You are on page 1of 9

INTEGRATION OF FWD AND GPR

Robert Parrillo and Roger Roberts Ph D,


Geophysical Survey Systems Inc., Salem, NH
parrillor@geophysical.com, roger@geophysical.com
Gary Sanati
Foundation Mechanics, Inc., El Segundo, CA
gsanati@jilsfwd.com

ABSTRACT
Falling weight deflectometers (FWD) are familiar and essential tools for road evaluation. FWD data is
required by design engineers to establish proper rehabilitation strategies and achieve desired load
capacities. Knowledge of existing asphalt thickness, commonly obtained using intermittent coring, is
necessary for subsequent data processing. Limitations of coring include data density (often 7 times less
frequent than the FWD data), accuracy (the actual asphalt thickness under the FWD load plate is
unknown), expense (coring crew costs approximately $2,000/day), safety (crews must operate in traffic),
and data integration. FWD manufacturer Foundation Mechanics has integrated Geophysical Survey
Systems RoadScan ground penetrating radar (GPR) system with their JILS truck-mounted FWD vehicle.
GSSIs 2 GHz air-launched horn antenna is used to determine asphalt thickness. Typical GPR asphalt
thickness accuracy is within 5% and the system is FCC certified for use in the United States. Advantages
of the integrated FWD/GPR system include cost savings (coring costs are practically eliminated),
accurate thickness information at each FWD location, dual-purpose vehicle (can perform high speed GPR
surveys independent of FWD), and data integration (FWD and asphalt thickness are stored in a single
database).

INTRODUCTION
There are over 3.96 million public centerline road miles in the U.S. and of this, 2.50 million miles (or
about 63 percent) are paved (FHWA, 2002). Pavements are normally classified into two categories;
flexible pavements and rigid pavements. Flexible pavements are asphalt pavements which may
incorporate underlying layers of stabilized or un-stabilized granular materials on a prepared sub-grade.
These types of pavements are called "flexible" since the total pavement structure bends (or flexes) to
accommodate traffic loads. Rigid pavements are portland cement concrete (PCC) pavements, which may
incorporate underlying layers of stabilized or un-stabilized granular materials. Rigid pavements do not
flex appreciably to accommodate traffic loads. Flexible or asphaltic pavements comprise about 82.2
percent of U.S. paved roads while rigid pavements comprise about 6.5 percent [1].
Composite pavements are a combination of hot mix asphalt (HMA) and PCC pavements. Frequently
they are the result of pavement rehabilitation (e.g., HMA overlay of PCC pavement) [1].
Pavement surface deflection measurements are the primary means of evaluating a flexible pavement
structure and rigid pavement load transfer. Deflection measurements can be used in back-calculation
methods to determine pavement structural layer stiffness and the sub-grade resilient modulus. The
general principal is to apply a load of known magnitude to the pavement surface and analyze the shape
and magnitude of the deflection basin to assess the strength of the pavement structure [1].
The most common type of equipment for measuring pavement deflection is the falling weight
deflectometer (FWD). This device applies dynamic loads to a pavement surface, similar in magnitude
and duration to that of a single heavy moving wheel load. The FWD can either be mounted in a vehicle or
on a trailer and is equipped with a weight and several velocity transducer sensors (see Figure 1). To
perform a test, the vehicle is stopped and the loading plate (weight) is positioned over the desired
location. The sensors are then lowered to the pavement surface and the weight is dropped. Multiple
tests can be performed on the same location using different weight drop heights [1].

Truck-mounted FWD

Trailer-mounted FWD
Figure 1 Falling Weight Deflectometers

Deflection variation between test points within a section may be quite large; ranging from 15 percent
to more than 60 percent. This variation reflects changes in layer thickness, material properties, moisture
and temperature conditions, sub-grade support, and contact pressure under the load plate [2].
FWD data, combined with layer thickness information, may be used to obtain the "in-situ" resilient
elastic module of a pavement structure. Selecting the type of rehabilitation to be implemented on a given
pavement is of considerable economic significance and to reach that decision without an adequate
knowledge of the structural condition of the pavement may have very costly consequences [1].
State Departments of Transportation estimate the cost to apply 1 inch of asphalt per lane mile at over
$50,000; making an over-design rehabilitation strategy requiring excess asphalt clearly undesirable.
Conversely, an under-design strategy using insufficient asphalt can result in premature pavement failure
and the need to rehabilitate sooner than expected.
One of the most critical elements in the accurate interpretation of deflection testing results is the use of
accurate pavement layer thickness data in the process. The elastic layer model is extremely sensitive to
such layer thickness. Moduli can vary by several hundred percent, based on assumed thickness of
pavement layers [4]. This is illustrated in Figure 2 by using a 4 inch asphalt thickness, varied by 50%,
plotted against both the corresponding calculated composite stiffness, and the % error in composite
stiffness.
200

% e rro r in c o m p o sit e st iffn e ss

composite stiffness (psi)

2,500,000

2,000,000

1,500,000

1,000,000

500,000

175
150
125
100
75
50
25
0

0
-50

-40

-30

-20

-10

10

% error in asphalt thickness

20

30

40

50

-50

-40

-30

-20

-10

10

20

30

40

50

% error in asphalt thickness

Figure 2 Asphalt thickness error vs. composite stiffness


Traditionally, determination of layer thickness has relied on either construction records or on
destructive (coring) calibration. Either method has now been shown to be capable of significant error.
Construction records have been shown to vary from actual conditions by an order of magnitude or more,
and actual pavement thicknesses have been shown to vary by an order of magnitude within only 100
meters (328 ft) ; a rate of change not detectable by coring procedures [4].

Other limitations of coring include:


Data density (cores are taken about 7 times less frequent then the FWD data);
Accuracy (the actual asphalt thickness under the FWD load plate is unknown and is assumed to be
constant between cores);
Expense (coring crews cost approximately $2,000/day);
Safety (coring crews must operate in traffic);
Data integration (the pavement thickness information is obtained by performing a manual
measurement of the core then manually recording this data for subsequent data entry);
An automated method of measuring and storing pavement thickness at each FWD test location would
offer financial advantages as well as much improved accuracy and safety.
GROUND PENETRATING RADAR FOR PAVEMENT THICKNESS
Ground Penetrating Radar (GPR) is an accepted electromagnetic evaluation technique used for the
transportation infrastructure and a variety of other applications including concrete inspection, utility
detection, geology and archeology. The use of GPR for measuring pavement thickness is a well
established and verified technique.
GPR systems yield accurate data in a form ready for management consideration. They survey
pavements quickly, cost-effectively, and with minimal traffic disruption and safety risks. The Strategic
Highway Research Program (SHRP), the Federal Highway Administration (FHWA), and several states
and other agencies have carried out studies of GPR that demonstrate the advantages of this automated
surveying system. Denmark, Finland, and the United Kingdom are already using GPR in their pavement
evaluation programs, as are several states [3].
GPR systems used for pavement thickness measurement utilize either ground-coupled or airlaunched horn antennas (see Figure 3). As the name suggests, a ground-coupled antenna needs to
remain in contact with ground (or suspended very slightly above the ground) to properly couple the
electromagnetic energy to and from the antenna. This presents some obvious limitations in using the
ground-coupled antenna for high speed pavement surveys on roads in less than perfect condition. More
importantly, to calibrate the ground-coupled system it is necessary to obtain cores from the pavement and
physically measure the actual pavement thickness. The measured thickness value must be entered into
the GPR analysis program so that the appropriate velocity of the radar signal through the pavement
lavers may be derived to determine the pavement thickness. Since the composition of the pavement
changes, it is necessary to obtain cores at regular intervals (1 core per km is one GPR manufacturers
recommendation) to derive accurate pavement thicknesses. Even when the cores are acquired at regular
intervals, the composition of the pavement is assumed to be constant between cores. Therefore, any
change in pavement composition affecting radar signal velocity between cores is a source of error when
using ground-coupled antennas.

Ground-coupled Antenna
Air-launched Horn Antenna
Figure 3 Ground-coupled and air-launched antennas

The air-launched horn antenna is attached to the front or rear of a vehicle and suspended with the
bottom of the antenna approximately 18 inches (45.7 cm) from the pavement surface. Consequently, the
air-launched antenna is routinely used at highway speeds and is not physically affected by rough road
conditions. But most importantly, it is not necessary to obtain cores to calibrate the air-launched horn
antenna system. The system is calibrated by placing a metal plate under the antenna and collecting a
GPR data file. The calibration file data collection process includes recording metal plate reflections at the
different heights that the antenna may experience during data collection over pavement. This metal plate
file is later processed to produce a horn antenna calibration file that is used with subsequent data files to
calculate the velocity of radar signal through the pavement. (A detailed description of the technique used
for horn antenna calibration and data processing is available in Geophysical Survey Systems Handbook
for GPR Inspection of Road Structures.) Thus, when using an air-launched horn antenna with the metal
plate calibration technique the velocity through the pavement, and the corresponding thickness, is
calculated for each individual GPR scan acquired. Typically 8 or 10 scans per foot are acquired for files
collected for the FWD application. In addition, since the metal plate calibration file is applied to each
scan, changes in the composition of the pavement are accommodated as they occur so that the accuracy
of the system is not dependant on the last core location.
Another important advantage of the horn antenna is the ability to measure thin pavement layers.
Since the antenna is suspended above the pavement surface, the direct-coupling (the portion of the
transmitted energy radiated from the transmit antenna directly to the receive antenna) occurs at the
antenna and not at the pavement surface where the ground-coupling occurs. With the ground-coupled
antenna the direct-coupling and ground-coupling occur together, creating near-field interference that limits
the minimum detectable pavement thickness. The 2 GHz air-launched antenna can reliably resolve layer
thicknesses of 1 inch while a 1.5 GHz ground-coupled antenna is normally able to reliable resolve first
layer thicknesses greater than approximately 3 inches.
Geophysical Survey Systems RoadScan product is a complete GPR system for measuring pavement
thickness at highway speeds and assessing the road structure beneath the pavement (base and subbase). The RoadScan system includes GSSIs SIR-20 dual-channel data collection system with
integrated Panasonic Toughbook computer, air-launched horn antenna (1 GHz or 2 GHz), wheel mounted
Distance Measuring Instrument (DMI), universal vehicle antenna mount, and RADAN software with the
Road Structure Assessment Module. The system is FCC certified for use in the United States. The
RoadScan system output includes both a graphic representation of the pavement structure with layer
depths and a corresponding generic ASCII file containing data that is easily shared with other applications
(See Figure 7).
FWD AND GPR INTEGRATION
El Segundo California based Foundation Mechanics manufacturers the JILS truck-mounted FWD
system. In the truck mounted design, the FWD system is integrated into the bed of a truck rather than on
a separate trailer towed behind a vehicle.
In 2004 the Arizona Department of Transportation, an owner of a JILS truck mounted FWD system,
expressed their desire to implement a method of collecting and storing pavement thickness information
during the FWD data collection process.
Several years earlier Foundation Mechanics had experimented with integrating GPR with FWD for the
California Department of Transportation. However, these efforts were performed using ground-coupled
antennas (See Figure 4). The ground-coupled antenna needed to be raised and lowered. GPR data
collection could only occur at very slow speeds, which meant that high speed GPR surveys, independent
of FWD, could not be performed. Additionally, it was not possible to measure thin overlays and it was still
necessary to obtain cores for GPR calibration.

Figure 4 Earlier efforts integrating FWD & GPR using ground-coupled antennas
In a collaborative effort between Foundation Mechanics and Geophysical Survey Systems, the
RoadScan GPR system was integrated with the truck mounted FWD system. The result is a single
vehicle that collects both FWD and pavement thickness data (see Figure 5). The 2 GHz air-launched
horn antenna is attached to the front of the vehicle. The antenna is easily installed or removed as
required so that it is not necessary for the antenna to remain on the vehicle before or after data collection.
The GPR and FWD systems share the same laptop computer for a single control station. The pavement
thickness directly under the FWD load plate is determined and stored with the FWD data in a single
database.

Figure 5 Foundation Mechanics JILS Truck-Mounted FWD with integrated GPR


The combined FWD/GPR data collection process is similar to normal FWD operation. Approximately
50 feet before the vehicle reaches the desired FWD test location the GPR data collection process is
initiated from the same laptop computer that operates the FWD system. A short GPR file is collected
while the vehicle slows and reaches the FWD test location (see Figure 6).

Direct Coupling at antenna


Top of pavement surface
Bottom of asphalt overlay
Bottom of pavement layers

Bottom of base layer

Figure 6 - Short run GPR file (10 scans per foot)


The raw GPR file may either be processed immediately after collection or at later time. Processing
the GPR file consists of first applying a stored metal plate calibration file and then by identifying the
desired layers. Identifying the layers is performed in an interactive interpretation mode using a simple
innovative point and click technique called EZ Tracker. By creating a simple processing macro in GSSIs
RADAN software, the GPR processing may be fully automated. The output consists of both a graphic
display of the road structure and an ASCII output file (see Figure 7).
ASCII output file

Time-based radar data

Depth-based layer data


Figure 7 Processed GPR file with corresponding ASCII output
The JTEST FWD software program imports the GPR ASCII file to obtain the layer thickness
information. This may be accomplished either before or after FWD data collection.
The JTEST software provides the operator with a simple method for importing processed GPR data
prior to FWD data collection (see Figure 8). Once imported, the JEST software not only displays the
pavement layer thickness under the FWD load plate but also provides supporting information including
graphs of the pavement layer thickness, radar velocities, statistical confidence indicators, and an
interactive spreadsheet view of the GPR data where advanced users may experiment with and specify
data threshold values.

Figure 8 GPR results loaded prior to FWD data collection


The other (and perhaps better) alternative is to collect all the GPR and FWD data then process and
import the GPR data into the FWD database sometime later. This approach results in a shorter overall
data collection time. Collecting the GPR data does not add any significant data collection time since the
GPR data is collected as the vehicle approaches and stops at the FWD test location. After the GPR data
has been processed, the FWD database is opened in JTEST. The appropriate FWD test location is
selected and the corresponding GPR results file is specified. The JTEST software then imports the
specified GPR data (see Figure 9).
Using either method, the result is a single FWD database file with pavement thickness information
stored at each FWD test location.

Figure 9 GPR results loaded after FWD data collection


In an evaluation in Arizona, 24 individual cores and 24 individual GPR scans were performed on two
separate highways. In 12 of the locations, the asphalt was overlaid with a thin friction course. The asphalt
thickness obtained from the core data ranged from 2.88 to 7.13 inches. The average difference in asphalt

thickness between the measured cores and GPR calculated thickness was less than 4%. Although
occasional coring is still required to obtain material properties, the evaluation study confirmed that GPR is
a reliable method for obtaining accurate asphalt thickness without coring.
An additional benefit of the FWD/GPR integration is the ability to use the same vehicle to perform
long distance pavement studies at highway speeds. For this application, the RoadScan GPR system is
used independently of the FWD system. The RoadScan system accommodates an optional GPS system
used to collect GPS data simultaneously with the GPR data. The results may then be plotted to produce
a graphic representation of the pavement thickness using contour mapping software such as Surfer by
Golden Software or DPlot by HydeSoft Computing (See Figure 10).

>4
3.75
3.5
Pavement Thickness

3.25

0-2.49"

2.5-3.49"

2.75

3.49-5"

2.5
2.25
2
1.75
< 1.5

Figure 10 GPS data integrated with GPR thickness data


CONCLUSIONS
The successful integration of GSSIs RoadScan GPR system with the Foundation Mechanics truckmounted FWD resulted in a flexible vehicle satisfying the customers requirements to collect and store
pavement thickness data with the FWD data.
The benefits of the integrated FWD and GPR system include:
Cost Savings Coring costs are dramatically reduced. The dual-purpose vehicle provides
combined FWD and GPR data collection without significantly impacting FWD data collecting time;
Safety Less time for personnel operating in dangerous traffic environments to obtain cores;
Data Integration FWD and pavement thickness data are integrated into a single database;
Data Quality Pavement thickness at each FWD location provides designers with the data they
need to make confident maintenance and rehabilitation decisions; and
Flexibility The FWD vehicle may be used for independent project and network level GPR
surveys at highway speeds without the need or expense of traffic control.
Following the initial implementation at the Arizona DOT, the integrated technology has been adopted
the City of Los Angeles and Montana DOT who have experienced similar success (See Figure 11).

Figure 11 City of Los Angeles

Montana DOT

REFERENCES
[1] Washington State Department of Transportation Interactive Pavement Guide
www.wsdot.wa.gov/biz/mats/pavement/fwd.htm
[2] LTPP Manual for Falling Weight Deflectometer Measurements Operational Field Guidelines
Version 3.1 August 2000
[3] Ground Penetrating Radar For Measuring Pavement Thickness, U.S. Department of
Transportation, Federal Highway Administration, Publication Number: FHWA-HIF-00-015 FHWA
Publication Number: FHWA-HIF-00-015
[4] Eckrose, R. A., Ground Penetrating Radar Supplements Deflection Testing to Improve Airport
Pavement Evaluations Nondestructive Testing of Pavements and Back-calculation of Moduli,
ASTM STP 1026. A. J. Bush III and G. Y. Baladi, Eds., American Society for Testing and
Materials. Philadelphia. 1989. pp. 563-573

RECOMMENDED FURTHER READING


Noureldin, A.S., Zhu, K., Li, S., Harris, D. Network Pavement Evaluation Using Falling Weight
Deflectometer and Ground Penetrating Radar, Transportation Research Board 82nd Annual Meeting,
Washington, D.C.
Al-Qadi, I.L., Lahouar, S. Jiang, K., MeGhee, K. K., Mokarem, D., Validation of Ground Penetration
Radar Accuracy for Estimating Pavement Layer Thickness, Paper No. 05-2341, Transportation
Research Board, 84th Annual Meeting, January 9-13, 2005, Washington, D.C.

You might also like