International Symposium (NDTCE 2003) NonDestructive Testing in Civil Engineering 2003  
Start > Contributions >Lectures > Radar 1: 
A HIGH FREQUENCY PENETRATING RADAR FOR MASONRY INVESTIGATIONDaniele Mecatti, Massimiliano Pieraccini, Guido Luzi, Linhsia Noferini, and Carlo AtzeniDepartment of Electronics and Telecommunications, University of Florence via Santa Marta 3, 50139 Firenze, Italy dmecatti@ieee.org AbstractIn this paper a highfrequency largebandwidth syntheticaperture penetrating radar for inspecting masonry structures is described. The system operated in a noncontact manner by mechanically moving the transmitting and receiving antennas in order to realize a synthetic aperture. The radar signal was a continuous wave step frequency (CWSF) waveform,scanning a 7 GHz bandwidth at 9.5 GHz center frequency, thus providing high resolution images of the investigated structures. Laboratory tests in order to assess the performances of the system are reported. Furthermore some preliminary results of a case study on a wall of historical interest ("Salone dei Cinquecento" Palazzo Vecchio  Firenze) are presented.IntroductionIntrawall radar inspection offers a unique noninvasive means for the diagnosis of buildings where the assessment of the structural safety and durability needs information on the internal structure and on the localization and size of reinforcement bars, voids and defects [1].Penetrating radar is being used increasingly for the appraisal of structural safety and durability of historical buildings [2] and large structures such as dams [3] and bridges [4]. In general, low frequency systems are more penetrating but they can perform limited bandwidths, hence a low spatial resolution. High frequency systems can provide larger bandwidths but their penetration depth is small. This fact discloses as a critical tradeoff in ground penetration applications, that usually require penetration depths up to a few meters in strong attenuating media. Masonry investigation requirements are usually different under relevant respects: the required penetration depth rarely exceeds several tens of centimeters,and dry masonry is light attenuating. In effect, a number of works both theoretical [5] and experimental [6, 7] stated that attenuation of all ovendry building materials approaches zero,as electromagnetic waves are only absorbed when water is present in the wall components. In this paper a highfrequency largebandwidth syntheticaperture penetrating radar is proposed. Performances were experimentally evaluated and an experimental test was designed and realized to asses intrawall imaging capability on dry masonry structures. The radar systemA SAR system operating at 9.5 GHz center frequency, with a bandwidth of 7 GHz, was designed and built. The transmitting and receiving horn antennas were mechanically moved along a linear guide, in order to synthesize an aperture up to 2 meters. The radiated signal was a continuous wave step frequency (CWSF) waveform, that sampled the required bandwidth in programmable frequency steps [8, 9]. The radar system was based on the homodyne architecture sketched in Figure 1.
An external synthesizer (HP 8672A) produced continuous waves at discrete frequency values, with frequency steps of 5 MHz. It fed the transmitting antenna and provided a reference signal for the mixer that detected the inphase component of the received signal. Transmitting and receiving antennas were two exponential horns with 15° of half power beam width at 10 GHz of frequency. The demodulated signal was amplified and digitized by the acquisition board,that provided the digital samples to a personal computer. Data processingAs the signal propagated through different media, the effect of dielectric property changes had to be considered in data processing. Both attenuation and reduction of the propagation speed of the electromagnetic waves in the investigated wall had to be taken into account.Their effect on imagery was a defocusing caused by both an exponential decay not compensated and a defocusing due to the refraction at the surface of separation between air and wall. In order to take into account attenuation and electromagnetic speed variation in a wall the following algorithm was developed and tested. Let consider the electromagnetic waves propagation through ray tracing approach: starting from each point P_{n}, a number N_{r} of rays were traced inside an angular aperture b with respect to the direction normal to the vertical plane. This aperture b is smaller than the antenna field of view, as reduced by the Snell law. So, by assuming the half power beam width b_{A} as angular limit of the free space antenna field of view, the angular aperture respects the following equation:
with e_{r} the relative permittivity of the investigated medium. The traced rays intercept the vertical plane scanned by the mechanical positioning system, so N_{r} positions of the antennas that fit these interception points can be identified. The value of the complex radar image at point P_{n} is obtained by adding only these signal contributions, taking into account their distances calculated along the corresponding ray between the point P_{n} and the wall surface and between the wall surface and the position of the antenna. In order to consider the attenuation of the investigated medium, the attenuation coefficient is written as [10]:
where, c is the speed of light, f is the microwave frequency, tan d is the loss tangent and is approximately constant with frequency for a great part of ground and masonry materials [11].
The radar image at point P_{n} (I_{n}) can be obtained by adding all signal contributions, taking into account their amplitude and phase history governed by attenuation coefficient, path, and dielectric permittivity [9]:
where r^{(l)} _{n,k}is the distance between the point A_{k}, kth position of the antennas, and the
incidence point S_{k} on the wall surface
Laboratory testsExperimental tests were carried out by means of a radar system developed in our laboratory. A testwall was built in the laboratory (Figure 3). It consisted of a brick and lime wall at a controlled distance from a masonry wall.Figure 4 shows a typical inverse Fourier transform of a measured frequency sweep. The first peak (A) is due to the first interface airwall, the next two peaks (B,C) identify a hollow space of 10 cm between the two walls.
In order to test the capability of this radar technique a series of measurements by varying the thickness of the hollow space between the two walls was carried out.Figure 5 plots the time domain measured thickness by varying at 1 cm steps the thickness of the hollow space between the two walls.
In order to apply the focusing method, it was necessary to evaluate the dielectric characteristics of building materials (e_{r}). There are a number of available techniques of different accuracy and effectiveness [6, 7]. In this work a standard method for masonry of known thickness (d) accessible from both sides is used [6]. The real part was directly obtained from the inverse Fourier transform of a frequency sweep. As seen above, the peak A is due to the first interface airwall, the peak B identifies the end of the first wall. Then, the square root of e_{r} can be calculated by the time of flight (t) between the two faces of the wall by using the following relationship:
In order to estimate the attenuation coefficient (a) of the wall,the transmitting and receiving antennas were positioned one in front of the other at the opposite sides of the wall, in contact with it. From the inverse Fourier transform of a frequency sweep, the amplitude A_{tw} of the signal transmitted through the wall was obtained. The same measurement was carried out positioning the two antennas in free space one in front of the other at a distance equal to the wall thickness. Let A_{fs} be the measured signal amplitude in freespace arrangement. The attenuation coefficient can be evaluated as:
This method [6] gives a rough estimation of the dielectric properties of the wall but as the qualitative behavior of the timefrequency pattern is not dramatically affected by the imaginary part of the refractive index, an accurate evaluation of this one is not essential in this study. A case study: the walls of "Salone dei Cinquecento" in Palazzo Vecchio

Fig 6: Picture of the radar instrumentation at "Palazzo Vecchio" (Firenze).  Fig 7: Radar image of a Salone wall in "Palazzo Vecchio" (Firenze). 
Figure 7 shows a radar image acquired into the eastern wall of the Salone. It's a section of the wall in a vertical plane passing through the radar: the xaxis is the direction perpendicular to the plaster surface (the origin is the plaster), the yaxis is the vertical direction. The image was focused using the value 4.0 as dielectric permittivity (e_{r}), that is the value we obtained by a preliminary permittivity measurement on the same wall, as described in the previous chapter.As the back face of the wall is clearly visible in the radar image (0.7 m), the whole thickness of the wall is imaged. The most interesting feature is the interface 15 cm behind the plaster surface. It points out a discontinuity between two building layers. It confirms the available ancient documentation that affirm that a wall was subsequently built leaning against the preexistent wall. The interest about this discontinuity is related to the fact that if the lost Leonardo's fresco was not destroyed, it can be in a narrow cavity inside any wall of the "Salone dei Cinquecento".
