Download as pdf or txt
Download as pdf or txt
You are on page 1of 8

Inspection and Condition Assessment

Using Ground Penetrating Radar


Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY-HYDERABAD on 06/18/13. Copyright ASCE. For personal use only; all rights reserved.

Dar Hao Chen, Ph.D., P.E.1; and Andrew Wimsatt, Ph.D., P.E.2

Abstract: The nondestructive mapping of anomalies and voids under roadway pavements is critical to highway authorities because of the
potential loss of support that would lead to safety hazards. 400 MHz ground-coupled penetrating radar 共GCPR兲 was used in this study to
characterize the subsurface conditions of three roadway pavements 共SH359, IH40, and U.S. 290兲. The extents of the anomalies in
horizontal and vertical directions were visible in GCPR images. Coring, boring, and lab testing were performed to verify the settlement
and source of the moisture on SH359. The source of the moisture was from the leaking water pipe, as indicated by the high chloride and
chlorite contents. A 1.8-m deep void 共3.8 m3 in volume兲 under IH40 and a 1.8 m ⫻ 4.6 m ⫻ 3.7 m 共30.6 m3 in volume兲 void under U.S.
290’s reinforced concrete pavements were successfully identified by GCPR and verified by field boring and coring. Fortunately, the voids
near the drainpipes were detected by GCPR in time. Otherwise, the void would have increased in size, and that could have led to a severe
hazard. This study has successfully demonstrated that the GCPR is able to identify anomalies and voids. Therefore, engineers can utilize
the information from GCPR to undertake remedial actions with confidence.
DOI: 10.1061/共ASCE兲GT.1943-5606.0000190
CE Database subject headings: Void; Radar; Nondestructive tests; Inspection; Pavements.
Author keywords: Void; Ground penetrating radar; Nondestructive testing; Anomaly.

Introduction Chen et al. 2007兲. Leaks, pipe breaks, or dislocated joints allow
for fines to be carried away, resulting in local base or foundation
In order to preserve the highway infrastructure and to provide erosion, and the formation of weak areas, and eventually voids.
safe transportation facilities, their condition must be known. This These voids keep increasing in size until the load carrying capac-
means that condition assessment procedures should be optimized ity of the roadway is compromised. The development of voids
so that any deterioration 共whether caused by external impacts beneath roadways is a serious safety hazard, and detecting the
and/or hidden construction defects兲 can be detected at an early presence and extent of subsurface voids under a roadway struc-
stage. Nondestructive testing 共NDT兲 has advanced to a level ture is critical in preventing tragedies from occurring. Among the
where the subsurface condition of a roadway can be diagnosed range of available tools, ground penetrating radar 共GPR兲 has been
with confidence 共Helmerich et al. 2008兲. Thus, only a few used by many researchers and practitioners 关Federal Railroad Ad-
samples 共collected destructively through coring or auguring兲 are ministration 共FRA兲 2005; Lesto and Hatcher 2002; Chen and
needed for verification. The advantage of combining NDT and Scullion 2008兴. For example, GPR has been used successfully in
field verification is that it provides a comprehensive evaluation of a variety of highway applications, including: 共1兲 detecting air-
subsurface conditions throughout the entire project, not only at filled and water-filled voids; 共2兲 locating subsurface vertical
locations where coring or auguring are performed. Other advan- cracks; 共3兲 locating subsurface anomalies including buried ob-
tages are the speed of the data collection and the immediate avail- jects, peat deposits, and near-surface bedrock; and 共4兲 analyzing
ability of the results. rutting mechanisms 共Loken 2007兲. Air voids and water-filled
Normally, voids develop because of subsidence and erosion of voids are both detectable using GPR because the dielectric con-
the base and subgrade materials. Historically, void-related road- stants of both air 共=1兲 and water 共=81兲 are substantially different
way problems have often developed near water supply pipes or than most pavement materials 共Chen and Scullion 2008兲. If the
drainpipes 共Hauser and Howell 2007; Chamberlain et al. 2000; void is air-filled, a large negative peak will appear in the wave-
form, since the dielectric constant of air is much lower than for
1 any pavement material. A large positive peak in the waveform
Professor, Changsha Univ. of Science and Technology, Chiling Rd.
45#, Changsha, Hunan 410076, People’s Republic of China; and, Texas will appear if there is a water-filled void, because the dielectric
Dept. of Transportation, 4203 Bull Creek #39, Austin, TX 78731 共corre- constant of water is much greater than for pavement material.
sponding author兲. E-mail: dchen@dot.state.tx.us Tomographic imaging of subsurface pavement features using
2
Division Head, Materials and Pavements Div., Texas Transportation GPR has been pursued by the Texas Department of Transportation
Institute, 501 CE/TTI Building, Texas A&M Univ., College Station, TX 共TxDOT兲 for over two decades. The depth of GPR imaging de-
77843. E-mail: a-wimsatt@ttimail.tamu.edu
pends on soil type and GPR antenna frequency. For example, clay
Note. This manuscript was submitted on January 23, 2009; approved
on June 22, 2009; published online on July 1, 2009. Discussion period soils with high moisture content will quickly attenuate the radar
open until June 1, 2010; separate discussions must be submitted for indi- signal and decrease its depth of penetration. High frequency an-
vidual papers. This paper is part of the Journal of Geotechnical and tennas, in the order of 1 to 2 GHz, produce high resolution that
Geoenvironmental Engineering, Vol. 136, No. 1, January 1, 2010. can detect small anomalies but can only penetrate up to approxi-
©ASCE, ISSN 1090-0241/2010/1-207–214/$25.00. mately 0.6 m 共2 ft兲 共Hunaidi and Giamou 1998兲. Low frequency

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2010 / 207

J. Geotech. Geoenviron. Eng. 2010.136:207-214.


antennas, on the order or 10 to 400 MHz, can penetrate tens to Table 1. Dielectric Constants of Common Roadway Materials
hundreds of feet, depending on soil conditions, but may not be Material Dielectric constant
able to locate small objects or anomalies. Although one can in-
Air l
crease exploration depth by increasing the transmitter power, un-
Granite 9
fortunately the power must increase exponentially with increased
Limestone 6
depth 共Daniels 1996; Hunaidi and Giamou 1998; Loken 2007兲. In
Sandstone 4
addition, the Federal Communication Commission 共FCC兲 regu-
Dry sand 4to 6
lates the level of radio emissions that can be generated. If the
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY-HYDERABAD on 06/18/13. Copyright ASCE. For personal use only; all rights reserved.

Wet sand 30
GPR transmitter signals become too large, they may interfere with
Dry clay 8
other devices such as TVs, radios, and cell phones. For applica-
Wet clay 33
tions on highway infrastructure, the new FCC regulation adopted
Asphalt 3to 6
in 2002 only allows GPR to be operated below 960 MHz or
Concrete 9to12
between 3.1 and 10.6 GHz, unless it has a grandfather exemption
Water 81
or its power is significantly reduced to an almost unusable level.
Metal 10,000
For highway pavements, where the depths of interest vary
from a few inches to 10 m 共⬃30 ft兲, two antennas may be re-
quired, one around 1 GHz and one around 200 to 400 MHz.
c ⫻ ⌬t1
Depending on antenna type, GPR systems are classified as air- Ha = 共1兲
coupled or ground-coupled systems. The TxDOT air-coupled sys- 2冑␧a
tems 共1 GHz兲 with the antennas 356 mm 共14 in.兲 above the where Ha=thickness of the asphalt or concrete surface layer;
surface allow for highway speed surveys. In contrast, a ground- c=constant 共speed of light in air= 0.30 m / ns兲; ␧a⫽dielectric con-
coupled system’s antenna fully contacts the ground, which limits stant of the pavement layer; and ⌬t1=time interval between peaks
the speed of the survey. The results presented in this paper are which represents the two-way travel time through the pavement
limited to 400 MHz ground-coupled penetrating radar 共GCPR兲 layer. Ranges of dielectric constants for typical pavement materi-
which 共with typical Tex. soils and degrees of saturation兲 scan als are given in Table 1.
approximately the top 4 m of the pavement structure.
The main purpose of this paper is to document the results of Case Study 1—Settlement Adjacent to a Culvert on
GPR surveys conducted by TxDOT and the Texas Transportation SH359
Institute in the last several years. Three field projects are pre-
sented in this study to demonstrate the application of GCPR for Laredo District maintenance personnel observed that riprap on the
eastbound shoulder of SH359 had collapsed 共refer to Fig. 1兲 and
void and anomaly detection. Boring and auguring were used to
verify the suspicious areas detected by the GCPR. Three road-
ways were showing localized problems, and the responsible
TxDOT districts were concerned about the safety of the roadway
structures and the extent of the problems. The combination of
GCPR and a few field core and auger samples provide a continu-
ous evaluation of subsurface conditions along the entire project.
The techniques demonstrated in this study are widely used within
TxDOT and are applicable to a wide range of roadway forensic
studies.

Basics of GPR

GPR is a nondestructive geophysical device used for subsurface


exploration. A GPR system operates by transmitting short pulses
of electromagnetic energy downward into the ground. These
pulses are reflected back to the antenna with amplitudes and ar-
rival times that are related to the dielectric constant of the mate-
rial layers. Across the layer interfaces, part of the energy is
reflected and part is absorbed, depending on the dielectric contrast
of the materials. The time delays and the amplitude of reflected
signals are used to evaluate the subsurface pavement conditions.
Objects or areas in the subsurface with different electrical prop-
erties will reflect the pulse differently and appear as anomalies.
Since the theoretical aspects of the GPR are covered in other
papers 共Sarrenketo and Scullion 2000; Hunaidi and Giamou 1998;
Loken 2007; Helmerich et al. 2008兲, only the thickness calcula- Fig. 1. Roadway condition near the studied area on SH359. 共a and b兲
tion 关Eq. 共1兲兴 is presented, as the thickness calculation will be Collapse of riprap on the eastbound lane; and 共c兲 distorted guardrail
discussed in the subsequent sections and patching on the westbound surface.

208 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2010

J. Geotech. Geoenviron. Eng. 2010.136:207-214.


Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY-HYDERABAD on 06/18/13. Copyright ASCE. For personal use only; all rights reserved.

Fig. 3. GCPR image and the related futures. 共a兲 GCPR image; 共b兲
Fig. 2. Roadway condition and GCPR testing. 共a兲 Roadway surface settlement with 100-mm overlay; and 共c兲 10 ft by 12 ft culvert with
condition on SH359; 共b and c兲 GCPR testing near weak spot and standing water.
water pipe.

lies can be observed in the eastbound lane. However there are


several anomalies in the westbound lane. So the focus of the
the westbound lane needed constant patching to level up the pave- discussion in this paper will be limited to the westbound lane.
ment surface due to settlement 共refer to Figs. 1 and 2兲. The settle- Although large volumes of GPR data were collected on this
ment on the westbound lane was evident from the distorted project, only areas with anomalies are presented. The GCPR im-
guardrail, patching, and cracking on the surface, as shown in Fig. ages shown in Fig. 3共a兲 clearly depict the settlement and patches
1. There was a water pipe leak near the eastbound culvert where
the city had worked to fix a leak about 1 year ago. There are fire
hydrants on the ditches near both the east and westbound lanes.
According to the utility design plan, there is a water pipe across
the pavement section with a depth of approximately 3.35 m from
the surface, as shown in Fig. 2.
The responsible district requested an investigation to deter- Culvert

mine the extent of the damage and the mechanism that is causing
the deterioration. An investigation was performed using GCPR,
among other tests. The study area is above a culvert that con-
stantly has water pooled inside it, as shown in Fig. 3共c兲. The
culvert box is 3.05 m by 3.66 m 共10 ft by 12 ft兲. Since Laredo is warehouse

in a very dry and hot area, pooling of water inside a culvert is


very unusual. An effort was made to determine if the pooling
water inside the culvert was natural runoff or water leaking from
the pipe. A chlorite test was performed. A chlorite content of 0.1 warehouse

ppm was found in the water under the westbound lane, and 0.19
to 0.3 was found under the eastbound lane. A chlorite content of
0.08 ppm is in excess of what occurs in nature. Thus, it was
determined that the water is from the leaking water pipe. During
the investigation, the city utility department was asked to check if
there is a leak in the water pipe. There is a water leak near a
warehouse that is about 600 ft from the culvert, as shown in Fig.
4共a兲. The green grass is observed from the source of the leak to
the culvert, as shown in Fig. 4共b兲. No other green grass was
observed in the area.
A 400 MHz GCPR antenna was used to survey and to map the Fig. 4. Studied area on SH359 with leaking water pipe near a ware-
subsurface condition, as shown in Fig. 2. No significant anoma- house

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2010 / 209

J. Geotech. Geoenviron. Eng. 2010.136:207-214.


A B C

10

Normal High Dielectric


Metal
Condition High Moisture 20

Time (ns)
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY-HYDERABAD on 06/18/13. Copyright ASCE. For personal use only; all rights reserved.

30

40

50

60

Fig. 5. GCPR testing and auguring verification 共A兲 GCPR testing


near an anomaly; 共B兲 GCPR images anomaly-metal; and 共C兲 augur-
ing verification
Fig. 6. Comparison of the Radar Signal among Locations A, B, and
C 关locations of A, B, and C can be found in Figs. 3共a兲 and 5共b兲兴

adjacent to the culvert. Fig. 3共a兲 shows the 100-mm settlement


near the culvert, as an accumulation of hot mix overlays have
been placed over the years to level up the pavement. metal is an abandoned metal object left during the construction.
Fig. 3共a兲 also shows that there are high dielectrics 共moisture兲 Fig. 6 shows the dielectric constant comparisons between the
on both sides of the culvert, especially from 20 to 70 m. In this anomaly 共metal兲, at 6 m 共normal area兲, and at 35 m 共near culvert
color scheme, a higher dielectric material is indicated by a red with settlement and high moisture兲. These three locations are ref-
band, indicating a positive voltage reflection. Boring was con- erenced in Figs. 3共a兲 and 5共b兲 as Locations A, B, and C, respec-
ducted at five locations to determine the chloride and moisture tively.
content in the soil. Chloride content tests were performed and Efforts were made to measure the moisture content of the soil
found that the chloride contents range from 500 to 800 共ppm兲. samples obtained from the weak spots. Fig. 7 shows how the
Note that TxDOT specification 关Texas Department of Transporta- moisture content increased with depth, as expected. Moisture con-
tion 共TxDOT兲 2004兴 requires that the chloride content for the tents from six other roadways 共U.S.83, SH44, FM2050, FM1472,
backfill material should be less than 100 ppm. Jayawickrama et and FM649兲 in the same county at depths from 1 m to 1.22 m 共3
al. 共2005兲 also indicated that for the backfill embankment material to 4 ft兲 were compared. The moisture content from the weak spot
chloride content should be less than 100 ppm to prevent corrosion at the depth from 1 m to 1.22 m 共3 to 4 ft兲 was about 5.3% higher
of structural elements. In general, high chlorite content, above than the average from six other roadways. Thus, the high moisture
naturally occurring in the native soils, was what proved that the content measurements from the weak spot confirmed the GCPR
water leaked from the water pipe and high chloride and moisture image and dielectric given in Figs. 3共a兲 and 6, respectively.
contents were attributed to creating a corrosive environment.
Not all boring was taken to the same depth because of obstruc-
tions 共rocks and abandoned metal兲 and a malfunction of the bor- Moisture Content (%)

ing equipment. The area with severe distress 共a pothole兲 is


0 10 20 30 40
strongly correlated to the GPR results and it is indicated as weak
spot in Figs. 2 and 3. There is significant bouncing of vehicles 0
In average
when they pass by the weak spot, indicating insufficient structural 5.3% higher than
500
support. As indicated in Figs. 2 and 3, the weak spot is adjacent to six other locations
the water pipe that runs across the west and eastbound lanes.
Depth (mm)

1000 in the same County


GCPR results also show high dielectrics and settlement near the
culvert and in the weak spot 关refer to Fig. 3共a兲兴. Boring was 1500
performed at the weak spot to collect sample up to 2.44 m 共8 ft兲
deep. 2000

Fig. 5共a兲 shows the GCPR testing near the culvert where there
is a significant anomaly 共marked with a square兲. Fig. 5共b兲 shows 2500

the GCPR images for the anomaly. Based on the shape of the
anomaly, it was judged as a metal. Verification was performed 3000

with boring and it was confirmed by strong resistance at a depth


approximately 0.6 m 关refer to Fig. 5共c兲兴. It was theorized that the Fig. 7. Moisture content with depth from the weak spot

210 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2010

J. Geotech. Geoenviron. Eng. 2010.136:207-214.


A B
90 Distance (m) 120
Rebar Pavement Surface

10 0.64
Top of Void
Bottom of CRCP

Depth (m)
Time (ns)
Bottom of Void
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY-HYDERABAD on 06/18/13. Copyright ASCE. For personal use only; all rights reserved.

rebar

40 2.56
void

Fig. 9. GCPR image indicating a 3.8 m3 void under IH40

Fig. 8. GCPR survey on IH40 and 3.82 m3 void under CRCP兲 this location. The anomaly starts directly under the CRCP. The
anomaly shows a significant drop in material dielectrics, and
based on the reflection pattern, it was suspected to be a void.
Therefore, the test results show that the problematic areas 共20 to Based on the GCPR image, the size of the suspected void was
70 m兲 not only have high chloride content, but also high moisture estimated to be significant.
content. The indication of the voids was an inversion of the GPR volt-
The GCPR was able to map the extent of the problematic area age, indicating a change in dielectric 共high to low兲. Normally, the
as highlighted in Fig. 3共a兲. This area generated many strong re- dielectric increases with layers at greater depth, yielding a posi-
flections and has smaller areas of high moisture content. It was tive reflection at the depth of the interface between layers. Fig. 10
theorized that moisture was wicking upward through the culvert’s shows a comparison between the suspected void and no-defect
pooled water that was from the leaking water pipe. The water in areas. The locations of the no-defect and suspected void are la-
the culvert and the moisture in the pavement was from the leaking
water pipe near the warehouse, not the pipe that runs through the
pavement. It is because the pipe that runs through the pavement
A B
had been examined with excavation and found no evidence of
leaking. The falloff of the rip rap in the east bound lane was
believed due to heavy rain that washed out the underlying mate-
rial.

Case Study 2—3.8 m3 Void under IH40


Amarillo District maintenance personnel reported that after rain- Top of Void
fall, fines were observed on the IH40 pavement surface near a
storm drainpipe. The typical cross section of the IH40 highway Normal
consists of 150 mm hot-mix asphalt overlay on top of 200 mm
continuously reinforced concrete 共CRCP兲 pavement. The location
where fines are pumping out of the pavement is near a storm
drainpipe buried approximately 3 m 共10 ft兲 under the CRCP. It
Bottom of
was suspected that there may be openings in the storm drainpipes
Void
that caused the washout. Amarillo District was concerned about
roadway safety, so they requested a GCPR survey to determine if
there are voids under the CRCP. Ten parallel GCPR tests were
performed surrounding the suspected area, as shown in Fig. 8共a兲.
The test length for each run was approximately 762 m 共2500 ft兲.
Fig. 8 also shows the verification hole dug to validate the GCPR
results.
Fig. 9 shows the GCPR results. The reflections from the rebar
are readily visible as a series of hyperbolas 共refer to Fig. 9兲. The
reason that a transverse bar appears as a hyperbola is because the
metal is imaged not only when the GCPR antenna is immediately
above, but also when the antenna is approaching and leaving the
target 共Daniels 1996兲. The apex of the hyperbola represents the
top of the steel rebar, and its measured depth correlates well with
what was given in the design plan.
A significant anomaly adjacent to the drainpipe was found, as Fig. 10. Comparisons of the radar signals between Locations A and
shown in Fig. 9. There is a clear change in the GCPR image at B 关locations of A and B can be found in Fig. 9兴

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2010 / 211

J. Geotech. Geoenviron. Eng. 2010.136:207-214.


75mm-100mm drop-off
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY-HYDERABAD on 06/18/13. Copyright ASCE. For personal use only; all rights reserved.

75mm-100mm drop-off

75mm-100mm drop-off

inlet

Cracks

Fig. 12. Conditions of U.S. 290 showing 75- to 100-mm drop-off,


cracks, and inlet on top of a mechanically stabilized earth 共MSE兲
Fig. 11. 3-m steel rod verifies existing void
retaining structure

beled in Fig. 9 as A and B, respectively. There are inversions and


significant negative reflections in the GPR image where the defect lized earth 共MSE兲 retaining structure in the westbound direction
is. A 25-mm diameter hole was drilled through the CRCP at the of U.S. 290 in southwest Austin, as shown in Fig. 12. In addition,
suspicious location, and a void was found. A 3-m steel rod was there are longitudinal cracks on the pavement surface 共refer to
used to estimate the extent 共depth兲 of the void, as shown in Fig. Fig. 12兲. The lane drop-off or fault measured up to 100 mm and
11. The 3-m rod reached something solid at the depth of approxi- the joint separated up to 25 mm. Maintenance personnel applied a
mately 2.4 m. This means there was over 1.8 m of thickness narrow asphalt concrete patch in the lane drop-off area. The lane
immediately under the CRCP. Twenty-five mm diameter holes drop-off and longitudinal cracks near the MSE wall prompted
were drilled in other locations, and no voids could be found. district personnel to request an investigation to assess the safety
A 0.6 m by 0.6 m hole was cut and removed at the void of the structure and to determine if there are significant voids
location, as shown in Fig. 8共b兲. The main purpose for the 0.6 m under the CRCP.
by 0.6 m hole was to examine the void condition and fill the void The locations of cracks and lane drop off are near an inlet
with cement grout. Fig. 8共d兲 shows the schematic of the void where has storm drainpipe buried approximately 1.8 m 共6 ft兲
under the CRCP. The CRCP had bridged over the 1.8-m deep under the CRCP. The main concern was the lane drop-off and
void. 3.8 m3 of cement grout was used to fill the void. If the void longitudinal cracks near the inlet and storm drainpipe may asso-
under IH40 was not found and was allowed to increase in size, it ciate with potential voids in the MSE structure, making the MSE
may have later collapsed. retaining structure unsafe. GCPR data was taken in the longitudi-
An effort was made to estimate the size of the void through nal directions. Additional GCPR tests were conducted in the
GPR images and physical measurements. In review of Fig. 9, the transverse direction at locations are suspicious. The GCPR data
vertical axes are the time of travel in nanoseconds and depth indicated very shallow voids along the faulted longitudinal joint
estimation 共in meters兲. The start of the pavement surface
共⬃2 ns兲 has been labeled in Fig. 9 for clarity. The bottom of the
CRCP slab 共or top of the void兲 occurs at approximately 12 ns and rebar
appears to terminate at 24 ns. The selection of the bottom of the
void is subject to interpretation, as there are multiple reflections to
choose from. However, the strong white reflection at 24 ns was
selected. Therefore, the time of travel in the void was 12.0 ns.
Using Eq. 共1兲 with c = 0.3 m / ns, ⌬t1 = 12.0 ns, and the dielectric
of the air= 1, would estimate a void depth of 1.88 m 共=0.3
⫻ 12/ 2兲. The depth of a void is sometime difficult to estimate,
because of the lack of a clear reflection from the bottom. How-
ever, the void estimation is sufficient to indicate “deep” or “shal- 4.6*3.7*1.8m void
low” which is adequate for most inspections. This case study
demonstrates the successful application of GCPR to locate voids
that may lead to a serious safety hazard if not identified in time.

Case Study 3— Lane Drop-Off on U.S. 290


The Austin District personnel have been concerned with the lane Fig. 13. GCPR image showing anomaly and distortion under CRCP
drop-off 共varying from 75 to 100 mm兲 in a mechanically stabi- pavement with 4.6 m ⫻ 3.7 m ⫻ 1.8 m 共30.6 m3 in volume兲 void

212 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2010

J. Geotech. Geoenviron. Eng. 2010.136:207-214.


GCPR signals and causes the dielectric properties to increase sub-
3
stantially with increased water intrusion. The extent of the sub-
4.6*3.7*1.8m (30.6 m ) void
surface problem on SH359 is readily visible in the GCPR images,
which show locations with settlement and high moisture content.
The 100 mm 共4 in.兲 settlement near the culvert was confirmed
Disjointed pipe
through core samples. Through auguring and lab testing, the
source of the high moisture was determined to be from a leaking
pipe, and not from natural rainwater, as determined by the high
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY-HYDERABAD on 06/18/13. Copyright ASCE. For personal use only; all rights reserved.

chloride and chlorite contents. In general, the high chlorite con-


tent, above naturally occurring in the native soils, was what
proved the water leaked from the water pipe and high chloride
and moisture contents were attributed to creating a corrosive en-
vironment.
In addition, GCPR images can outline the condition with re-
spect to depth in the problematic area. Thus, the engineer can
utilize the information with confidence to develop remedial strat-
Lon
gitu
75~
100
Dro
mm egies that consider the approximate dimensions. As shown in this
din p-o
al C
rac
ks
ff
paper, the radar signals generated by air-filled voids beneath
pavement sections are inverted and relatively easy to detect.
4.6
Void

*3.7
Are
a These anomalies were clearly identified in GCPR images. A
1.8-m deep void 共3.8 m3 in volume兲 on IH40 and a 1.8 m
*1.8
m

⫻ 4.6 m ⫻ 3.7 m 共30.6 m3 in volume兲 void on U.S. 290 was


MS
successfully identified below the CRCP pavement and verified by
E W
a ll
Dis
join
field boring and coring. If the voids had not been identified in
Inle ted
t
Dra
in P
ipe
time, they would have increased in size, which could have lead to
a catastrophic failure.

Fig. 14. Confirmed 30.6 m3 共4.6⫻ 3.7⫻ 1.8 m兲 void under CRCP
with disjointed pipe
Acknowledgments

共less than 25 mm兲 and this was confirmed with drilling through The support and assistance from Jenny Li, John Bilyeu, Hector
the concrete. The voids were not considered severe. Cantu, Jo Ann Garcia, Russell Luther, Joe Chappell, Miguel Arel-
A transverse storm drain was placed underneath the CRCP lano, and Mike Taylor of Texas Department of Transportation and
near the end of the lane drop-off, so GCPR data was collected in Tom Scullion and Lee Gustavus of Texas Transportation Institute
this area. The GCPR data indicated a significant anomaly over the is much appreciated.
transverse storm drain, as shown in Fig. 13. A core were taken,
revealing a void approximately 1.8 m 共6 ft兲 deep, 4.6 m 共15 ft兲
long, and 3.7 m 共12 ft兲 wide as shown in Fig. 14. Thus, the References
estimated void is approximately 30.6 m3. As can be seen in Fig.
14, the transverse storm drain had separated. It is thought that this Chamberlain, A. T., Sellers, W., Proctor, C., and Coard, R. 共2000兲. “Cave
separation caused water to erode the area around the drain. The detection in limestone using ground penetrating radar.” J. Archaeol.
resulting moisture intrusion in this area and moisture flow from Sci., 27, 957–964.
the grassy median through the embankment into the void may Chen, D.-H., Nazarian, S., and Bilyeu, J. 共2007兲. “Failure analysis of a
have resulted in the embankment settling where the longitudinal bridge embankment with cracked approach slabs and leaking sand.” J.
joint faulted. Perform. Constr. Facil., 21共5兲, 375–381.
Fortunately, the disjointed storm pipe has been identified in Chen, D.-H., and Scullion, T. 共2008兲. “Detecting subsurface voids using
time. Otherwise, the 30.6 m3 void could have grown and col- ground coupled penetrating radar.” Geotech. Test. J., 31共3兲, 217–224.
lapsed resulting in human casualties and property losses. Borings Daniels, D. J. 共1996兲. Subsurface-penetrating radar, The Institute of
are planned to be taken in the area of the longitudinal joint fault- Electrical Engineers, Los Alamitos, Calif., 1–62.
ing and separation to see if there are other problems with the Federal Railroad Administration 共FRA兲. 共2005兲. “Non-destructive evalu-
underlying embankment material such as deeper voids that cannot ation of railway track using ground penetrating radar.” RR05-06,
具http://www.fra.dot.gov/downloads/Research/rr0506.pdf典 共Nov. 7,
be detected with GCPR data. District personnel are planning to let
2009兲.
a project to repair this area.
Hauser, E. C., and Howell, M. J. 共2007兲. “Ground penetrating radar sur-
vey to evaluate roadway collapse in North Ohio.” 具http://www.
wright.edu/geology/People/faculty/ehauser/papers/p087ech.pdf典 共July
Conclusions 2007兲.
Helmerich, R., Algernon, D., Niederleithinger, E., Streicher, D., and
This study has successfully demonstrated that the GCPR is able to Wiggenhauser, H. 共2008兲. “Bridge inspection and condition assess-
identify anomalies and void locations. GCPR was used in three ment in Europe.” Proc., Transportation Research Board Annual Meet-
projects to identify anomalies under roadway pavements. The ing, Washington, D.C.
presence of moisture under pavement has a significant impact on Hunaidi, O., and Giamou, P. 共1998兲. “Ground penetrating radar for detec-

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2010 / 213

J. Geotech. Geoenviron. Eng. 2010.136:207-214.


tion of leaks in buried plastic water distribution pipe.” Proc., 7th Int. Orleans. 具https://transportation.wes.army.mil/triservice/navfac/reports/
Conf. on Ground-Penetrating Radar, Lawrence, Kan., 具http:// WhitehousePVD06.pdf典 共June 11, 2009兲.
www.watersanitationhygiene.org/References/EH_KEY_ Loken, M. C. 共2007兲. “Use of ground penetrating radar to evaluate Min-
REFERENCES/WATER/Piped%20Water%20Schemes/Water nesota roads.” Research Rep. No. MN/RC-2007-01, Minnesota Dept.
%20Leak%20Detection/Ground%20Penetrating%20Radar%20for of Transportation, St. Paul, Minn.
%20Leak%20Detection%20共ICGPR兲.pdf典 共Nov. 7, 2009兲. SAARENKETO, T., and Scullion, T. 共2000兲. “Road evaluation with
Jayawickrama, P., Jackson, A., Clifford, F., Morse, A., and Rajagopalan, ground penetrating radar.” J. Appl. Geophys., 43, 119–138.
S. 共2005兲. “Use of alternative water sources.” TxDOT Research
Project Rep. No. 0-4412-1, Texas Dept. of Transportation, Austin, Texas Department of Transportation 共TxDOT兲. 共2004兲. Standard specifi-
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY-HYDERABAD on 06/18/13. Copyright ASCE. For personal use only; all rights reserved.

Tex. cations for construction and maintenance of highways, streets, and


Lesto, J., and Hatcher, J. 共2002兲. “Airfield void detection.” Nas, New bridge, Dept. of Transportation, Tex.

214 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2010

J. Geotech. Geoenviron. Eng. 2010.136:207-214.

You might also like