unattended beyond the next anticipated inspection, would likely become a clear and present danger. This flag is also used to report the actual or imminent failure of a noncritical structural component, where such failure may reduce the reserve capacity or redundancy of the bridge, but would not result in a structural collapse by the time of the next scheduled inspection interval (see Figures 3.10 and 3.11).

Red flag due to serious cracking in abutment stem. (Courtesy of New York State Department of Transportation.)

Figure 3.8   Red flag due to serious cracking in abutment stem. (Courtesy of New York State Department of Transportation.)

Red flag due to overextended bearing. (Courtesy of New York State Department of Transportation.)

Figure 3.9   Red flag due to overextended bearing. (Courtesy of New York State Department of Transportation.)

Yellow flag due to a crack in the primary redundant member. (Courtesy of New York State Department of Transportation.)

Figure 3.10   Yellow flag due to a crack in the primary redundant member. (Courtesy of New York State Department of Transportation.)

Yellow flag due to loss of bearing area in a pedestal. (Courtesy of New York State Department of Transportation.)

Figure 3.11   Yellow flag due to loss of bearing area in a pedestal. (Courtesy of New York State Department of Transportation.)

Nonstructural conditions are reported using a “Safety Flag.” The safety flag is used to report a condition presenting a clear and present danger to vehicle or pedestrian traffic, but is in no danger of structural failure or collapse. Safety flags can also be issued on closed bridges whose condition presents a threat to vehicular or pedestrian traffic underneath the bridge (see Figure 3.12).

The notification and response procedures vary depending on the flag type.

US Federal Rating System: The federal ratings result from an overall condition assessment of each bridge’s three or four major components and do not require the multielement evaluations mandated by other states, such as NYSDOT inspection program. The federal ratings are used to identify bridges that do not meet contemporary Federal Highway Administration (FHWA) standards. Those bridges are classified as either “structurally deficient” or “functionally obsolete.” Bridges are considered “structurally deficient,” according to the FHWA, if significant load-carrying elements are found to be in poor or worse condition due to deterioration and/or damage, the bridge has inadequate load capacity or repeated bridge flooding causes traffic delays. The fact that a bridge is “structurally deficient” does not imply that it is unsafe or is likely to collapse.

A “structurally deficient” bridge, when left open to traffic, typi-cally requires significant maintenance and repair to remain in service and eventual rehabilitation or replacement to address the deficiencies. To remain in service, structurally deficient bridges are often posted with weight limits. “Functionally obsolete” refers to a bridge’s inability to meet current standards for managing the volume of traffic it carries, not its structural integrity. For example, a bridge may be functionally obsolete if it has narrow lanes, no shoulders, or low clearances.

Safety flag due to exposed wiring. (Courtesy of New York State Department of Transportation.)

Figure 3.12   Safety flag due to exposed wiring. (Courtesy of New York State Department of Transportation.)

Table 3.7   Count of Bridges by Construction Material

Count of Bridges

Concrete

Steel

Prestressed Concrete

Wood

Othersa

Total stock

248,739

188,551

132,033

26,682

3802

Structurally deficient

18,506

38,419

5036

9855

710

Functionally obsolete

28,187

34,678

12,773

3482

685

The federal bridge rating scale addresses both structural condition and functional adequacy, using different criteria than the above mentioned NYSDOT inspection condition rating scale. Table 3.7 and Figure 3.13 show structurally deficient and functionally obsolete highway bridges in the United States as categorized by con-struction material. Table 3.8 and Figure 3.14 show structurally deficient and func-tionally obsolete bridges in the United States as categorized by bridge type.

We end this section by observing that a good state inspection practice should in-clude both Federal ratings as well as the state’s own ratings, if available (Figures 3.15 and 3.16).

3.3.3.2  Healthy Life Causes as Metrics

We can quantify the health of a bridge as

3.3 H L T H = H L T H E + H L T H T
Note that the health of the bridge measure HLTH is a qualitative measure that should be used for DM situations. It should be used as a complementary measure to other bridge performance measures such as bridge ratings. The main difference between HLTH and bridge rating is that bridge rating evaluated the field condition of bridge. On the other hand, HLTH evaluates other factors that lead to the current bridge condition. Both measures account for different set of parameters, thus the consideration of both can give a more complete picture of the bridge health.

Count of bridges by construction material. (From FHWA, National Bridge Inventory, NBI, Federal Highway Administration, Washington, DC,

Figure 3.13   Count of bridges by construction material. (From FHWA, National Bridge Inventory, NBI, Federal Highway Administration, Washington, DC, 2007.)

Table 3.8   Count of Bridges by Structure Type

Count of Bridges

Slab

Stringer/Multi Beam or Girder

Girder and Floor Beam System

TEE Beam

Box Beam or Girders (Single or Spread/Multiple)

Truss (Deck and thru)

Suspension/Stayed Girder

Othersa (Culvert)

Total stock

79,879

249,238

7432

36,444

55,330

12,608

133

158,726 (126,401)

Structurally deficient

6600

39,911

2594

4770

3189

7161

36

8262 (2944)

Functionally obsolete

9612

39,819

1792

7807

6660

2455

44

11,609 (5543)

Count of bridges by structure type. (From FHWA, National Bridge Inventory, NBI, Federal Highway Administration, Washington, DC,

Figure 3.14   Count of bridges by structure type. (From FHWA, National Bridge Inventory, NBI, Federal Highway Administration, Washington, DC, 2007.)

Percent of deficient bridges. (Courtesy of New York State Department of Transportation,

Figure 3.15   Percent of deficient bridges. (Courtesy of New York State Department of Transportation, NYSDOT 2008.)

Number of red flagged bridges. (Courtesy of New York State Department of Transportation,

Figure 3.16   Number of red flagged bridges. (Courtesy of New York State Department of Transportation, NYSDOT 2008.)

To give an example of the use of HLTH let us consider an example of two bridges. The official needs to compare the health of the bridges. By making an in-depth study of all enabling and triggering causes, the different estimated factors were computed (qualitatively) as shown in Tables 3.9 and 3.10.

Notice that the relative weights in both Tables 3.9 and 3.10 are assigned by the official. More studies are needed for the nature and the values of these weights. The final estimates for HLTH are 695 and 763 for bridges A and B, respectively. This outcome indicates that bridge B should be healthier than bridge A.

Table 3.9   Example of Enabling Causes for Healthy Life of Two Bridges

Enabling Cause

Weight

Objective Rating of Bridge A

Objective Rating of Bridge B

Sound management

10

7

9

Design

8

6

6

Detailing

10

4

7

Construction

10

7

5

Scheduled maintenance

12

8

6

Inspection/monitoring

12

8

9

Total

420

438

Table 3.10   Example of Triggering Causes for Healthy Life of Two Bridges

Triggering Cause

Weight

Objective Rating of Bridge A

Objective Rating of Bridge B

Retrofits

10

5

5

Rehabilitation

10

6

5

Special maintenance

15

8

8

Replacement

15

3

7

Total

275

325

3.3.4  Role of SHM

Structural health monitoring/structural health in civil engineering can play a major role in almost all enabling and triggering bridge healthy life. Some specific examples are shown in Table 3.11.

The project that was described by Kosnik and Hopwood (2008) provides an ex-ample of healthy bridge life that was aided by HSM monitoring. The John F. Ken-nedy Memorial Bridge, a large cantilever through truss opened in 1963, carries In-terstate 65 across the Ohio River between Louisville, KY, and Jeffersonville, IN. An overall view of the bridge is shown in Figure 3.17. The bridge is restrained by pairs of bearing assemblies on each bank of the river. The bearings resist considerable designed-in uplift forces, particularly on the Indiana side. In 2006, one of the anchor bolts on the northwest bearing was found to have been fractured (see Figure 3.18). The washer on the bolt opposite the failed bolt could be spun in place by hand, indicating that there was very little tension in that bolt. The bearing assembly with the fractured bolt also moved visibly under live traffic.

Live strain, displacement, and acceleration data were collected on uplift bearing anchor bolts for a total of approximately 17 hours over several weeks under a varie-ty of weather and traffic conditions and both before and after replacement of a fractured anchor bolt (see Figure 3.19). On the basis of initial data it was determined that the compromised North West, NW bearing assembly was subject to large live strains in two of the three remaining anchor bolts. The fractured North West-North East (NW-NE) anchor bolt was replaced with a threaded rod instrumented with strain gauge arrays, and strain data on the three original anchor bolts and the replacement bolt were recorded before and after installation and tightening of the replacement. The live strains in the original anchor bolts decreased after tightening of the replacement bolt, and no further indications of bending of the anchor bolts were observed.

Table 3.11   SHM Role in Some Bridge Healthy Life

No.

Healthy Life Issues

SHM Applications

1

Rehabilitation/repair choice, closing/opening/usage-restriction of the structure, evacuation for safety

Monitoring and evaluating condition state as is. STRID and DMID projects

2

Evaluation/implementation of new designs/construction methodology at local or global level

New and existing engineering paradigms

3

Quality of life (improving mobility, etc. —opening appropriate lanes or redirecting traffic in-time)

Bridge testing and life cycle analysis.

4

Maintenance decisions (routine to using just-in-time concept—say sending a salt truck to a bridge only when you see that such conditions warrant that will help structural durability, saves money, and improves safety)

Bridge management tools and methods Applying reliability and risk methods for maintenance decisions

5

Safety (users and structural safety)

Engineering design and analysis paradigms

6

Evaluation (durability, design/analysis option evaluation, etc.)

Deterioration monitoring and decision making tools

7

Research, development, and technology transfer

  • Improving the state-of-the-art
  • Improving the state-of-the-practice
  • Bridging the gap between the state-of-the-art and state-of-the practice

SHM/SHCE in civil infrastructure is in its infancy (as of writing of this volume) more research is needed in all components of the fields

8

Security (relatively new in bridge field, but probably common for security applications you deal with)

There are many common needs and applications for bridge security and SHM (see Chapter 11 of Ettouney and Alampalli 2012)

Overall view of the I-65 John F. Kennedy Memorial Bridge. (Reprinted from ASNT Publication.)

Figure 3.17   Overall view of the I-65 John F. Kennedy Memorial Bridge. (Reprinted from ASNT Publication.)

3.4  Why Bridges Fail/Die?

3.4.1  Definitions of Failure

There are several definitions of bridge failure. Most of those definitions are qualitative. A general failure definition was provided by NYSDOT (2001), which categorized failure into three groups as follows

  • Catastrophic: The structure is vulnerable to a sudden and complete collapse of a superstructure span or spans. This failure maybe the result of partial or total failure of either the superstructure or the substructure. A fail-ure of this type would endanger the lives of those on or under the structure.
  • Partial Collapse: The structure is vulnerable to major deformation or discontinuities of a span. (This would result in loss of service to traffic on or under the bridge.) This failure may be the result of tipping or tilting of the substructure causing deformation in the superstructure. A failure of this type may endanger the lives of some of those crossing or those under the structure.
  • Structural Damage: The structure is vulnerable to localized failures. This failure may be the result of excessive deformation or cracking in the primary superstructure or substructure members of the bridge. A fail-ure of this type maybe unnoticed by the travelling public but would require repair once it is discovered. This type of damage would also make a bridge more susceptible to overload failures.

Anchor bolt identification. (Reprinted from ASNT Publication.)

Figure 3.18   Anchor bolt identification. (Reprinted from ASNT Publication.)

Anchor bolt strain gauge locations. (Reprinted from ASNT Publication.)

Figure 3.19   Anchor bolt strain gauge locations. (Reprinted from ASNT Publication.)

Wardhana and Hadipriono (2003) showed the relationship between classes of failure that are similar to the above classification of NYSDOT and the phasing (con-struction vs. in-service phases). The number of failures (up to 2003) according to failure phasing is shown in Table 3.12.

Table 3.12   Number of Bridge Failures According to Failure Categories

Types of Failures

Construction

Service

Unknown

Distress

0

17

0

Partial collapse

3

80

13

Total collapse

5

12

21

Unknown

0

277

75

Total

8

386

109

Note that each of the above failure classifications/groups has two attributes: (a) The scale and extent of the damage, and (b) The consequences of such failure. This makes those definitions to be closer to a risk-based definition that concerns itself with vulnerabilities, threats, and consequences. Such general and encom-passing definition makes it possible to quantify failure. A general risk-based ap-proach would be

3.4 F R = T · V · C
where
  • FR = Failure severity estimate
  • T = Threat/hazard level estimate
  • V = Vulnerability of structure to failure due to T
  • C = Consequences of failure

We recognize Equation 3.4 as the classical risk equation FEMA 452 (2010) and FEMA (2009). We now try to make Equation (3.4) more specific for quantification of failure risk by defining

3.5 S T V = T · V
where
  • STV = Propensity to failure

Thus, a suitable failure severity estimate can be expressed as

3.6 F R = S T V · C
Assuming that the limits of STV and C to be
3.7 1 S T V 10
3.8 1 C 10
The failure estimate should have limits as
3.9 1 F R 100
Assuming that the propensity to failure of a given bridge “A” is STV = 4.5, while the consequence of this failure is STV = 7.5 the failure severity estimate is FR = 33.75. For a different bridge “B,” if the propensity to failure is STV = 7.3, while the consequence of this failure is STV = 3.0 the failure severity estimate is FR = 21.9. So, the severity of failure of bridge “A” is much more than the severity of failure of bridge “B,” even though bridge “B” is more vulnerable to the postulated threats than bridge “A.” Equations 3.4 through (3.9) can be used to quantify severity of failure across a number of bridge networks.

3.4.2  Classification of Failure Causes

Failure causes have been classified by Wardhana and Hadipriono (2003) into enabling and triggering causes. The differences between the two types of causes were discussed earlier; in general, enabling causes occur over longer period of time, while triggering causes occur suddenly. We discuss different enabling and triggering failure causes first. Next we try to associate those failure causes with Equations 3.4 through 3.9 for estimating f FR. We finally discuss the role of SHM in evaluating failure causes.

3.4.2.1  Triggering Causes

Triggering causes of failure include hydraulic/flood (scour, debris, drift, etc.), colli-sion/impact, overload (Figures 3.20 through 3.22), fire, ice, earthquakes (including tsunamis), wind (hurricanes/tornadoes, etc.), and soil failures. Note that all of these causes occur suddenly, and last relatively short period of time. Figures 3.23 and 3.24 show percentages of bridge failure rates according to different triggering caus-es in the United States and New York State, respectively. Note that this database is not a comprehensive database of all the failures but only the failures recorded by the NYSDOT based on data collected from periodic survey of other states.

Impact failure.

Figure 3.20   Impact failure.

Overload failure. (Courtesy of International Association of Structural Movers.)

Figure 3.21   Overload failure. (Courtesy of International Association of Structural Movers.)

Traffic overload. (Courtesy of International Association of Structural Movers.)

Figure 3.22   Traffic overload. (Courtesy of International Association of Structural Movers.)

Failure rates (US). (Courtesy of New York State Department of Trans-portation.)

Figure 3.23   Failure rates (US). (Courtesy of New York State Department of Trans-portation.)

Failure rates (New York State). (Courtesy of New York State Department of Transportation.)

Figure 3.24   Failure rates (New York State). (Courtesy of New York State Department of Transportation.)

3.4.2.2  Enabling Causes

Wardhana and Hadipriono (2003) subdivided enabling causes (they called it princip-al causes) as design, detailing (Figures 3.25 and 3.26), construction, maintenance, material related (including deterioration, normal wear and tear, fatigue, and/or corro-sion). Of these, maintenance and material related causes were the largest causes for collapse or distress. Note that all of these causes occur over relatively long pe-riod of time. Figure 3.27 shows effects of construction material type on bridge fail-ures. Figure 3.28 shows the effects of bridge construction dates and Figure 3.29 shows bridge type on failure rates.

3.4.2.3  Estimating Bridge Failure Propensity to Causes

We just presented a plethora of bridge failure causes. For a given bridge, many of those causes might not be applicable; yet several others might have an effect on the bridge. We present a method for evaluating total failure severity estimate when there are several potential causes of failure that we define as FR_TOTAL. A generalized form of Equation 3.4 can be offered as

3.10 F R _ T O T A L = α ( i = 1 i = N C A U S E ( F R i ) n ) 1 / n
where
  • FRi = Failure severity estimate for ith failure cause
  • NCAUSE = Number of pertinent failure causes
  • n = Suitable power number
  • α = Scaling factor

The form of Equation 3.10 was used in FEMA (2009) to sum the individual risks of uncorrelated threats; it is reasonable to use to sum the severity of failures due to uncorrelated failure causes in our current endeavor. Note that n = 2 reduces Equation 3.10 to the popular square root of sum of squares (SRSS) method. How-ever, we propose to use a much higher value for n, say n = 10, which would produce more realistic results, see FEMA (2009) for more discussion of this approach. Proceeding as before

3.11 F R i = T i · V i · C i
where
  • Ti = Threat/hazard level estimate ith failure cause
  • Vi = Vulnerability of structure to failure due to T ith failure cause
  • Ci = Consequences of failure of the ith failure cause
Also, defining
3.12 S T V i = T i · V i
where
  • STVi = Propensity to failure due to Ti

Failure due to concrete details (a) overview of failure site and (b) failed box girder. (Courtesy of New York State Department of Transportation.)

Figure 3.25   Failure due to concrete details (a) overview of failure site and (b) failed box girder. (Courtesy of New York State Department of Transportation.)

Failure due to steel details. (Courtesy of New York State Department of Transportation.)

Figure 3.26   Failure due to steel details. (Courtesy of New York State Department of Transportation.)

Failure of bridges according to material of construction. (Courtesy of New York State Department of Transportation.)

Figure 3.27   Failure of bridges according to material of construction. (Courtesy of New York State Department of Transportation.)

Failure of bridges according to construction date. (Courtesy of New York State Department of Transportation.)

Figure 3.28   Failure of bridges according to construction date. (Courtesy of New York State Department of Transportation.)

Failure of bridges according to bridge type. (Courtesy of New York State Department of Transportation.)

Figure 3.29   Failure of bridges according to bridge type. (Courtesy of New York State Department of Transportation.)

Thus, a suitable failure severity estimate can be expressed as

3.13 F R i = S T V i · C i
Assuming that the limits of STV and C to be
3.14 1 S T V i 10
3.15 1 C i 10
By adjusting the value of α, the total failure severity estimate would have limits as
3.16 1 F R _ T O T A L 100
The limits in Equations 3.14 through 3.16 are arbitrary. Careful studies should be conducted to evaluate the effects of those limits, and provide realistic limits, for practical bridge situations.

As an example of estimating failure severity estimate for multiple causes, let us consider two bridges, “A,” and “B,” with different estimated failure causes, propensity of failures, and consequences of failure and shown in Table 3.13. In this example, the power number is assumed to be n = 10. The number of pertinent hazards are NCAUSE = 6 and NCAUSE = 5 for bridges “A,” and “B,” respectively. Bridge “A” does not have hydraulics as a realistic failure cause, since it does not cross a water body. Bridge “B” is not located in a seismically prone area; it does not have wind as a potential cause of failure. The failure severity estimates of bridges “A,” and “B,” are 30.15 and 25.65, respectively. Studying the numbers in Table 3.13, it is clear that the consequences of bridge “A” failures caused these results even though bridge “A” seem to have lesser threats and vulnerabilities than bridge “B.”

Equations 3.10 through 3.16 can be used to quantify severity of failure due to any number of failure cause combinations across a number of bridge networks.

Table 3.13   Failure Severity Estimates for Two Bridges

Bridge “A”

Bridge “B”

Cause

STVi

Ci

Fri

STVi

Ci

Fri

Construction

2

9

18

4

4

16

Hydraulic

NA

NA

NA

6

5

30

Seismic

4

9

36

NA

NA

NA

Wind

2

9

18

NA

NA

NA

Collision/impact

2

7

14

6

3

18

Fatigue

6

4

24

6

3

18

Overload

3

5

15

7

3

21

Ncause

6

5

α

0.84

0.85

Scaled totals

30.15

25.65

Table 3.14   SHM Roles in Bridge Failure—General

Causes of Failure

Sensing/Measurements

STRID

DMID

Enabling

Design Detailing

Use sensing to verify design procedures for unusual geometries that are not within the bounds of conventional codes such as bridges with large skew angles

Modal or parameter identification methods can be used to enhance analytical bridge models, especially for existing bridges

Using in-field damage information, such as remaining fatigue life, provide for an efficient and safe retrofit bridge designs

Construction Maintenance

See Table 3.15 NDT methods can compliment conventional visual inspection practices

Different STRID methods produce more accurate analysis results for maintenance projects

Detecting damage in an accurate and timely manner would focus maintenance efforts in safer directions

Material related

See Chapters 3 and 4(for concrete deterioration), Chapter 5 (for fatigue) and Chapters 6 and 7 (for fiber reinforced polymers) in Ettouney and Alampalli (2012)

Triggering

External events

See Table 3.15

3.4.2.4  SHM/SHCE Role

Structural health monitoring plays a major role in all aspects of bridge failure. Some examples of the different roles each of the three components of SHM can play in enabling and triggering causes of bridge failure are shown in Table 3.14 through 3.16.

3.4.3  Attributes of Failure

The failure event of a bridge can be used as an information source to help under-stand failure mechanisms, causes, attributes, and so on. When bridge fails, it might be instructive to collect information regarding the event as follows:

  • Bridge Type: The structural type, including superstructure, substructure, foundation, and soil
  • Bridge Attributes: Span length, support conditions (before and after the failure)
  • Material: Record the material of construction of different bridge com-ponents
  • Abutments and Piers (before and after the failure): Type, height, size, bearing types, foundations, seat widths, and so on.
  • Age: Year Built, year failed, and if there were any indications of sud-den deterioration or loss of function
  • Failure Type: Describe the type of failure (as described throughout this section)
  • Failure Cause: The cause (or causes) of failure (as described throughout this section)
  • Consequences: This includes direct consequences such as number of Fatalities and/or number of injuries. Additional consequences might be of interest such as economic, social or other local/regional consequences
  • Multimedia: Recording failure aftermath with photographs and/or video can provide valuable database. This is particularly important since failure sites are usually reconstructed soon after the failure event

Table 3.15   SHM Roles in Bridge Failure—Triggering Causes

Hazard

Sensing/Measurements

STRID

DMID

Collision/accidents (Figure 3.30)

On bridge vibration monitoring can reveal occurrence and extent of collision events

Before event and after event STRID efforts can give estimates of extent of global, and perhaps local, damage

Global or local DMID methods can aid in producing safe and cost-effective retrofit efforts

Concrete deterioration

See Chapters 3 and 4 in Ettouney and Alampalli (2012)

Construction

Sensing behavior of systems during construction projects would provide an invaluable QA/QC insight that helps in improving safety during construction and potential clarifications of future behavior

Modal identification methods can show abnormal dynamic behavior during construction. Thus providing additional safety measures to the construction process

DMID techniques can show extent of damages after construction accidents

General deterioration

SHM procedures for general deterioration are similar to those procedures in Chapters 3 and 4 in Ettouney and Alampalli (2012)

Fire (Figures 3.31 and 3.32)

After fire events, sensing would produce valuable information regarding extent of fire damage (deflections, tilt, etc.)

Modal identification methods can reveal potential changes in global bridge behavior after fire

NDT methods such as acoustic emission (local damage) or thermography (regional damage) can provide damage information

Hydraulic/scour/flood

See Chapter 1 in Ettouney and Alampalli (2012)

Nature

See Chapters 1 through 5 in Ettouney and Alampalli (2012)

Overload

Capacity and demands can be correlated in real time by measuring bridge response (capacity: strains, displacements, etc.) and live load demands (truck loads, volume and frequency)

STRID methods offer a unique opportunity for improving modeling technique of overloaded systems

Global or local DMID methods can aid in producing safe and cost-effective retrofit efforts when a system is overloaded

Steel deterioration

See Chapters 3, 4, and 5 in Ettouney and Alampalli (2012)

Seismic (Figure 3.33)

See Chapter 2 in Ettouney and Alampalli (2012)

Fatigue

See Chapter 5 in Ettouney and Alampalli (2012)

Table 3.16   SHM Role in Bridge Failure—Age of Bridge

Bridge Age

Sensing/Measurements

STRID

DMID

Older

Monitoring systems on older bridges can be costlier than their counterparts on new systems

Can validate rehabilitation and existing analytical models

Different DMID procedures can help in detecting damages of older bridges, thus extending their life span

Newer

Newer bridges offer opportunities of placing monitoring systems that are more interconnected and consistent with the bridge system than older bridges

Can validate rehabilitation and existing analytical models. They can help in validating and improving accuracy of construction-related analytical models

Newer bridges are less exposed to aging-type damages, such as deterioration or increased traffic demands. However, they can be vulnerable to abnormal hazards that exceed their design limits, such as an abnormally high flood or abnormally strong earthquake

Failure due to accidents. (Courtesy of New York State Department of Transportation.)

Figure 3.30   Failure due to accidents. (Courtesy of New York State Department of Transportation.)

Fire failures: steel girders have deformed extensively. (Courtesy of New York State Department of Transportation.)

Figure 3.31   Fire failures: steel girders have deformed extensively. (Courtesy of New York State Department of Transportation.)

Fire failure: deck damage. (Courtesy of New York State Department of Transportation.)

Figure 3.32   Fire failure: deck damage. (Courtesy of New York State Department of Transportation.)

Seismic failure. (Courtesy of New York State Department of Transportation.)

Figure 3.33   Seismic failure. (Courtesy of New York State Department of Transportation.)

3.4.4  Structural Modes of Failure

Bridges, as well as any other infrastructure can fail only in one of two temporal structural modes: either brittle or ductile. Spatially, the structures can either fail pro-gressively, or globally. We explore those structural modes of failure next. At the end of this section, we examine the role and limitations of SHM in each of those modes. Note that we differentiate between damage and failure in this discussion. Damage would be a local event in spatial terms, while failure is more extensive in spatial terms.

3.4.4.1  Ductile Failure

Ductile structural failure occurs over long period of time. Signs of distress are usually observable. As such most conventional SHM techniques can be used to monitor the damages and alert for any impeding failure. Examples of ductile failure are failures due to foundation settlements, or excessive wind-induced deflections and vibrations. The failures of Tacoma Narrows Bridge (Figure 3.45) and the Tay Rail Bridge (Figure 3.42) are ductile-type failures. The Tacoma Narrows collapse due to excessive wind vibrations lasted long enough to prevent any human causality. Unfortunately, the Tay Rail Bridge collapse occurred suddenly, thus causing great loss in life. We consider the Tay Rail Bridge failure as a ductile failure since the signs of poor performance were observed long before the collapse, such as rattling of steel connections, and excessive vibrations when trains crossed the bridge.

3.4.4.2  Brittle Failure

Brittle structural failure occurs suddenly. Signs of distress are usually subtle, and difficult to observe. As such care is needed when applying conventional SHM tech-niques. It is important to clearly identify the brittle modes of failure that the SHM project aims to uncover, and ensure that the techniques used are capable of identi-fying those brittle failure modes. Examples of brittle failure are failures due to corrosion (Figure 3.34) or fatigue (Figure 3.35) since both types of failure occur suddenly, even though the damage occurred over a long time period. The failure of the I-35 Bridge in Minnesota (Figure 3.59) is clearly a brittle failure; it occurred suddenly and caused major catastrophic consequences.

Corrosion failure. (Courtesy of New York State Department of Transportation.)

Figure 3.34   Corrosion failure. (Courtesy of New York State Department of Transportation.)

Fatigue failure. (Courtesy of New York State Department of Transportation.)

Figure 3.35   Fatigue failure. (Courtesy of New York State Department of Transportation.)

3.4.4.3  Global Failure

Brittle or ductile failure can lead to global bridge failure. Global failure occurs when the whole structural system becomes unstable. Global instability is perhaps the most severe failure mode and the most difficult to detect or monitor. When the bridge is newly constructed, its global stability condition is usually investigated and ensured. As the bridge ages, this global stability condition changes: the global stability safety factor starts decreasing. After a suddenly applied abnormal loading condition, such as an earthquake event, the structural stability safety factor decreases even further, due to the nonlinear changes in the system. Such nonlinearities would decrease global stability resilience. Ettouney et al. (2006) explored the effects of such nonlinearities on the global stability of structural systems. For highly redundant structural systems reductions of global stability safety factors are slow. For low redundant systems, investigating global structural stability should be performed often; for such investigations to be beneficial, they should be done using actual bridge properties, not theoretical properties. An example of the use of SHM in observing stability condition of a steel truss is given in chapter 8.

3.4.4.4  Progressive Failure (Collapse)

Progressive collapse condition occurs when the failure starts locally, then progress further away from the initial location of the failure. Ettouney et al. (2004) presented a progressive collapse theory that stated that progressive collapse event can lead to one of two outcomes: (a) if the failure front is arrested, say by reaching a resilient support, or (b) if large portion of the structure have failed such that the remaining structure becomes globally unstable. The latter scenario was discussed earlier. We discuss now the first outcome, that is, the progression of failure front. Extent of progressive collapse in bridges can be described in terms of four scales, as follows.

  • Limited Progression: The collapse progression is limited in scale. Examples are fatigue cracks, local buckling, and limited corrosion
  • Local Progression: Collapse progressed to a single and or continuous spans; superstructure only
  • Regional Progression: Collapse progressed to a single and or continuous spans; superstructure, as well as substructure
  • Global Progression: The collapse would propagate across many spans and over many piers and supports

Figure 3.36 shows schematics of the failure scales. Clearly, SHM techniques can be of value in monitoring and assessing progressive collapse at different collapse scales, as shown in Table 3.17.

3.4.4.5  Mitigation Strategies and Failure Modes

Obviously, optimal failure mitigation strategies should depend on the mode of failure. An optimal strategy should aim at delaying or eliminating the failure under consideration at reasonable costs. Quantitatively, this means that we should strive to reduce probability of failure. Since in this section we are concerned with bridge (or system) failure, we recall immediately that an optimal failure mitigation strategy should aim at increasing the bridge system reliability. Note that for optimal strategy, reduction of threat (corrosion prevention measures, for example) might offer a more cost-effective solution than reducing vulnerability. Thus an optimal mitigation strate-gy might be defined as the one that would reduce risk, provide for higher benefit to cost ratio, or produce the optimal life cycle analysis (costs, benefits and life spans). Chapters 8, 9, and 10 in Ettouney and Alampalli (2012) discuss those issues in greater detail. In what follows, we explore specific optimal mitigation needs for different structural failure modes.

Ductile Failure: The goal of ductile failure mitigation should be to eliminate or reduce the chance of such a failure. Inspection (Figure 3.38) or monitoring can help in such a situation. Regular maintenance and rehabilitation efforts should be adequate in mitigating ductile failure. The bridge in Figure 3.39 is being rehabilitated after a general deterioration in its conditions was observed over the course of several years. Note that the rehab efforts are striving to preserve the historic nature of the bridge.

Scales of bridge failures.

Figure 3.36   Scales of bridge failures.

Table 3.17   SHM Role in Different Progressive Collapse Scales

Failure Scale

SHM Role

Example

Sensing

STRID

DMID

Limited

Fatigue crack

See Chapter 5 in Ettouney and Alampalli (2012)

Local

Barge collision

  • Before event: proximity sensors
  • During event: vibration signatures (time histories of accelerations, velocities, or strains)
  • After event: different NDT procedures to detect damages

Before and after event STRID model evaluation can help in damage assessment, accurate retrofit design, and decision-making processes

Global DMID methods, e.g., thermography or more localized DMID methods, e.g., ultrasonic or acoustic emission can aid in detecting presence, type, location and extent of damage

Regional

Scour failure of footing or underlying soils (Figure 3.37)

See Chapter 1 in Ettouney and Alampalli (2012)

Global

Failure of single span will propagate to other spans by dynamics, or catenary action

Placing vibration sensors over the whole length of the bridge (in an optimal manner as in Ettouney and Alampalli [2012]) can help in accurate understanding of the progressive collapse event

Simulating progressive collapse of any structural system in an accurate fashion is not an easy task. Using different STRID methods, coupled with optimal sensing data, would result in an improved accuracy of progressive collapse modeling

NDT methods can help in forensic analysis of failed structure

Scour failure. (Courtesy of New York State Department of Transporta-tion.)

Figure 3.37   Scour failure. (Courtesy of New York State Department of Transporta-tion.)

Bridge inspection.

Figure 3.38   Bridge inspection.

Brittle Failure: Brittle failure is much more difficult to mitigate. This is due to the difficulty in continuously assessing the actual global stability condition. We immediately notice two contributing factors:

  • Governing Redundancy Factor: It is essential to identify all pertinent potential failure modes and the redundancies that control each of these modes. Traditionally, theory of structures provided some help in assessing some structural redundancies. For example, configurations of members in trusses (Figure 3.40) or the number of supports (Figure 3.41) can indicate the degree of redundancy in a given system. Noting that such general assessment rely on the implicit assumption that all structural com-ponents are in pristine condition. Sudden damages or normal wear and tear can alter this situation, reducing what might have been a highly redundant system into a system with low redundancy that is susceptible to brittle fail-ures.
  • Time Factor: As indicated earlier, passing of time can reduce the governing redundancy of a bridge system. An optimal brittle failure mitiga-tion strategy should account for the as-built condition, not the designed condition.

Rehabilitation to avoid ductile failure. (Courtesy of New York State De-partment of Transportation.)

Figure 3.39   Rehabilitation to avoid ductile failure. (Courtesy of New York State De-partment of Transportation.)

Internal redundancy.

Figure 3.40   Internal redundancy.

An effective mitigation strategy should follow some or all of the following steps

  1. Identify all governing redundancies in the system. This should be done us-ing a well-designed STRID effort
  2. Identify and monitor actual conditions of items that control governing re-dundancies of #1. A well integrated system of inspection and SHM is needed
  3. Use the different SHM principles in this volume: duality, scaling and seren-dipity to help identifying low redundancy that might not have been detected in #1 or #2. For example, a subtle change in local resonance might indicate a reduction in a governing redundancy

External redundancy.

Figure 3.41   External redundancy.

When embarking on brittle failure mitigation efforts, the decision maker needs to ask the following questions:

  • Can visual inspection alone detect brittle failure potential?
  • Can sensing strains alone detect brittle failure potential?
  • What kind of sensing can detect brittle failure potential?

The detailed answers to these questions can ensure an effective and optimal brittle failure mitigation strategy.

Global Failure: Global failure mitigation strategies share same attributes to both ductile and brittle failure mitigation strategies.

Progressive Failure (Collapse): First, we need to define progressive collapse mitigation strategy (PCMS). We propose two types of strate-gies. An active PCMS would try to prevent local failures that might cause progres-sive failures from happening in the first place. A passive PCMS would try to prevent the scale of failure from increasing. Thus, passive PCMS implies accepting the fact that initial failure might/can occur.

Passive PCMS include the use of two concepts: (a) improving redundancy of the structural system, and/or (b) hardening different components of the structural sys-tem. Detailed and careful studies are needed to ensure that

  • The PCMS is adequate to meeting the project objectives
  • The PCMS does not have side effects that can reduce resiliency during other types of threats (multihazards effects)

The active PCMS aims at reducing or eliminating local failures that can cause pro-gressive collapse. Different SHM techniques, such as NDT methods (chapter 6), bridge security methods (see chapter 11 in Ettouney and Alampalli 2012) or bridge management methods (see chapter 9 in Ettouney and Alampalli 2012) can be used for optimal active PCMS.

Decision Making Role: Mitigation of bridge failure is complex and overreaching subject, as evident from the above overview. There are numerous problems to resolve, and there are even more choices for mitigation approaches. Issues such as hardening versus improving redundancy, governing redundancies, passive PCMS versus active PCMS provide overwhelming choices to the decision maker. Pairings of problems and solutions are not easy to find. We recommend a utilization of different quantitative DM techniques in finding such pairings. Not only DM offer potential of using optimal mitigation solution, it also can be automated (due to its quantitative nature) to any SHM, or management practice. The cost saving and consistency of decisions are two obvious benefits of using DM techniques in this field.

3.5  Examples of Bridge Failures

3.5.1  Introduction

Competent personnel using well-established material and structural specifications design bridges. In recent years, most bridge designs do employ quality control and assurance measures. Similarly, they are also inspected routinely to identify issues that affect their structural safety to carry the in-service loads as well as routine over-loads imposed on them. Appropriate actions are taken based on inspection results depending on the urgency, effect of the critical findings on structural safety and durability, and available resources. But, most of these are reactive in nature and applies to slow form of deterioration such as corrosion and age-related issues. But, bridges do face several hazards and are designed to some of these hazards based on a certain return interval while not designed for other hazards that the structure has a very low probability with low consequence of encountering in its service life. Due to several uncertainties and due to the limitations associated with specifica-tions, constructability, design/analysis difficulties and time, variability of materials used, and other factors, designs are generally very conservative in nature.

Despite all the conservatism, failures do happen often and in some cases with loss of life and significant economic impacts, due to several factors. These include design errors, unanticipated level of loads or hazards, material issues, failure of quality control or assurance, limitations associated with inspection, communication issues between different groups involved with bridge service life, neglect. But, very few failures can be attributed to hazard level exceeding the designed hazard level. In most cases, bridges do not fail due to a single factor but a combina-tion/convergence of a number of factors. Even though some of these can be de-tected through inspection process several such as design errors, items concealed from general inspections, and severe problems that can develop rapidly before the next cycle are hard to detect once bridge is in service. This section discusses some of the historic failures and documented reasons to show that the bridge failures in general are due to convergence of several factors progressing for several years than a simple, sudden, single reason.

3.5.2  Tay Rail Bridge, UK

The 3.5-km-long Tay Rail Bridge spanned the Firth of Tay in Scotland, between the city of Dundee and the suburb of Wormit in Fife, replacing an early train ferry. On December 28, 1879, the centre section known as the “High Gird-ers” (thru truss) collapsed killing 75 people when a train on it went down with the bridge.

The Tay Rail Bridge was a lattice-grid (truss) design, combining cast and wrought iron and was opened on June 1, 1878. The bridge had several issues early in the construction that was attributed to original bedrock surveys. During the construction, it was found that the bedrock near the banks was deeper than predicted earlier. This led to the redesign of piers and reduced number of piers making superstructure spans longer that originally planned. The bridge was a combination of deck truss (“Low Girders”) and thru truss sections.

The official forensic investigations noted that the bridge was “badly de-signed, badly built and badly maintained, and that its downfall was due to inherent defects in the structure, which must sooner or later have brought it down” (Court of Inquiry 1879). Thus several factors contributed to the failure including the following: (a) Allowance for wind load had been made by the designer Bouch (Seim 2008) as he was advised that this was unnecessary for girders shorter than 200 ft. The designer apparently did not make any allowances when the revised design in-volved longer girders. The middle section of the bridge was a thru truss to allow a higher clearance for the passage of ships underneath and was potentially top heavy and very vulnerable to high winds; (b) The cylindrical cast iron columns supporting the 13 longest spans of the bridge, each 75 m long, were of poor quality. Many had been cast horizontally, with the result that the walls were not of even thickness, and there was some evidence that imperfect castings were disguised from the quality control inspections. In particular, some of the lugs used as attachment points for the wrought iron bracing bars had been “burnt on” rather than cast with the columns. However, no evidence of the burnt-on lugs has survived, and the normal lugs were very weak. They were tested for the Inquiry by David Kir-kaldy and were proved to break at only about 20 t rather than the expected load of 61 t. These lugs failed and destabilized the entire centre of the bridge during the storm; (c) There was clear evidence that the central structure had been deteriorating for months before the final accident. A few months after the bridge was opened, the bridge inspector noticed that several joints had loosened making many of the tie-bars useless for bracing the cast-iron piers; and (d) Recent research also indi-cates that the cast iron used to join the columns of the bridge together might have become brittle under great strain and might have contributed to the failure.

This bridge failure shows the value of preliminary site data collection, proper de-sign, selection of structural material, proper quality control (during the design, fabri-cation, construction, and inspection), and proper maintenance when deficiencies are noted during the bridge inspection. It is interesting to that some of the same errors contributed to the failure of the recent collapse in Minnesota that failed more than100 years after the failure of the Tay Rail Bridge (Figures 3.42 and 3.43).

3.5.3  Quebec Bridge Failure

One of the oldest failures of major bridges, killing 75 of the 86 people on the bridge, during the construction can be the First Quebec Bridge in August of 1907. Failure of this bridge shows the value of checking the design calculations after any significant change in structures design, quality control in the design process, involvement of the design engineer during construction through field visits, and supervision of a competent reengineer during the construction phase.

During 1907, the Quebec Bridge was supposed to be the bridge with the longest central span as well as longest cantilever span once built. But, the southern half of the bridge that was nearing completion collapsed suddenly. Theodore Cooper, then famous engineer, directed the project. The engineer soon after his appointment recommended increasing the span from 1600 ft to 1800 ft to reduce the vulnerability of the piers to scour by moving them to shallow water and to reduce the project time by making it easy to construct. To offset the increase steel costs, he also modified specifications that would allow for higher unit stresses. Because the St. Lawrence was a shipping lane, the 2800-ft bridge needed almost 150 ft clearance above the water to allow the ocean-going vessels to pass. Further, the bridge was to be multifunctional and was required to be 67-ft-wide to accommodate two railway tracks, two streetcar tracks, and two roadways.

Tay Rail Bridge after construction.

Figure 3.42   Tay Rail Bridge after construction.

Tay Rail Bridge after collapse.

Figure 3.43   Tay Rail Bridge after collapse.

In the rush to provide shop drawings for steel fabrication preliminary weight cal-culations prepared in early stages were never properly checked when the design was finalized using revised specification. The real bridge weight exceeded signifi-cantly than the preliminary estimated weights and thus the bridge dead load capaci-ty. Due to the inability of the project director to visit the field, an engineer with less experience for the major bridge construction was in charge of the field site. By the time, a material inspector reported the excess weights, south anchor arm, tower, and two panels of the south cantilever arm were fabricated and six panels of the anchor arm were already in place. The project director concluded that the increased stresses are within allowable limits and proceeded with the construction of the bridge. When the bridge was nearing completion during the 1907 summer, the effect of increased dead loads were revealed in the form of distortions of lower chord compression members that are key structural members. Signs of buckling were ob-served in them and also in the splices between some lower chord members. The deflection of one of the chords of the south anchor arm grew from 0.75 in to 2.25 in. But the work continued as it was felt that there is no immediate danger and some attributed the bends presence before the installation. By the time, project director was informed, a decision was made and conveyed to field personnel to stop the construction, two compression chords of the anchor arm buckled and the subse-quently bridge collapsed in matter of seconds. For detailed discussions about the Quebec Bridge collapse, see Lienhard (2008), Smith (2008), and Ricketts (2008) (Figure 3.44).

3.5.4  Tacoma Narrows, WA

In 1937, the Washington State legislature created the Washington State Toll Bridge Authority and appropriated $5,000 to study the request by Tacoma and Pierce County for a bridge over the Narrows. The first Tacoma Narrows Bridge, a cable-supported bridge, was opened to traffic on July 1, (See Figures 3.45 and 3.46). It collapsed 4 months later on November 7, 1940, at 11:00 AM (Pacific time) due to a physical phenomenon known as aeroelastic flutter caused by a 67 km per hour wind in a moderate windstorm (Ketchum 2011c).

The bridge was the third longest suspension bridge when built with a center span of 2800 ft, 29-ft-wide, and had the greatest ratio of length to width ever built. The false bottom type of caissons were used in sinking of cellular piers in water 120-ft-deep. Total height of west and east piers was 198, and 247 ft, respectively. Shore anchorages for suspension cables each contained 2500 cubic yards of con-crete weighing 52, 500 tons. Two steel towers, 425 ft in height and each weighing 1927 tons, surmount the piers and support the suspension cables. The two suspen-sion cables, 17.5 in diameter and more than 1 mile long, each consists of 6308 #6 parallel wires formed into 19 individual strands of 332 wires each. This forms a total of 14,000 miles of wire weighing 3817 tons (Washington University 2008a and 2008b).

Collapsed Quebec Bridge: sequence of events that led to collapse.

Figure 3.44   Collapsed Quebec Bridge: sequence of events that led to collapse.

Completed bridge, before opening. (Courtesy of University of Washington Libraries, Specail Collections.)

Figure 3.45   Completed bridge, before opening. (Courtesy of University of Washington Libraries, Specail Collections.)

Opening of bridge for traffic. (Courtesy of University of Washington Libraries, Specail Collections.)

Figure 3.46   Opening of bridge for traffic. (Courtesy of University of Washington Libraries, Specail Collections.)

The Tacoma experience taught engineers that wind causes not only static loads on the bridge, but also significant dynamic actions. It is commonly presented as an example of failure due to resonance when fundamental frequency of the bridge coincides with the external periodic frequency of the load. A cable-supported bridge is subject to wind-induced drag (the static component), flutter (the instability that occurred at Tacoma Narrows), and buffeting (where gusts “shake” the bridge). Adequate aerodynamic performance is required with respect to each of these effects. For modest span bridges, drag generally controls the strength required to resist wind. Flutter becomes critical when the wind acting on the structure reaches a critical velocity that triggers a self-excited unstable condition. The task in design is to assure that the critical wind velocity is high enough so that it has a very low probability of occurrence. This can be achieved by providing a stiff structure and/or an aerodynamically streamlined superstructure shape. Buffeting influences fatigue of the bridge materials as well as users’ comfort. The magnitude of buffeting response under higher probability wind conditions must be controlled. Addressing these issues in an engineering context requires the use of wind tunnel models. Current practice is converging on use of such models for the aerodynamic properties of the bridge shape only. The mechanical properties of the bridge, and the final wind evaluation, are performed using computer models that incorporate the wind tunnel results (Ketchum 2008). Figure 3.47 illustrates the sequence of events that led to the bridge collapse.

Sequence of bridge collapse (a) interrelationship between geometry and increased wind demands and (b) wind flutter.

Figure 3.47   Sequence of bridge collapse (a) interrelationship between geometry and increased wind demands and (b) wind flutter.

3.5.5  Silver Bridge, OH

The Silver-Point Pleasant Bridge (known in history as Silver Bridge) collapsed on December 15, 1967 evening killing 46 people and injuring 9 when 31 of the 37 ve-hicles on the bridge plunged into water. Since the bridge on the U.S. Highway 35 connects Point Pleasant, WV and Kanauga, OH over the Ohio River, a major trans-portation route connecting both the States was destroyed causing a major disruption to many and caused enormous sensation across the nation on the safety of bridges.

The Silver Bridge got its name as it was the country’s first aluminum painted bridge. The 2235-ft-long Silver Bridge was a two-lane eye-bar suspension type bridge designed according to American Society of Civil Engineers specifica-tions and constructed in 1928 by the General Corporation and the American Bridge Company. It was designed with a 22-ft roadway and one 5-ft sidewalk with “High Tension” eye-bar chains, a unique anchorage system, and rocker towers. At the time of its construction, the maximum permitted truck gross weight was about 20,000 lb compared to the large truck limit of 60,000 lb or more in 1967. See Figure 3.48.

The Silver Bridge was the first eye-bar suspension bridge of its type to be con-structed in the United States. The bridge’s eye bars were linked together in pairs like a chain. A huge pin passed through the eye and linked each piece to the next. Each chain link consisted of a pair of 2 × 12 in bars and was con-nected by an 11 in pin. The length of each chain varied depending upon its location on the bridge. Such bridges had usually been constructed from redundant bar links, using rows of four to six bars, sometimes using several such chains in parallel. The eye bars in the Silver Bridge were not redundant as links were composed of only two bars each of high strength steel (more than twice as strong as common mild steel), rather than a thick stack of thinner bars of modest material strength “combed’ together as is usual for redundancy. With only two bars, the failure of one could impose excessive loading on the second, causing total fail-ure—unlikely if more bars are used. While a low-redundancy chain can be engineered to the design requirements, the safety is completely dependent upon correct, high quality manufacturing, and assembly.

“Rocker’ towers were used that was also a unique feature of the Silver Bridge. About 131-ft-high towers allowed the bridge to move due to shifting loads and changes in the chain lengths due to temperature variations. A curved fitting was placed next to a flat one at the bottom of the piers. The rocker was then fitted with dowel rods to keep the structure from shifting horizontally. With this type of connection, the piers were not fixed to the bases. These allow the bridge to respond to various live loads by a slight tipping of the supporting towers that were parted at the deck level, rather than passing the suspension chain over a lubricated or tipping saddle or by stressing the towers in bending. Thus the towers required the chain on both sides for their support, so failure of any one link on either side, in any of the three chain spans would result in the complete failure of the entire bridge.

The forensic investigations concluded that the heat-treated carbon steel eye-bar (#330, on the north of the Ohio subsidiary chain, the first link below the top of the Ohio tower) broke due to a small crack that formed through fretting wear at the bearing. Over the years, partly due to increased loads (Figure 3.49) when compared to design loads, stress corrosion, and corrosion fatigue allowed the crack to grow, and when became critical, the member broke in a brittle fashion. This placed undue stress on other members of the bridge. The remaining steel frame buckled and fell due to the newly concentrated stresses. The entire structure collapsed in the span of a minute.

Silver Bridge collapse—before collapse.

Figure 3.48   Silver Bridge collapse—before collapse.

Changes in demands for Silver Bridge.

Figure 3.49   Changes in demands for Silver Bridge.

The visual inspections performed cannot detect such cracks formed in the eye bar without disassembling the eye bar. These defects are still hard to detect even in present days. This failure prompted President Johnson to initiate a taskforce charged to determine procedures available to preclude future disasters and imple-ment changes, if needed. This led to the establishment of National Bridge Inspection Standards by the Federal Government and current routine bridge inspections with qualified inspection personnel. Details of the bridge and its failure can be found at National Transportation Safety Board (NTSB) (1971) and LeRose (2001). Figure 3.50 illustrate some of the factors that led to collapse. Figure 3.51 shows the aftermath of the collapse.

3.5.6  Mianus River Bridge Collapse

Mianus River Bridge carried Interstate 95 over the Mianus River near Greenwich, CT. A 100-ft suspended eastbound span between piers 20 and 21 collapsed on June 28, 1983, killing three and seriously injuring three others when their vehicles fell into the river with the bridge. The collapse was attributed to the failure of the pin and hanger connection that was commonly used decades ago as this offered easier analysis and low construction costs. For more details, see Nationmaster.com (2011d).

The suspended span that collapsed was attached to the bridge structure at each of its four corners. A pin and hanger assembly attached each corner to the girders of the cantilever arm of an adjacent anchor span to support the weight of the northeast and southeast corners of the suspended span, The pin and hanger assembly includes an upper pin attached through the 2 1/2-in-thick web of the girder of the cantilever arm and a lower pin attached through the 2 1/2-in-thick web of the girder of the suspended span. One and one half-inch-thick steel hangers connect the upper and lower pins-one on the inside and one on the outside of the web (NTSB 1984). Figure 3.52 shows a pin and hanger system that is similar to the one used in Mianus River Bridge.

Silver Bridge collapse—reasons for collapse.

Figure 3.50   Silver Bridge collapse—reasons for collapse.

Silver Bridge collapse—after collapse. (With permission from National Highway Institute.)

Figure 3.51   Silver Bridge collapse—after collapse. (With permission from National Highway Institute.)

Sometime before the collapse of the suspended span, the inside hanger in the southeast corner of the span came off of the inside end of the lower pin. This action shifted the entire weight of the southeast comer of the span onto the outside hanger. The outside hanger gradually worked its way farther outward on the pin, and over a period of time, a fatigue crack developed in the top outside end of the upper pin. The shoulder of the pin fractured off, the pin and hanger assembly failed, and the span collapsed into the river. The NTSB determined that the probable cause of the collapse of the Mianus River Bridge span was the undetected lateral displacement of the hangers of the pin and hanger suspension assembly in the southeast corner of the span by corrosion-induced forces due to deficiencies in the State of Connecticut’s bridge safety inspection and bridge maintenance program (NTSB 1984). Figure 3.53 shows the failed bridge. Figure 3.54 illustrates the sequence of failure.

Typical pin and hanger construction.

Figure 3.52   Typical pin and hanger construction.

Failed span. (With permission from National Highway Institute.)

Figure 3.53   Failed span. (With permission from National Highway Institute.)

The investigations cited corrosion from water build up due to inadequate drainage as a cause. The highway drains had been deliberately blocked during road mending some 10 years before and water leaked down through the pin bearings, causing them to rust. The outer bearings were safety-critical and nonredundant, a design flaw of this particular type of structure. The bearings were difficult to inspect close up, although traces of rust could be seen near the affected bearings.

This highlights the need for thorough inspection and maintenance program, use of nondestructive procedures when needed, and the value of redundancy.

3.5.7  Schoharie Bridge Collapse

On April 5, 1987, two spans of the five-span Schoharie Creek Bridge collapsed causing five vehicles falling into the river with 10 fatalities. A third span collapsed 90 minutes later. The bridge was built in 1954 and owned by the New York State Thruway Authority. It was located in Mohawk valley northwest of Albany, NY. The 540-ft-long and 112-ft-wide bridge carried four lanes of traffic, median area, and shoulders over Schoharie creek. The average height of the bridge above the creek was about 80 ft. The bridge was designed to 1949 edition of AASHTO Standard Specifications for Highway Bridges. It was a five-span steel superstructure with reinforced concrete deck. The piers consists of two slightly tapered reinforced con-crete columns connected by a tie beam near column tops and supported on a com-mon pedestal and footing. Two of the piers were in the Creek with other two piers on the creek banks to support the structure along with two end abutments.

Failure sequence of a pin and hanger system (a) lateral and longitudinal expansion of hangers, (b) as-designed pin-hanger system, and (c) failure of bridge due to lack of system redundancy.

Figure 3.54   Failure sequence of a pin and hanger system (a) lateral and longitudinal expansion of hangers, (b) as-designed pin-hanger system, and (c) failure of bridge due to lack of system redundancy.

The bridge collapsed on April 5, 1987 during the spring flood caused by a rainfall totaling 150 mm combined with snowmelt to produce an estimated 50 year flood (WJE Associates 1987). The collapse was initiated by the toppling of pier three, which caused the progressive collapse of spans three and four into the flooded creek. Pier two and span two fell 90 minutes after span three dropped, and pier one and span one shifted 2 hours after that (Thornton-Tomasetti 1987). The NTSB suggested that pier two collapsed because the wreckage of pier three and the two spans partially blocked the river, redirecting the water to pier two and increasing the stream velocity (NTSB 1988).

According to investigations, scouring under piers began shortly after the bridge was built in 1955 as the bridge footings experienced floodwater flows unanticipated in the design of the bridge, a 100-year flood. It is believed that the majority of the scouring energy was dissipated into moving the original riprap layer from around the footings. Once the backfill had been exposed, the years of peak flows removed the backfill material, and the backfill material in turn was replaced by sediment settling into the scoured. Furthermore, the riprap placed at construction had probably been washed away during the 1955 flood, and had not been replaced (WJE Associates 1987).

The investigations concluded that the collapse of the Schoharie Creek Bridge was due to the extensive scour under pier three that was affected by four important factors (Thornton-Tomasetti 1987): (a) the depth of shallow footings, bearing on soil, was not enough to take them below the probable limit of scour; (b) the foundation of pier three was bearing on erodable soil that allowed high velocity floodwaters to penetrate the bearing stratum; (c) the as-built footing excavations and backfill could not resist scour; and (d) riprap protection, inspection and maintenance were inadequate.

Besides scour issues, several factors including two common practices when the bridge was designed in the 1950s contributed to the severity of the collapse (Thorn-ton-Tomasetti 1987). These were the bridge bearings allowed the spans to lift or slide off of the concrete piers and the simple spans were not redundant. Thus, the use of continuous spans, rather than simple spans, would have provided redundancy once pier three failed, and perhaps allowed for the redistribution of forces between the spans. The failure also shows that it is important for bridge owners to identify the critical features that can lead to the collapse of a bridge and to ensure that those critical features are inspected frequently and adequately (NTSB 1988). Although the Schoharie Creek Bridge had been inspected annually or biennially since 1968, an underwater inspection of the piers footings had never been performed. The bridge was scheduled for an underwater inspection in 1987, but the bridge collapsed before the inspection took place (NTSB 1988). The failures also emphasize how important it is to design footings deep enough to avoid loss of support capacity due to scour (Shephard and Frost 1995) (Figures 3.55 and 3.56).

3.5.8  I-35 Bridge, Minneapolis, MN

On August 1, 2007, evening the eight lane, 1907-ft-long I-35W highway bridge over the Mississippi River in Minneapolis, Minnesota, experienced a catastrophic failure in the main span of the deck truss killing 13 and injuring 145 persons. One thousand feet of the deck truss collapsed, with about 456 ft of the main span falling 108 ft into the 15-ft-deep river.

The I-35W Bridge was located about 1 mile northeast of the junction of I-35W with Interstate 94. In addition to spanning the Mississippi River, the bridge also extended across Minnesota Commercial Railway railroad tracks and three roadways: West River Parkway, 2nd Street, and the access road to the lock and dam. The bridge was opened to traffic in 1967 with 141,000 ADT in 2004. The 14-span structure used welded built-up steel beams for girders and truss members, with riveted and bolted connections. The bridge was 1907-ft-long and carried eight lanes of traffic, four northbound and four southbound. The 1064-ft-long deck truss portion of the bridge encompassed a portion of span 5; all of spans 6, 7, and 8; and a portion of span 9. The deck truss was supported by four piers (see Figure 3.57).

Collapsed bridge, from

Figure 3.55   Collapsed bridge, from http://ny.water.usgs.gov/projects/scour/text.html

Scour hole, from

Figure 3.56   Scour hole, from http://ny.water.usgs.gov/projects/scour/fig5.html

I-35 Bridge spanning the Mississippi river. (

Figure 3.57   I-35 Bridge spanning the Mississippi river. (NTSB 2008.)

Riveted steel gusset plates at each of the 112 nodes (connection points) of the two main trusses tied the ends of the truss members to one another and to the rest of the structure. The gusset plates were riveted to the side plates of the box mem-bers and to the flanges of the H members. All nodes had at least two gusset plates, one on either side of the connection point. A typical I-35W main truss node, with gusset plates, is shown in Figure 3.58.

The NTSB report (NTSB 2008) attributed the probable cause of the collapse of the bridge (Figure 3.59) to the inadequate load capacity due to a design error of the gusset plates at the U10 nodes (see Figure 3.60), which failed under a combination of substantial increases in the weight of the bridge, which resulted from previous bridge modifications, and the traffic and concentrated construction loads on the bridge on the day of the collapse. Contributing to the design error was the failure of quality control procedures by the design company to ensure that the appropriate main truss gusset plate calculations were performed for the I-35W Bridge and the inadequate design review by Federal and State transportation officials. Contributing to the accident was the generally accepted practice among Federal and State transportation officials of giving inadequate attention to gusset plates during inspections for conditions of distortion, such as bowing, and of excluding gusset plates in load rating analyses.

Typical five-member node (two upper chord members, one vertical member, and two diagonal members) on I-35W Bridge. (

Figure 3.58   Typical five-member node (two upper chord members, one vertical member, and two diagonal members) on I-35W Bridge. (NTSB 2008).

Collapsed bridge. (

Figure 3.59   Collapsed bridge. (NTSB 2008).

Locations of components. (

Figure 3.60   Locations of components. (NTSB 2008.)

Count of bridges in the United States: (a) Number of bridges built - sorted by construction material, and (b) total number of bridges built.

Figure 3.61   Count of bridges in the United States: (a) Number of bridges built - sorted by construction material, and (b) total number of bridges built.

Table 3.18   Count of Bridge Inventory in the United States

Year Built

Concrete

Steel

Prestressed Concrete

Wood

Before 1900

1738

4196

642

731

1901–1905

283

825

126

34

1906–1910

882

2129

222

69

1911–1915

1556

1597

161

47

1916–1920

3248

2899

381

182

1921–1925

5564

2395

403

152

1926–1930

12,121

6529

852

716

1931–1935

11,874

9255

752

1650

1936–1940

13,883

11,566

657

1994

1941–1945

4463

2949

198

898

1946–1950

11,824

9509

418

2121

1951–1955

15,185

11,268

728

2255

1956–1960

24,584

16,684

5278

3090

1961–1965

24,345

18,621

9176

2453

1966–1970

22,017

18,002

11,709

2309

1971–1975

15,333

14,304

12,513

1631

1976–1980

13,801

10,990

12,195

1701

1981–1985

14,186

9321

13,510

1286

1986–1990

15,409

10,742

15,685

1589

1991–1995

14,265

9611

15,580

1534

1996–2000

14,500

8845

15,178

870

2001–2007

14,545

9275

15,977

568

3.6  Appendix I: Count of Bridges in the United States by Construction Material

The number of highway bridges in the United States is a massive number indeed as shown in Figure 3.61 and Table 3.18. The primary materials used in bridge construc-tion are concrete, pre-stressed concrete, steel, and wood, as shown in Figure 3.61a. The utilization of wood as a bridge construction material peaked in the mid-twentieth century. Around the same time, the utilization of pre-stressed concrete increased. At present, the use of steel, concrete, and pre-stressed concrete dominate bridge con-struction. Figure 3.61b indicates the total bridges built in 5-year intervals. As of the writing of this volume, the total number of bridges in the United States is over 600,000. Note that the numbers in Figure 3.61b does not account for the decommis-sioned, failed, or demolished bridges.

References

Alampalli, S. and McCowan, P., (2008) “Addressing Critical Findings from Highway Bridge Inspection: New York State Procedures,” NDE/NDT for Highways and Bridges: Structural Materials Technology (SMT), ASNT, Oakland, CA.
Court of Inquiry, (1879) “Tay Bridge Disaster,” Report Of The Court of Inquiry, and Report Of Mr. Rothery, “Upon the Circumstances Attending the Fall of a Portion of the Tay Bridge on the 28th December 1879” place this as the second reference”. The Railways Archive. Retrieved on 2007-02-05.
Ettouney, M., and Alampalli, S., (2012) Infrastructure Health in Civil Engineering: Applications and Management, CRC Press, Boca Raton, FL.
Ettouney, M., Smilowtz, R., Tang, M., and Hapij, A., (2006) “Global System Considerations for Progressive Collapse with Extensions to Other Natural and Man-Made Hazard,” Journal of Performance of Constructed Facilities, 20(1), 403–417.
FHWA., (2007) “National Bridge Inventory, NBI,” Federal Highway Administration, Washington, DC.
FEMA., (2009) Handbook for Rapid Visual Screening of Buildings to Evaluate Terrorism Risk, Federal Emergency Management Agency, FEMA Publication, Washington, DC.
FEMA., (2010) A How-To Guide to Mitigate Potential Terrorist Attacks, Federal Emergency Management Agency, FEMA Publication, Washington, DC.
Garg, N., Larkin, A., and Ricalde, L., (2008) “Evaluation of the Application Potential of PSPA in determination of Modulus of Concrete HMA pavements at the FAA’s National Airport Pavement Test Facility,” NDE/NDT for Highways and Bridges: Structural Materials Technology (SMT), ASNT, Oakland, CA.
Hansen, W. and Surlaker, S., (2006) Embedded Wireless Temperature Monitoring Systems For Concrete Quality Control, University of Michigan, An Arbor, MI.
Ketchum, M., (2011c) “Tacoma Narrows” http://www.ketchum.org/wind.html [accessed February 20, 2011].
Kosnik, D. and Hopwood, T., (2008) “Monitoring of In-Situ Strains in Bearing Assembly Anchor Bolts on a Large Through - Truss Bridge,” NDE/NDT for Highways and Bridges: Structural Materials Technology (SMT), ASNT, Oakland, CA.
LeRose, C., (2001) The Collapse of the Silver Bridge, West Virginia Historical Society Quarterly Vol. XV, No. 4.
Lienhard, J., (2008) “Quebec Bridge Collapse,” http://www.uh.edu/engines/epi220.htm [accessed February 20, 2011].
National Transportation Safety Board (NTSB)., (1988) “Collapse of New York Thruway (1–90) Bridge over the Schoharie Creek, near Amsterdam, New York, April 5, 1987.” Highway Accident Report: NTSB/HAR-88/02, Washington, D.C.
Nationmaster.com (2011d) “Mianus River Bridge,” http://www.nationmaster.com/encyclopedia/Mianus-River-Bridge [accessed February 20, 2011].
Ni, Y., Xia, Y., Liao, W., and Ko, J., (2008) “Technology Innovation in Developing the Structural Health Monitoring System for Guangzhou New TV Tower,” International Symposium on Structural Control and Health Monitoring, Taichung, Taiwan.
NTSB., (1971) “Mianus River Bridge Collapse” http://www.ntsb.gov/Publictn/1971/HAR7101.htm, accessed on September 2, 2008. [accessed February 20, 2011].
NTSB., (1984) “Silver Bridge Collapse” report date 1984 accessed on 5 – 15 – 2008, http://www.ntsb.gov/publictn/1984/HAR8403.htm [accessed February 20, 2011].
NTSB., (2008) “Highway Accident Report, Collapse of I-35W Highway Bridge Minneapolis, Minnesota, August 1, 2007,” National Transportation Safety Board, Washington, D.C., NTSB/HAR-08/03, PB2008–916203, Notation 7975C.
NYSDOT., (2008) SFY 2006–07 Annual Report of Bridge Management and Inspection Programs, New York State Department of Transportation, Albany, NY.
Ricketts, B., (2008) “The Collapse of the Quebec City Bridge,” accessed from the web on 2008-02-05, http://www.mysteriesofcanada.com/Quebec/quebec_bridge_collapse.htm [accessed February 20, 2011].
Seim, C., (2008) “Why Bridge Have Failed Throughout History,” Civil Engineering Vol 78 (5): 64–71, 84–87, http://www.ASCE.org [accessed February 20, 2011].
Shepherd, R. and Frost, J. D., (1995) Failures in Civil Engineering: Structural, Foundation and Geoenvironmental Case Studies. American Society of Civil Engineers, New York.
Smith, C., (2008) “A Disaster in the Making,” accessed from the web on 2008-02-05, http://www.civeng.carleton.ca/ECL/reports/ECL270/Prelude.html [accessed February 20, 2011].
Thornton-Tomasetti, P. C., (1987) “Overview Report Investigation of the New York State Thruway Schoharie Creek Bridge Collapse.” Prepared for New York State Disaster Preparedness Commission, December.
Wardhana, K. and Hadipriono, F., (2003) “Analysis of Recent Bridge Failures in the United States,” ASCE J. of Performance of Constructed Facilities, Vol 17, No. 3.
Washington University., (2008a) Data sheet on the 1940 Tacoma Narrows Bridge, 1940, http://content.lib.washington.edu/cdm4/item_viewer.php?CISOROOT=/farquharson&CISOPTR=108&CISOBOX=1&REC=1
WJE, Associates, Inc., and Mueser Rutledge Consulting Engineers., (1987) “Collapse of Thruway Bridge at Schoharie Creek,” Final Report, Prepared for: New York State Thruway Authority, November.
Search for more...
Back to top

Use of cookies on this website

We are using cookies to provide statistics that help us give you the best experience of our site. You can find out more in our Privacy Policy. By continuing to use the site you are agreeing to our use of cookies.