Bundesanstalt für Materialforschung und -prüfung

International Symposium (NDT-CE 2003)

Non-Destructive Testing in Civil Engineering 2003
Start > Contributions >Lectures > Radar 1: Print

Development of an Utility Finding Impulse Radar

Thomas Kind, Ingo Krause, Christiane Maierhofer
Federal Institute for Materials Research (BAM), Berlin, Germany


Within the EC-project Smart-RAD: Development of an Innovative Ground Penetrating Radar for Recognition and Identification of Subsurface Buried Objects an impulse radar for utility finding is developed by five European project partners. The objectives and first results of the project will be presented in this paper. The main focus of BAM is on the antenna development which will be briefly discussed. The project is financially supported by the EU in the fifth framework program Competitive and Sustainable Growth (GROWTH) for three years and will last till May 2004.


The need of non-destructive detection of buried objects in building materials has continuously increased in the last twenty years within the sector of civil engineering and construction industries. Therefore impulse radar [i] has been applied for non- destructive damage assessment and also in recent years to quality assessment. Typical applications are the detection of voids and reinforcement in concrete, the detection of detachments at pavement structures [ii],[iii] and subsoil investigations.

The Smart-RAD project consortium has focused its efforts to the area of utility finding in building ground and has identified the main limitations of existing systems as:

  • Unreliable detection confidence pending on ground and underground conditions, especially of soil type and moisture content
  • Coarse accuracy in the localization of detected buried objects
  • Unavailability of an efficient and cost-effective survey technique
  • Difficult on-field radar data interpretation, today still only restrained to expert operators

A main point which motivates the Smart-RAD project is that the cost for damages on risky lines like gas or power lines are about 50 million EURO per year only in Europe.

Up to now, the market of impulse radar also known as ground penetration radar (GPR) or surface penetration radar is limited due to the system and service cost and user acceptance. Another point for the market limitation is that GPR is only used for buried object detection up to a depth of 1m. But the need of utility companies is to detect pipes and tubes of risky lines (gas, power and telecommunication) in depth up to 5 m. Before utility companies are doing excavation work they need a subsurface map showing the presence and accurate location of the lines. So beside the hardware development to improve the GPR performance the project consortium will also develop algorithm to interpret the conventional radargramms (B or C scan) of existing system into the desired utility map, so that less training of personal is required.

Project Objectives

The main objective of the project is the development of GPR technology for precise and fast no-dig areas tracing, which would allow timely decisions to be taken on where and how deep to dig, reducing sensibly the number of borings and excavation to be performed, and the high cost associated. The scientific and technological research objectives of the project related to the main objective can be summarized in four groups as follows:

  1. Antenna Design and Development
  2. Transmitter and Receiver Hardware Design
  3. Survey Control and Antenna Positioning Technique Design
  4. Software Design and Development

In the transmitter and receiver hardware design the activities will focus on the design of a high performance GaAs pulse generator, a wide band GaAs pre-amplifier and a ultra fast sampler unit with jitter better than 40ps which will lead to an improved precision in the determination of buried object depth. The centre frequency of the radar system has a major impact on the depth of penetration of the radar system. For low frequency the attenuation in the ground is lower than for high frequency. So the centre frequency for the Smart-RAD system has been chosen to 500MHz, which suits the specifications met in the Smart-RAD project.

The survey control will be designed for fast dragging of the GPR antenna over the surface which will require a design for a fast data transfer between sampler unit and processing unit. For the antenna positioning technique a wire displacement transducer, a differential global positioning system and a video based system will be tested.

The software design and development has been divided in a radar data interpretation (RDI) module and a neural imaging processing (NIP) module. The RDI module will process the raw data and display the results in a clear form to the user. The general requirement of the RDI is to make the processing simpler and quicker for the user in the way that less training is required. The NIP module will support the processing of the radar data by using the raw data and the RDI results. It will provide the user with information about the presence, geometry depth and dielectric properties of the buried objects in real time.

Antenna Development

The main problems of GPR antennas are the weak reflection signal and the limited resolution capability. The week reflection signal is caused by radiation power losses during the propagation of the electromagnetic pulse wave from transmitting antenna towards the reflector (e. g. buried objects, abrupt changed ground layers) and back to the receiving antenna. Sources for the power losses of the electromagnetic waves are: mismatched feeding line and antenna impedance, internal power losses of the antenna especially for resistive loaded antennas, large radiation pattern of the transmitting and receiving antenna and thus a small antenna gain, unwanted reflection at the surface, ground medium absorption and most cases the buried object is not a perfect reflector. The last two sources are inherent of the application and can not be changed. All other power loss sources can be optimized for the application within limits.

The limited resolution capability of the antenna is related to the limited bandwidth of the antenna, the chosen center frequency of the antenna and the not tuned frequency band of the antenna to the frequency spectrum of the electrical signal from the impulse generator. The chosen center frequency influences also the power losses due the frequency dependency of the ground medium absorption.

The ongoing antenna development is focused on the resolution capability first, by incorporating conductive foam absorbers [iv] in the antenna housing. In a second step this foam absorbers will be optimized for maximum penetration depth.

For the development of antenna design, a analytical simulation tool for linear wire dipoles has been developed [v] to calculate the electro-magnetic field inside a dielectric space. The objective of the simulation tool is to study the influence of the antenna geometry on the shape of the reflected impulse.

In [vi] is shown that the transmission of a thin wire dipole as shown in fig. 1 can be calculated by a closed formulation (s. eq. 1) derived from the transmission line equation. The electrical field strength for any point can be calculated by the contribution of the current flow at the antenna feed and the antenna tips .

Fig 1: Linear wire dipole with spherical coordinate system of the radiated electromagnetic field for the antenna tips and feed; denotes the unit vector in each of the three coordinate system.


Antenna Prototype and Test
The transmitting and receiving antenna construction are similar. Each antenna consists of a metal housing soldered by tinned sheet metal where one of the two larger sides is open for the radiation. The antenna itself is made by brass and each bow tie triangle has equal edges of 13 cm. The brass is mounted on a cardboard by a glue strip and the cardboard itself fixed on the bottom side of the antenna housing by a glue strip again. Direct to the brass triangle tip a ferrite core transformer balun vii is soldered, which is mounted on a small PCB for stability reasons. The ferrite core transformer has an impedance matching of 1:1 and corresponds to an autotransformer. The balun is connected to the antenna female SMA connector by a 2 cm rigid line. The female SMA connector is centered at the antenna backside (s. fig. 2). The size of the antenna housing is 28 cm x 17 cm x 7.5 cm.

Fig 2:
Antenna prototypes; left: antenna Type A without absorber foam; right: antenna with 7.5 cm absorber foam above the antenna brass sheet.

Antenna tests have been done with and without foam absorbers. The foam absorbers have a specified conductivity and are commonly applied for reducing reflections in radar applications. These 7.5 cm thick absorbers were direct mounted to the bottom side of the antenna (s. fig. 2 right). The absorber is a composite of three absorber foams layers.with different conductivities and the foams are arranged in the ways that by increasing distance to the antenna sheet the conductivity of the foam layers are also increasing, thus the absorber is acting as a slough.

In the following, the antenna without absorber foam is denoted as Type A antenna, while the antenna with 7.5 cm absorber foam is Type B.

Fig 3: Setup of the A-Scan measurement with metal sheet reflector;. d and h denotes the antenna separation and the distance of the metal sheet reflector above the antennas.

The resolution capability of the antenna has been characterized by the recording of A-scans in air with the antenna opposite to a metal sheet reflector. Different A-Scans were collected for different distances (h) of the metal sheet reflector (s. fig. 3). Type A and B antenna were used for this recording. The height of the metal sheet reflector was altered in three steps: 50, 100 and 150cm and antenna separation was chosen as 23.5cm. The collected A-scans for different metal sheet reflector height were stringed together in a B-Scan. The antenna surface clutter was removed by subtraction of a reference A-Scan, which was taken at the beginning of each measurement. This approach can be used as the antenna surface clutter is independent from the metal sheet reflector.

Fig 4: B-Scans for the setup in fig. 3 (left: type A antenna; right: type B antenna). Metal sheet distance was altered in sequence 100cm (I), 50cm (II), 150cm (III).

The B-Scans with background removal show clearly the echo of the metal sheet reflector. The appearance of the reflection at different times is correlated to the height of the reflector. The duration of the reflections due to the ringing of the impulse is obviously shorter for the type B antenna than for the type A antenna. The different ringing of the reflection (h=150cm) is shown in fig. 5. The ringing of the type A antenna is caused by interference of reflection inside the antenna housing and direct transmitted pulse wave. For the type B antenna the reflection inside the antenna housing is much more damped due to the absorbers. But it is shown that also the direct transmitted pulse wave is damped due to the absorber. This can be optimized by choosing a appropriated graded conductivity profile along the foam thickness with increasing conductivity.

Fig 5: A-Scans for metal sheet distance of 150cm (left: type A; right type B).

For evaluation of the penetration depth the antenna prototype with foam absorbers (type B) has been applied to a utility test site. The test side consists of six metal and PVC pipes arranged in different depth and horizontal position. The pipes are buried in sand with gravel underneath. Fig 6 shows the resulting B-scan with reflection hyperbola of all six pipes and sand/gravel interface at a depth of about 1.8m.

Fig 6: B-Scan of a utility test site recorded by type B antenna; P1-P6 denotes the pipes of the test site.


The evaluation of the present commercial used GPR systems for utility finding shows a considerable potential for further development. Beside further technical development of GPR system components the Smart-RAD project objectives aims to develop a more user friendly system using by setting up a adequate Man-to-Machine Interface (MMI).

Tests with a antenna prototype show that the incorporation of conductive foam in GPR antenna housing results in an increase of the resolution capability. Further antenna development will be focused on optimization of the conductivity profile along the foam thickness to minimize the impact of the absorber foam to the radiated signal intensity.

In the remaining project period the very good cooperation with the partners will be continued to finalize the radar prototype and to perform intensive field test of the protoype on different test sites.

The project consortium consists of seven partners: D'Appolonia S.p.A. (I), University of Pavia - Department of Electronics (I), Mala Geoscience AB (S), Federal Institute for Material Research and Testing - BAM (D), Building Research Establishment - BRE (UK), Structural Testing Service - STS LTD (UK) and G Impuls Praha (CZ). D'Appolonia S.p.A. is the project coordinater.

Project number: GRD1-2000-25155


  1. Daniels, D.J., Surface Penetrating Radar, London, U.K., The Institute of Electrical Engineers, 1997
  2. Maierhofer, Ch., Kind, Th., Application of Impulse Radar for Non-Destructive Investigation of Concrete Structures, Ninth International Conference on Ground Penetrating Radar, April 29 - May 2, 2002, Santa Barbara, California USA
  3. Merkblatt B10, Merkblatt über das Radarverfahren zur zerstörungsfreien Prüfung im Bauwesen, Deutsche Gesellschaft für zerstörungsfreie Prüfung e.V. (DGZfP), Nov. 2001
  4. Ugur O. , Gürel L. , "Modelling of Ground-Penetrating-Radar Antennas With Shields and Simulated Absorbers", IEEE Trasnactions on Antennas and Propagation, Vol. 49, no. 11, p. 1560-1567, Nov. 2001
  5. Göttel R., Mönich G., Maierhofer Ch., "Antenna design and simulation for the application of radar to concrete structures", 1. International Radar Symposium, Munich, 15.09.- 17.09.1998
  6. Mönich G , "Impulse Radiation by Wire Bends", International Conference on Antennas and Propagation, Warwick, 1989
  7. Balanis, C. A., 1997, Antenna theory - Design and Analysis, New York, John Wiley & Sons Inc.
STARTPublisher: DGfZPPrograming: NDT.net