Al-Quadi Paper
Al-Quadi Paper
Abstract
Introduction
frequency, the transmitted signal can be of any shape and can have any
desired frequency, which will allow for a better penetration and
recognition of some detailed features. However, the data acquisition rate
is very limited. A pulsed system transmits a short time domain EM pulse
at a well-defined pulse repetition frequency and then records the pulses
reflected from the various layer interfaces.
The pulsed systems are the most used and are the most commercially
available because of their ease of use. The principle of the impulse
systems is based on transmitting an EM pulse and analyzing the reflected
pulses from the layer interfaces where there is a contrast in the dielectric
properties. Figure 1 shows the major components of a pulsed radar
system. A sequence of trigger pulses is generated in the control unit by
the radar circuitry. These trigger pulses are sent through the control
cable to the antenna, where each trigger pulse is transformed into a
bipolar transmit pulse. These transmitted pulses encounter different
materials in the subsurface with different dielectric properties. At the
layer interfaces, the incident pulses are reflected back to the surface,
where they are collected by the receiving antenna and are then sent to the
control unit for processing and display.
Ground penetrating radar technology has been used for the past 20
years for a variety of applications to assess pavement performance (2-6).
Yet the main issue after all these years remains: “How well does GPR
work and under what conditions?” Results show that GPR works well
for some situations but is not an appropriate tool for other situations. It
is not currently used on a routine basis by the U.S. Departments of
Transportation primarily because of the difficulties encountered in data
interpretation, as well as the expenses involved in conducting GPR
surveys. While it is expected that there will be a growing interest and
Al-Qadi, Lahouar
demand for GPR surveys, a number of limitations exist that are related
mainly to interpreting the results of GPR testing:
• The images obtained from the reflected signals are not
photographs of the features that are beneath the surface being
investigated. The images show the amplitude of the radar-
reflected signals from the interfaces with different dielectric
properties. These amplitudes are plotted in colors using user-
defined color codes. Therefore, a considerable amount of
experience and operator skill may be required to interpret sub-
surface radar results correctly.
• Extensive amount of data.
• Determination of the exact location of a reflecting feature
beneath the surface relies upon a prior knowledge of the
dielectric properties of the material.
• Change of the dielectric constant with depth mainly because of
the presence of moisture.
• Losses in the pavement materials especially with the presence of
moisture or conducting subgrade soils.
• Reflections from thin layers may overlap depending on the GPR
system resolution.
This paper explores the feasibility and necessary conditions for using
GPR as a nondestructive technique to assess flexible pavements. The
paper also presents the analysis of GPR data collected from different
flexible pavement systems.
The GPR system used in this research was a SIR-10B control unit
that could be connected to air-coupled and ground-coupled antennas. An
air-coupled antenna was used in the first phase of this study. The
antenna was composed of a pair of separate horn antennae (one serves as
a transmitter and the other as a receiver) with a frequency bandwidth of 1
GHz, which corresponds to a pulse width of 1 nanosecond. As depicted
in Figure 2, the antenna was mounted behind the survey van with the
control unit set inside. To precisely locate the collected GPR data
longitudinally on the road, a high-resolution distance-measuring
instrument (DMI) connected to the survey-vehicle wheel was used to
control the trigger pulses generated by the GPR system and, therefore, to
control the spatial GPR data acquisition frequency. In this case, data
were collected as a function of distance (i.e., n scans every meter) and,
thus, was independent of the survey speed.
Use of GPR
HMA Unbound
OGDL CTA
Sec. WS BM-25.0 SM-9.5A (mm) Aggregate
(mm)
(mm) (mm) (mm) (mm)
A 38 150 - 75^ 150 175
B 38 150 - 75^ 150 175/ GT
C 38 150 - 75^ 150 175/ GT
D 38 150 - 75^ 150 175/ GT
E 38 225 - - 150 75/ GT
F 38 150 - - 150 150
G 38 100 50 - 150 150/ GT
H 38 100 50 75^ 150 75
I 38* 100/RM 50 75^ 150 75
J 38 225 - 75^/MB - 150
K 38 225/SR - 75+ - 150
L 38 150/RM - 75+ 150 75
* High laboratory compaction ^ Asphalt treated + Portland cement
treated WS: Wearing surface; SR: Stress Relief Geosynthetic; GT:
Woven Geotextile/Separator; RM: Reinforcing Steel Netting; MB:
Moisture Barrier
12000
10000
Surface Reflection
8000
HMA/Base Reflection
6000
HMA Base
Amplitude
4000
2000
0
-2000
-4000
-6000
-8000
0 2 4 6 8 10 12 14 16 18 20
Time (ns)
Surface Reflection
5
4
3
Spurious Reflection
2 OGDL/21-A Reflection
1
FIGURE 4 Linescan view showing the Reflections
Multiple copper plate reflections at the
(1) 21-B/ Subgrade, (2) 21-A/21-B, from
(3) OGDL/21-A,
copper plates (4) BM-
25.0/OGDL, and (5) WS/BM-25.0 interfaces
Figure 5. Linescan View Showing the Copper Plate Reflections at the
(1) 2-B/Subgrade,(2) 21-A/21-B, (3) OGDL/21-A, (4) BM-25.0/OGDL
and (5) WS/BM-25.0 Interfaces
cti
di =
2 ε r ,i
(1)
where:
di is the thickness of the ith layer,
ti is the EM wave two-way travel time through the ith layer as
shown in Figure 6,
c is the speed of light in free space: c = 3×108 m/s,
εr,i is the dielectric constant of the ith layer.
Tx/Rx
A0
A1
A2
Time
Subgrade, εr,3
Similarly, for the third layer the dielectric constant εr,3 is found
according to equation (4).
2
A 2 A1 A2
1− 0 + γ +
A 1
Ap Ap
εr ,3 = εr , 2 p
(4)
2
A A1 A2
1 − + γ1
0
−
Ap Ap Ap
where:
εr,3 is the dielectric constant of the third layer,
A2 is the amplitude of the reflection obtained at the interface
between the second and third layers as shown in Figure 5.
γ1 is the reflection coefficient at the interface between the first
and second layers (9) and is given by Equation (5).
ε r ,1 − ε r , 2
γ1 = (5)
ε r ,1 + ε r , 2
The same process can be repeated iteratively to compute the
dielectric constant of the nth layer using Equation (6).
Al-Qadi, Lahouar
2
A 2 n−2 Ai An−1
1− 0 + γ
A ∑
+
Ap
i
p i =1 Ap
ε r ,n = εr ,n-1 2
(6)
A n −2 Ai An−1
1 − + ∑ γi
0
−
Ap Ap Ap
i =1
section was a good candidate for testing the effects of thin layers on GPR
thickness estimation performance because the pavement was more than
two years old and the layers had different densities—thus resulting in
different dielectric constants. For example, among the three layers, the
OGDL layer had the highest air-voids and lowest asphalt content.
As shown in Figure 4, it is difficult to find the reflection locations of
the individual layers that compose the HMA layer. This problem is
caused, on one hand, by the overlap between the reflected pulses from
the WS/BM-25.0 interface and the pavement surface and, on the other
hand, by the reflected pulses from the BM-25.0/OGDL interface and the
OGDL/base interface. Therefore, using the traditional thickness
estimation technique from GPR data would yield the total thickness of
HMA instead of the thicknesses of the individual layers. The total
thickness of the HMA layer found from GPR data along the test section
is presented in Figure 7. For comparison purposes, the HMA design
thickness for the same section is also shown in the figure. According to
Figure 7, GPR results tend to overestimate the design HMA thickness by
approximately 30 to 80 mm, which represents 10 to 30 percent of the
total design thickness. Based on comparison of the GPR results to
thicknesses measured directly on cores taken from the test section, the
HMA thickness overestimation is only 11 percent. This overestimation
is mainly caused by the following:
• The assumption that the three HMA layers have the same
dielectric constant even though they have different compositions.
• The dielectric constant computation technique, which is based on
the amplitude of the surface reflection, as given by equation (2).
In fact, because of the overlap between the surface reflection and
the reflection at the WS/BM-25.0 interface, the measured
amplitude of the surface reflection would be greater or less than
the real reflection amplitude. In other words, because of the
different peak polarities of the transmitted GPR pulses (positive
peak surrounded by two negative peaks), the superposition of the
two adjacent reflections might have an additive or a subtractive
effect, which will either increase or decrease the measured
surface reflection amplitude.
finding the exact reflection locations, equations (2) to (7) can be used to
iteratively estimate the dielectric constants of the different layers. Then,
Distance (m)
0 10 20 30 40 50 60 70 80 90 100
0
100
150
Depth (mm)
200
250
300
350
400
equation (1) can be used to find the thicknesses of the individual layers.
It should be noted that applying deconvolution to the reflected GPR
signals usually produces an output signal with an increased noise level.
Because a high noise level generally makes it more difficult to detect
reflected pulses, the probability of missing some interfaces might be high
in this case. The effects of the additive noise can be reduced
considerably by passing the reflected GPR signal through appropriate
digital filters during a preprocessing phase. Furthermore, for the
deconvolution operation to succeed, the coupling pulse, usually inherent
in air-coupled GPR data, should be removed by performing a template
matching followed by a subtraction operation of the coupling pulse
shape. The coupling pulse is collected during the GPR survey by
pointing the antennas towards the sky.
Figure 86 shows the HMA thicknesses of the WS, BM-25.0, and
OGDL layers found from the same GPR data as before but after applying
deconvolution and using the aforementioned thickness estimation
technique. For comparison purposes, the figure also shows the design
thicknesses for each layer. According to the figure, the thicknesses
estimated by GPR are very comparable to the design thicknesses.
Moreover, when comparing the GPR estimated thicknesses to the
thicknesses measured directly on cores, as shown in Table 2, an average
thickness error of 21.1 percent is found for the WS layer, 0.6 percent for
Use of GPR
the BM-25.0 layer, 1.0 percent for the OGDL layer, and 3.7 percent for
the overall HMA layer. These results show the improvement made in the
thickness estimation accuracy when the pavement’s individual HMA
layers were considered instead of being assumed as a single layer.
Distance (m)
0 10 20 30 40 50 60 70 80 90 100
0
50
100
150
Depth (mm)
200
250
300
400
Figure 9
7 illustrates a simple configuration that can be used to estimate
the velocity v in a layer, simultaneously using monostatic (collocated
transmitter and receiver) and bistatic (separate transmitter and receiver)
Al-Qadi, Lahouar
vt1 = 2h (9)
2
x
vt 2 = 2 h 2 + (10)
2
where h is the thickness of the layer, t1 and t2 are the two-way travel
times of the monostatic and bistatic systems, respectively, and x is the
distance between the transmitter and receiver of the bistatic system.
x
T T/R R
t1
HMA t2 h
εr1
P
Figure 9. Common Midpoint Geometry using Ground-Coupled
Monostatic and Bistatic Systems
c x
v= = (11)
εr t − t12
2
2
Use of GPR
Air-
coupled
Ground-
coupled
x0
T R
Air
εr0=1
h0
θi x1
T/R
t1
HMA θt h1
FIGURE 9 Modified common midpoint geometry using a gr
εr1 t2
P
Figure 11. Modified Common Midpoint Geometry Using a Ground
Coupled Monostatic System and an Air-Coupled Bistatic System
Distance (m)
0 282 536 784 1024 1269 1499
0
100
200
Thickness (mm)
300
400 Bridge
500
600
Figure 12. HMA thickness, I-81 North, passing lane from milepost 145
(0 m) to milepost 146
Distance (m)
0 233 469 689 912 1153 1378
0
100
200
Thickness (mm)
300
400
Full depth repair
500
600
Figure 13. HMA thickness, I-81 South, passing lane from milepost
153 (0 m) to milepost 152
easier than in the case of thin layers. Furthermore, it guarantees that the
detected layers are homogeneous, thus reducing thickness-measurement
errors caused by dielectric constant variations within the layer. For rigid
pavements, the condition of thick layers is usually verified for both the
concrete slab and the supporting base layer. In contrast, for flexible
pavements, the HMA layers are usually composed of at least one thin
layer, which is the wearing surface or any newly placed overlay.
Therefore, the thick layers condition is generally not applicable for
flexible pavements. However, in the case of newly built and non-aged
flexible pavements, the HMA layers could be considered as a relatively
single homogeneous thick layer, especially when they are composed of
the same aggregate type (13).
To validate flexible pavement layer thickness estimation using GPR,
different GPR surveys were conducted in a quality control-quality
assurance (QC/QA) study of a newly-built, three-lane pavement section
of Route 288 in Richmond, Virginia (14). The pavement system was
composed of three HMA layers (HMA base, HMA intermediate 1, and
HMA intermediate 2), with a total thickness of 240 mm and a 150-mm
aggregate base layer. In order to ensure that the HMA layers were
relatively homogeneous for the GPR survey, the GPR measurements
were conducted on the HMA pavement layers just a few hours after they
were placed.
Figure 12 shows a comparison between the GPR estimated
thicknesses, calculated using the data analysis technique presented in
equations (1) through (7), and the design thicknesses for part of the test
section of Route 288. The layer thicknesses were estimated from GPR
data collected over the different layers just a few hours after their
construction. As shown in Figure 12, the GPR estimated thicknesses are
very comparable to the design thicknesses. To verify the accuracy of the
GPR data analysis, stationary GPR measurements were collected at
specific locations where some cores were taken for density and material
property tests. The cores were also used to directly measure the HMA
layer thicknesses and to compare them to the thicknesses estimated by
GPR.
Figure 13 depicts the correlation between the layer thicknesses
estimated by GPR and the thicknesses measured directly from cores for
the three HMA layers. This figure shows that the layer thicknesses
found by both techniques are very comparable (data points congregated
around the line of equality). Actually, the average error between GPR
thickness and core thickness for all three HMA layers was 2.9 percent.
This error is similar to the error found when measuring layer thicknesses
directly on cores (13).
Al-Qadi, Lahouar
Distance (m)
40 45 50 55 60
0
50 HMA
Base
100
HMA Design
150 Base Design
Thickness (mm)
200
250
300
350
400
450
280
260 HMA Base
HMA Intermediate 1 Avg. Error = 2.16%
240
HMA Intermediate 2
GPR Thickness (mm)
220
200
180
Avg. Error = 2.90%
160
140
120
Avg. Error = 3.71%
100
80
80 100 120 140 160 180 200 220 240 260 280
Core Thickness (mm)
Conclusions
Recommendations
References
Discussion