Bibliography Glossary Geophysical
About Engineering Applications Geophysical Methods Glossary Bibliography
Vehicle Mounted Ground Penetrating Radar Systems

Basic Concept: Vehicle mounted GPR is a rapid, continuous data collection system that obtains highly accurate and repeatable data as it is moved over a pre-defined path. These GPR systems are used for condition evaluation of new and existing decks and pavements.

Data Acquisition: There are basically three primary GPR methods used for data collection and analysis:

GPR Method One: (High-Resolution, High-Speed Survey (48-72 km/h) Using Dual-Polarization of 1 GHz Horn Antennas

This method employs two 1.0 GHz air-coupled (horn) antennas (sensors), mounted behind the vehicle-in line, and in a manner where one antenna is responsive to signals influenced primarily by the longitudinal steel, and the other is responsive primarily to the transverse steel. In Method One, signals from both sensors are used, with their offset in longitudinal position along the same path accounted for, to simulate a higher resolution, ground-coupled antenna of Method Two used in the lower speed, high-resolution surveys.

Figure 72 shows the dual-polarization deployment of Method One (high speed, high-resolution), with the antennas mounted in-line. Currently, antennas are positioned for data collection in the right half of any given lane. Depending on current lateral location of the laser (or camera) positioning system, this survey line will be collected either 0.3 m, 0.9 m, or 1.5 m to the left of any given pavement stripe (lane marker) on which the laser (or camera) is focused. Lateral positioning of lines is within ±5 cm on any desired spacing interval.

Ground Penetrating Radar Method One:  
	Dual-Polarization horn antenna method (antenna configuration):  Note two 1 GHz antennas, polarized 90-degrees apart.  
	The front antenna (nearest the vehicle) is sensitive to longitudinal bars; the rear antenna is most sensitive to 
	transverse steel (high speed, high-resolution).  (GEOVision Geophysical Services)

Figure 72. Ground Penetrating Radar Method One: high speed, high resolution,
Dual-Polarization horn antenna method. (GEOVision Geophysical Services)

GPR Method Two: (High-Resolution, Low-Speed Survey (8-16 km/h) Using One or More 1.5 GHz Antennas Each Collecting Data on Unique Scan Paths

This method uses a single 1.5 GHz antenna traveling along each survey line (although several can be placed beside each other at a maximum spacing of 60 cm) in a manner where the transverse steel in the top mat is crossed perpendicularly, and its reflection amplitude is accurately measured. It requires no signal modifications, and simulation of a higher frequency antenna is not needed because this antenna has been determined to be the most sensitive to conditions within a deck that are coincidental with corrosion-induced d amage, and it easily discriminates between the signals generated from the desired transverse steel and signals generated from longitudinal steel (undesired) in the top mat of reinforcement.

GPR Method Two is practical and/or economical only on relatively short bridge decks, and in rural locations or limited metropolitan areas where either light or moderate traffic volume and lane closures do not pose problems for transportation officials (figure 73). It provides highly accurate GPR survey results whose deterioration predictions for quantity, location, and extent cannot be exceeded. However, its slower data collection speed necessitates lane closures, making it less attractive in congested urban areas or on network-level projects where many decks have to be assessed in a short period of time, or at a lower cost. Survey speeds are limited to about 8 km/h on most bridge decks, but can be increased to about 16 km/h when deck surfaces are free of debris, potholes, or rough patches that might damage the antenna, or de-couple it from the deck surface. Also, this method requires the shallowest reinforcing bars in the upper steel mat be oriented perpendicular to the survey direction so that unbiased and accurate imaging of this layer and measurement of its properties can be performed. It is economically limited to those decks where transverse steel is tied above longitudinal steel in the upper reinforcing mat.

Ground Penetrating Radar Method Two:  
	low speed, high-resolution) 1.5 GHz ground-coupled, multiple antennas (antenna configuration).  
	(GEOVision Geophysical Services)

Figure 73. Ground Penetrating Radar Method Two: low speed, high-resolution, using four
1.5 GHz antennas spaced equal distance apart. (Geophysical Survey Systems, Inc.)

GPR Method Three: Lower-Resolution, High-Speed Survey (32-72 km/h) Using Single-Polarization 1 GHz Horn Antenna.

A distinctly separate high-speed GPR survey, using a single-polarization deployment of a single, 1.0 GHz horn antenna (or multiple antennas that are mounted on booms so that the sensors do not scan on the same paths), is probably the most common method of GPR evaluation. This method has been around the longest and provides good, overall results. It is used to rank a large number of decks based on a condition rating (good, fair, poor, etc.), and it provides fairly accurate deterioration quantity assessments that can be used to help decide among various maintenance/rehabilitation strategies on individual projects, or estimate project budgets.

Like the high-speed, high-resolution GPR survey that uses a dual-polarization deployment of two-horn antenna, a single-polarization deployment has the advantage of speed. More deck coverage, often several times as much (or many decks), can be surveyed in a single day even if the decks are open to traffic or separated by several kilometers of roadway. Unlike the dual-polarization method, however, accurately locating zones of deterioration with a high degree of confidence and pinpointing the boundaries between sound and poor concrete is not a reliable outcome of a single-polarization (high-speed) survey.

All three methods are useful for different applications, with Method One being the most practical for all-purpose GPR work, although there are some variables affecting economic application of all methods that may impact why one method is selected over another. These will not be covered here. Some general applications are highlighted:

Data Interpretation: GPR Method One yields nearly identical information to the Method Two 1.5 GHz ground-coupled sensor data, yet surveys can be carried out without the need for lane closures at speeds between 48 and 72 km/h. Typically, a chase vehicle following the GPR survey vehicle is the only traffic control requirement in most states. Figure 74 compares the two methods.

Comparison of Ground Penetrating Radar 
	Methods with "ground truth":  (a) method two (low speed, high-resolution, 1.5 GHz), (b) method one 
	(high speed, high-resolution—dual polarization), and (c) delamination map produced from hammer-sounding 
	the deck after overlay was removed.  Both Ground Penetrating Radar data sets were collected and predicted 
	prior to asphalt removal.

Figure 74. Comparison of Ground Penetrating Radar Methods with "ground truth": (a) method two (low speed, high-resolution, 1.5 GHz), (b) method one (high speed, high-resolution-dual polarization), and (c) delamination map produced from hammer-sounding the deck after overlay was removed. Both Ground Penetrating Radar data sets were collected and predicted prior to asphalt removal.

Advantages: Because of its unmatched imaging capabilities and speed of operation, GPR is the preferred method for condition evaluation of new and existing decks and pavements.

Limitations: FCC limitations restrict the use of the data acquisition units purchased after July 15, 2002, to 100 KHz; therefore no new vendors in the U. S. will have this capability until these regulations are relaxed. Also, it prevents the only manufacturer of the equipment and software for GPR Method One from selling data acquisition units (in the U. S.) that can function at the high data collection rates required to perform this survey (upwards of 400 scans/second/channel on a two-channel system that can operate two antennas simultaneously).

Understanding What a GPR Survey Cannot Do: Once this is understood, the basics behind what a well-designed survey can and will do are recognized, expectations will begin to match capabilities, and more confidence in the technology will begin to be established, as it should.

Although GPR can be very successful at identifying delaminated areas on a deck, or its results can correlate extremely well with half-cell corrosion potential tests, even maintenance reports that validate the accuracy of its prediction capabilities, it cannot distinguish among any of the following symptoms of an active corrosion process:

  1. Corroded reinforcement.
  2. Delaminations, unless water-filled (and 0.3 cm (1/8-inch) thick or greater) at the time of the survey.
  3. Spalling, unless water-filled and 0.3 cm (1/8 in) or thicker during the time of the survey.
  4. Punky, deteriorated concrete that has lost much of its compressive, tensile and shear strength.
  5. Otherwise moist, chloride-impregnated concrete.

It is important to make these distinctions, so that realistic expectations of survey results will predominate the current school of thought within the industry. Currently, most people accept that GPR is used to find delaminations, and that is an incorrect assumption. Rather, GPR is an ideal instrument for rapidly identifying that the environment for corrosion-induced damage exists within a deck, and the right methods can also be used to pinpoint and establish deterioration threshold levels and boundaries on a partially or almost fully deteriorated deck. When used as the primary NDT method in a comprehensive diagnosis, it provides important information that allows other NDT or destructive sampling to be utilized in a much more economic, efficient, and accurate manner, further improving the quality of the GPR results.

Proper Use of GPR Results With Other NDT Methods or Destructive Sampling

The value of all three GPR methods is significantly improved and should always be performed with additional "ground truth" (better called destructive examination) or nondestructive testing as supplementary information. A calibration of sorts can then be used to validate an initial interpretation and determine an accurate deterioration threshold, which will then be applied to the resulting relative deterioration map for an improved, final interpretation (contour map). Otherwise, expect the accuracy in the ability of GPR to estimate overall deterioration quantity on a deck and define locations of deteriorated zones and determine their physical boundaries to diminish, often significantly.

Typically, coring, half-cell corrosion potential testing (and contour-plotting of the results), or other destructive techniques, such as chloride ion content sampling, are performed in significantly limited quantities using guidance from the GPR results. Other NDT methods, and visual underside inspections of the deck, may also be used for this purpose, complementing the GPR data by providing information that supports or refines the initial GPR analysis. These geophysical methods may include ultrasonic seismic (SASW) techniques, such as ultrasonic body-wave (UBW), ultrasonic surface-wave (USW) and impact-echo (IE), alone or in combination. Although half-cell corrosion potential is typically thought of as a semi-destructive analysis technique because the lead at the anode is physically attached to the upper reinforcement mat for good electrical contact, it qualifies as a geophysical method because of the nature of its underlying principles. New equipment under evaluation does not require drilling into the deck to attach the (anode) lead so that it truly becomes a nondestructive, geophysical technique.

IMPORTANT: Using previously obtained data from cores or other sampling/testing results generally does not enhance the GPR analysis much, if at all. Sometimes, depending on the data, a GPR analysis is better off without it. In order for these data to be useful, they must be obtained after the GPR survey takes place and must be sampled (a) in regions where GPR contour plots show relatively uniform contour levels (not steep contour gradients), (b) at several contour levels representing varying degrees of relative deck deterioration on these initial plots, (c) at control points where little or no deck deterioration is initially expected to exist based on the GPR data and the experience of the interpreter, and (d) redundant contour levels representing (a), (b) and (c) above.