Detection of Water Leaks in Supply Pipes With Active Microwave GPR Sensors and Electromagnetic Modelling
Detection of Water Leaks in Supply Pipes With Active Microwave GPR Sensors and Electromagnetic Modelling
Detection of Water Leaks in Supply Pipes With Active Microwave GPR Sensors and Electromagnetic Modelling
0
}
Output
Ground Layer 2, {
r2,
0
}
Point
source
URSI EMTS 2004 111
- Receiver Sensitivity: -100 dBm
- Processing: Analog-to-Digital Conversion at 75Hz sampling rate, 4096-point FFT in a Pentium
Processor
The reason of combining FM-CW and CW is to achieve both detection of spatial inhomogeneities inside the
ground medium because of the presence of water caused by the leak and also detection of small Doppler shift
because of the flow of water which is of the order of few Hz. In addition to the two radar modalities, in the
framework of the LEAKING project a passive microwave radiometer is also used to detect independently
the presence of water leaks in supply pipes. The three independent modalities being supplementary are ex-
pected to provide better detection performance in diverse environmental conditions (high detection probabil-
ity and lower false alarm rates).
MEASUREMENT RESULTS
Testing measurements have been carried out in experimental sites and several features of leak detection were
evaluated. The aim of the field tests was to locate water leaks in underlying supply pipes by detecting the
slow leaking water movement and the voids caused by water concentrations. It has been shown that the use
of FM-CW in conjunction with CW radars provides promising results. The following figures provide several
measurements taken from both sensors after A/D Conversion and FFT processing with the aid of the Na-
tional Instruments LabView program.
Figure 2: CW radar measurements on leak point: 2V p-p signal (sec V) and spectrum (Hz dB)
Figure 3: CW radar measurements away from leak: 0,1V p-p signal (sec V) and spectrum (Hz dB)
The CW measurements prove that this active sensor can detect even the very slow movement of water cre-
ated by the leaks. The effect of this movement on the returned signal (Figure 2) is evident, compared to the
weak and noisy signal of Figure 3.
The FM/CW measurements (Figure 4) exhibit the capability to spot inhomogeneities in the ground medium.
The frequency domain can be mapped to the spatial domain linearly, thus providing information about the
structural defects and alterations caused by the presence of ample amounts of leaking water. Indeed, the in-
formation obtained by the linear FM on the vertical axis can be used as a test bed for obtaining results from
the EM modelling algorithm presented in the following section.
112 URSI EMTS 2004
Figure 4: FM-CW radar FFT magnitude (Hz dB): away from leak (left) and on the leak (right)
ELECTROMAGNETIC MODELLING
The two layered ground medium shown in Figure 1 was analysed by applying the Sommerfeld technique
[1,2]. A dipole source inside the earth medium is assumed and then the field in the air medium is determined.
The obtained result has the form of Sommerfeld integrals which cannot be computed analytically [1,4]. Tak-
ing into account the fact that near-field regime is valid, the electric field induced by a dipole source inside
the earth is computed by using a Taylor expansion of the integral terms [3]. By considering all possible di-
pole orientations, a Greens function of the structure is obtained. The result can be written in the following
convenient form:
( ) ( )
__
= =
' ' ' = ' ' '
I
i
J
j
ij
j i
z y x z y x G
1 1
, , o
(1)
where (x, y) is the source point inside the earth, while the observation point is on the origin of the coordi-
nate system on the surface of the earth. The
nm
coefficients are computed numerically and are readily avail-
able if they are computed once. In order to develop an imaging algorithm the inhomogeneities caused by wa-
ter leak are computed by minimizing the error function:
( ) ( )
2
1 1
, , ,
__
= =
_ A ' ' ' =
N
n
M
m
ij
ij l j i m n
x z y x G y x V c (2)
where V(x
n
, y
m
) is the measured voltage at the observation point x
n
, y
m
, and x
ij
is the dielectric permittivity
function described by a cubical basis function inside the earth.
CONCLUSIONS
The present work analyses a novel approach in leak detection techniques, based upon ground penetrating
radar technology. The method proposed here combines a microwave subsurface imaging technique with an
underground movement detector, in order to obtain a more accurate and robust leak detection system. The
measurements have shown the efficiency of the CW radar in detecting the very small Doppler shift of the
moving water, while the FM/CW sensor provides a suitable device for applying the subsurface imaging algo-
rithm shown here. Future work plans include the development of an expert subsystem that translates radar
measurements to simple indications about the probability of water leak existence, a task requiring extensive
field tests in diverse environmental conditions.
REFERENCES
[1] Alfredo Baos, Jr, Dipole Radiation in the Presence of a Conducting Half-Space, Pergamon Press,
New York, 1966.
[2] P. Yl-Oijala, M. Taskinen, Efficient Use of Closed Form Greens Functions for the Electromagnetic
Scattering by 3D Buried Objects, Proceed. 2001 IEEE AP-S Symposium, Vol. 4, 2001, pp. 834-837.
[3] Krzysztof A. Michalski, Extrapolation Methods for Sommerfeld Integral Tails, IEEE Transactions on
Antennas and Propagation, Vol. 46, No. 10, October 1998, pp. 1405-1417.
[4] N. Geng, L. Carin, Wide-Band Electromagnetic Scattering from a Dielectric BOR Buried in a Layered
Lossy Dispersive Medium, IEEE Transactions on Antennas & Propagation, Vol. 47, No. 4, pp. 610-
619, April 1999.
URSI EMTS 2004 113