Bundesanstalt für Materialforschung und -prüfung

International Symposium (NDT-CE 2003)

Non-Destructive Testing in Civil Engineering 2003
Start > Contributions >Lectures > Plenary 4: Print

Nondestructive Evaluation For Highway Bridges In The United States

Glenn Washer, Federal Highway Administration, United States Department of Transportation, Turner Fairbank Highway Research Center, McLean, VA 22101


This paper will describe developments and activities at the Federal Highway Administration's Nondestructive Evaluation Validation Center (NDEVC). The center is designed as a resource for the development and testing of nondestructive evaluation (NDE) technologies, and to provide State highway agencies in the United States with independent development, evaluation and validation. The center is comprised of three elements; the laboratory, located at the Turner Fairbank Highway Research Center acts as the nucleus of the center. Test bridges, located in Northern Virginia and Pennsylvania, act as field test specimens to evaluate technologies under realistic field conditions during practical trials. Sections of bridges containing defects, known as component specimens, are utilized to perform capability trials in the laboratory.

Activities of the center that will be discussed include a reliability study of the visual inspection of highway bridges. During this study, forty-nine bridge inspectors from 25 US states were asked to inspect the test bridges associated with the NDEVC. The results of the study describe the reliability of both routine and in-depth inspection practices. Conclusion from this unique study indicate that only 68% of condition ratings assigned by normal bridge inspectors lie within +/- 1 of the average rating for the deck, superstructure and substructure of bridges.

The development of innovative NDE tools will also be discussed. Current activities of the center are focused in the following areas:

  • Ultrasonic stress measurement in structural wires, strands and plate
  • Innovative load rating technologies
  • Field application of X-radiography
  • Ultra-wideband radar arrays for imaging concrete deck deterioration
  • Automated Ultrasonic Inspection (AUI) for steel bridge fabrication
  • Ultrasonic measurement of material properties


There are more than 590,000 bridges in the National Bridge Inventory (NBI) of the United States. Since shortly after the collapse of the Silver Bridge in 1967, the National Bridge Inspection Standards (NBIS) have provided guidance for the inspection of bridges. The standards require the inspection of all bridges on public roadways in the United States on a periodic basis, normally at least once every two years. Data from the inspections are maintained in the Nationally Bridge Inventory (NBI), a database of information on the size, construction and general condition of bridges and culverts in the United States.

In the more than thirty years since the program was adopted, bridge inspection has relied largely on visual inspection to evaluate bridge condition. Some nondestructive evaluation (NDE) methods have been used to compliment visual inspections, but widespread use of NDE tools has been limited for most State Departments of Transportation (DOT's). New NDE technologies are increasingly being sought to solve difficult inspection challenges that may be beyond the capability of normal visual inspections.

In 1996, the FHWA was mandated by the U.S. Congress to develop a center for evaluating NDE technologies for the inspection of highway bridges. The mandate was implemented through the development of the FHWA NDE Validation Center (NDEVC) at the Turner-Fairbank Highway Research Center in McLean, VA.

The objective of the NDEVC is to improve the state of the practice for highway bridge inspection. The Center is designed as a resource for the development and testing of NDE technologies, and to provide DOT's with independent evaluation and validation of those technologies. The NDEVC is also involved with the development of new NDE technologies to solve specific problems related to the inspection and evaluation of bridges.

The Center is comprised of three elements; the laboratory, located at the Turner Fairbank Highway Research Center acts as the nucleus of the Center. Test bridges, located in Northern Virginia and Pennsylvania, act as field test specimens to evaluate technologies under realistic field conditions during practical trials. Sections of bridges containing defects, known as component specimens, are used to perform capability trials in the laboratory and research related to the development of new NDE technologies.

This paper reports on recent activities at the NDE Validation Center. First, a study on the reliability of routine bridge inspection methods is described. The innovative study provides a backdrop to assess both the need for improved NDE technologies, and as a baseline to compare NDE technologies with current practices. The development of innovative NDE tools is also discussed. Current activities of the center discussed in this paper include:

  • Reliability of visual inspections
  • Laser measurement technologies
  • Bridge deck inspection
  • Inspection of composites
  • Automated ultrasonic testing

Reliability of Visual Inspection for Highway Bridges

Highway bridge inspections are conducted within the context of the NBIS that requires reporting of bridge condition in a standardized format. Condition ratings ranging from zero to nine are assigned to three components of the bridge; the superstructure, substructure and deck. The assignment of condition ratings to components is used to measure bridge performance at the national level, to forecast future funding needs, to determine the distribution of funds between states, and to determine the maintenance needs for a particular structure. Obviously, the accuracy of the condition ratings is important to ensure FHWA programs for funding bridge construction and renovation are equitable and meet the goal of reducing the number of deficient bridges, and to ensure bridges in need of maintenance and repair are identified.

A study of the accuracy of this inspection process was initiated by the NDEVC in 1998. The study provides overall measures of the reliability and accuracy of bridge inspection, identifies factors that may influence the inspection results, and determines some of the procedural differences that exist between various State inspection programs (Moore et. Al. 2001). The report, completed in 2001, is available from the U.S.D.O.T..

The study consisted of having a group of 49 practicing bridge inspectors from across the United States inspect the test bridges associated with the Validation Center. Each inspector performed 10 separate inspection tasks as part of the study, including both routine and in-depth inspections. An observer documented the performance and behavior of inspectors during the inspections. Each inspector was provided with a common set of hand tools to conduct the inspections that included a masonry hammer, a plum bob, a level, binoculars, flashlights, and other non-intrusive tools.

During the performance of six routine inspection tasks, inspectors were asked to provide a condition rating for the superstructure, substructure and deck. A typical example of the frequency distribution of condition ratings reported by inspectors for a reinforced concrete bridge is shown in figure 1. As the figure indicates, there was generally wide, normal distribution of results reported by inspectors. On average, between four and five different condition rating were assigned for each separate component examined in this portion of the study. Statistically, results indicated that only 68 % of reported condition ratings for these components would vary between +/- 1 from the average condition rating for a particular component, when the results are extrapolated to the entire population of bridge inspectors. This data, and other data from the study, indicates that there is wide variation in the manner in which routine inspections are conducted, and that the definitions of the particular condition states may not be refined enough to allow for accurate and reliable condition ratings.

Fig 1: Condition rating statistics as reported by inspectors for a reinforced concrete bridge.

The inspectors were also required to perform two in-depth inspections. An in-depth inspection is an up-close, arms-length inspection generally conducted to identify deficiencies not normally detected during routine inspections. Among the tasks performed by the inspectors was an in-depth inspection of a welded steel girder with fatigue-sensitive details. The inspectors were expected to search for and find seven crack indications that existed in the portion of bridge inspected. These indications were reported by inspectors at a rate of 3.9 %, or about 4 out of every 100 inspections of a particular crack indication correctly identified the indication. The results indicated that 86% of inspectors that correctly identified these defects used a flashlight for the inspection, and were on average 0.2 meters from the girder during the inspection. Among inspectors that did not correctly identify defects, only 38% used a flashlight, and averaged 2.7 meters from the girder during the inspection. These results indicate that the low crack detection rates found during the study may be related to how the inspections were performed, and additional training or retraining may help improve the use of appropriate inspection practices. It was concluded that a significant portion of in-depth inspections are unlikely to note defects beyond those found during a routine inspection.

Laser Measurement Technologies

Over the last four years, the NDEVC has developed numerous applications using laser-based distance measurements for highway infrastructure. A scanning laser system was built as part of an FHWA research and development program. The system is capable of measuring distances with sub-millimeter accuracy over a range of 30 meters. The mechanized scanning head enables the laser to be scanned over +/- 200° on one axis and +/- 60° on a second axis. Two angles and the distance measurement are combined to locate a point being measured in three-dimensional space. Targets are not required. Applications for this measurement technology include measuring bridge deflections under calibrated load to evaluate the structural behavior, measuring out-of-plane distortions in girder webs and flanges, and measuring as-built construction of large structures such as abutments.

Of the approximately 100,000 bridges classified as structurally deficient in the National Bridge inventory, more than 25% are classified as deficient due solely to a low structural appraisal rating, i.e. the bridges have a low load carrying capacity(Chase, 1999). This load carrying capacity is normally determined by theoretical calculation and may not accurately reflect the true capacity of the bridge. Experimental load rating is one way to ensure that calculations accurately represent bridge behavior. The laser system can be used to determine the deflections of a bridge as part of an experimental load rating scheme. Because the laser system requires no targets to make measurement, test set-up time can be greatly reduced. Access to the structure to connect gages or mount targets is not required. The measurement range of the system makes it possible to measure the deflection of a bridge over open lanes of traffic or above small waterways.

Fig 2: Out of plane deformations along a 1 x 6 meter section of a curved girder bridge under loading. An unstiffened web is shown in (A), a web with vertical stiffeners at the locations indicated is shown in (B).

Another recent application of this unique measurement technology is measuring out-of-place web distortions in a curved-girder bridge. There are several advantages to using the laser technology for this application. First, distortions of the web over a large field can be determined. Second, no interaction with the beam under test is required because the measurements are non-contact and made at a range of up to 30 meters (10 to 20 meters typically). Figure 2 indicates the results from testing of a curved girder bridge. The figure illustrates laser deformation data on a 1 x 6 meter section of a curved girder web under loading. Figure 2A indicates the out-of-plane distortion of a web without vertical stiffeners. Figure 2B indicates the out-of-plane distortions for a web with vertical stiffeners as indicated in the figure. The effect of stiffeners to constrain web buckling can be easily observed in the figure. Measurements of this type can be used to quantify the effects of attachments on beam performance, quantify local buckling phenomena, and track beam behavior during testing.

Bridge Deck Inspection

The National Bridge Inventory indicates that there are more than 3.2 billion square feet of bridge deck in the United States. The majority of these decks are made of reinforced concrete that provides the driving surface for the bridge. The service life of a deck can be much shorter than the substructure and superstructure of a bridge. Bridge decks deteriorate due to corrosion of reinforcing steel and the resulting delaminations and spalling that can make a deck structurally deficient. The ability to detect this deterioration in its early stages is critical in directing repairs to the most at-risk bridges and will help optimize use of limited funds. However, an effective and practical technology for detecting delaminations is not available, particularly for asphalt covered bridge decks.

To address this need, the FHWA has been active in the development of ground penetrating radar (GPR) systems for the detection and imaging of subsurface defects in bridge decks. In 1998, a system constructed by Lawrence Livermore National Laboratories (LLNL) was delivered to the FHWA for testing. The system, known as the HERMES (High Speed Electromagnetic Roadway Measurement and Evaluation System) consisted of a towable array of 64 GPR transceivers that operated at normal highway speeds. The goal of the HERMES project was to develop a GPR system that can reliably detect, quantify and image delaminations in bridge decks. The density of data collected by the 64 transceivers enabled synthetic aperture radar techniques to be used in the processing of the data, and two and three-dimensional images could be produced. The results of the testing indicated that the system could collect data at highway speeds, and reconstruct that data to generate two and three-dimensional images of the internal features of a bridge deck. However, the sensitivity of the antennas to very thin delaminations in a concrete deck needed to be improved.

Fig 3: PERES-II images of a delamination in a concrete bridge deck.

To make these improvements, a new high-frequency antenna was designed and implemented on a robotic cart capable of scanning over the surface of the bridge deck. The center frequency of the new GPR system, known as PERES II (Precision Electromagnetic Roadway Evaluation System) is 3 GHz (bandwidth ~ 3.2 GHz), providing a short-wavelength pulse capable of detecting and imaging thin features in a bridge deck such as delaminations. Figure 3 indicates a delamination detected during field testing with the PERES II system. The figure depicts a section of bridge deck approximately 0.8 x 3 meters in dimension in plan view. Figure 3A is the GPR image of the bridge deck surface. Notable in this image is a concrete patch in the deck shown as a bright, circular feature in the lower right of the figure. Figure 3B is an image at the depth of the top reinforcing bar (rebar) mesh. Rebars are clearly evident as bright lines in the image, except for an area of the figure where delaminated concrete is indicated by higher amplitude reflection between rebar locations and a loss for contrast for the rebar itself. Figure 3C is the image at the depth of the bottom rebar mesh indicating rebar imaged in the area of the figure outside the delaminated area, and a loss of the rebar image in the area where there is a delamination.

A broader series of field testing conducted with the PERES II systems indicated that while many delaminated areas can be detected with the high-frequency radar, others cannot. The issue of the thickness of a delaminated area of concrete relative to the wavelength of the GPR pulse is believed to be among the most significant issues to be resolved in accurately imaging delaminations. Future field testing is aimed at determining the delamination thickness required for effective imaging and how that may influence the practical application of this technology.

Inspection of Composite Structures

The growing use of composites in civil infrastructure presents many challenges in terms of post-construction inspection for both quality control and monitoring the long-term performance of the materials. The NDEVC is currently developing thermographic methods for evaluating the bond of composite laminates installed as retrofits on concrete structures.

Thermal loading provided by the concrete structure is used to detect anomalies in heat-transfer that occur due to delaminated or debonded sections of composites as the beam is heated during the day. An example is presented in Figure 4, which indicates the thermal signature of a delamination beneath a carbon fiber laminate. The laminate is installed on a prestressed box girder to strengthen the girder. A tap test was conducted and the delamination discovered is encircled in the photo (4A). The IR image of the defect is shown in figure 4B. This defect was later found to be a delamination between concrete patch material, installed to repair the girder prior to the installation of the composite laminate, and the original concrete of the beam. This figure illustrates several significant features of IR testing for composites. First, the image was obtained with a commercially available IR camera, commonly available from several sources. Second, no active heating was used to obtain results. The temperature gradient required to view this delamination was provided by the difference between diurnal warming of the air and the lagging temperature changes in the massive concrete member. Third, this defect in the concrete patch is detected even though it is entirely covered by composite laminate. Finally, the results are available real-time, and the IR image in the figure has not been post-processed in any way.

Fig 4: Thermal images of a delaminated area beneath a carbon-fiber composite laminate. Photo (A) is a video image showing area detected by a tap-test. Images (B) IR image of the same area.

Automated Ultrasonic Testing

Automated Ultrasonic Inspection (AUT) combines traditional ultrasonic testing (UT) methods with computer data acquisition and processing. AUT technologies have been available for several years, and the use of these methods as part of routine inspection practices for pipelines and aeronautical applications is increasing widespread. In general, AUT systems consist of an ultrasonic pulser/receiver, a motion control or tracking system, and supporting computer software that creates ultrasonic images that display the output of the ultrasonic pulser/receiver on a two-dimensional spatial image.

There are several advantages to AUT over traditional UT techniques including the following:

  • Images created by an AUT system can be easier to interpret, especially in areas of complex geometry.
  • AUT system preserve a record of the inspection that can be archived for future use and used to confirm completeness of inspections
  • AUT systems can be automated to provide efficient inspection procedures and reduce the influence of human factors on test results

Both UT and radiographic testing (RT) are used to inspect steel bridges during fabrication to ensure weld quality. The use of RT is more common for steel bridges in the United States than UT, although requirements vary by State and bridge member design. The health issues related to RT testing introduce logistical difficulties in the fabrication process that result in increased cost. However, RT is a well-proven method that provides a more complete record than manual UT. Because of these factors, RT is frequently preferred by bridge owners over UT. AUT provides a more complete record that can be archived, and therefore may provide an safer alternative to RT that meets the needs of bridge owners.

Over the last year, the NDE center has pursued the development and evaluation AUT systems for the fabrication inspection of steel bridges. The goal of the AUT project is to determine if AUT technology can provide an alternative to radiographic testing (RT) as a quality control tool in steel bridge fabrication. The study has examined the use of AUT technology for the inspection of butt-welds during steel bridge fabrication and compared AUT results with the traditional RT technique.

Fig 5: AUT images of a toe crack and a root crack in a butt weld. Crack locations are shown in A, AUT images shown in B and C.

Figure 5 indicates a typical result from the AUT system being used in the research. Figure 5A indicates the cross section of a butt-welded plate with two cracks, one at the toe of the weld and the second emerging from the root of the weld. These cracks were inserted in the plate. A projection C-scan is shown in figure 5B, along with dashed lines that indicate the geometry of the plate bevels prior to welding. Figure 5C indicates the B-scan of the defects, showing (essentially) results in an elevation view of the weld. As evident in the figure, both amplitude information and dimension (length) of the defect is represented, providing key information for defect classification

To date, more than 120 hours of in-line testing with industrial RT services have been conducted, with more than 30 flange welds inspected. Many of these included defects that required repair, primarily weld inclusions. No differences in defect classification between RT and AUT have been realized in this testing, although additional testing on a wider variety of defects may be required to provide full results. A final report detailing the testing conducted under this study is being developed and should be available in the fall of 2003.


The goal of the FHWA NDE Validation Center is to improve the state-of-the-practice for highway bridge inspection. The facility is staffed with a multi-disciplinary team that evaluates the reliability and accuracy of existing NDE technologies and works to develop new technologies. This article has presented a brief overview of some of the current technologies and efforts that are underway at the Center to improve the tools available for the inspection and evaluation of highway bridges. For more information on projects at the NDEVC, the reader is referred to the Center's website at www.tfhrc.gov/hnr20/nde/home.htm.


  1. Moore, M., Phares, B., Rolander, D., Graybeal, B., Washer, G., "Reliability of Visual Inspection for Highway Bridges," USDOT Report FHWA-RD-01-020, Washington, D.C., 2001
  2. Chase, S. "Dynamics and Field Testing of Bridges in the New Millinium: A Look forward," White paper prepared for Transportation Research Board Committee A2C05, January 1999, Washington D.C.
STARTPublisher: DGfZPPrograming: NDT.net