For a long time, microwaves have been considered as a possible sensing agent for nondestructive testing/evaluation purposes. This trend has still been reinforced these last years with the advent of new microwave penetrable materials, such as composites. Inspection of materials via a mechanically scanned probe has proven to offer a convenient, but time consuming, way to measure local reflexion or transmission coefficients and, hence, to evaluate defects, faults, etc... High speed measurements are now possible by using arrays of fixed probes, resulting in attractive imaging equipments. Indeed, the availability of amplitude/phase data allows us to consider different processing techniques, the complexity of which can be selected according to the required performances in terms of contrast, spatial and time resolutions. This paper reviews some of the most promising approaches, such as non-linear inverse scattering techniques and neural networks. Prospective considerations are devoted to the future of such sophisticated microwave sensing techniques.

The potential of microwaves for defect and flaw detection in glass reinforced polymer composites is investigated. Specially fabricated thick glass reinforced polymer composite samples with several types of embedded defects are used for these measurements. An open- ended rectangular waveguide is utilized as a sensor. Microwave images were created by scanning the samples. These images demonstrate the ability of microwaves for detecting and locating defects and voids of different sizes and shapes in such reinforced plastic composites. The standoff distance and the frequency are studied as a means of increasing detection sensitivity. Comparative measurements illustrate the superiority of this technique over other nondestructive inspection (NDI) methods used on some samples.

The useful life of a glass fiber/epoxy composite subjected to impact fatigue loading is an important issue in the future design of numerous industrial components. Lifetime predictions have been a problem particularly due to the difficulties encountered in monitoring damage accumulation in composites. It is hypothesized that there is a build up of micro damage, such as matrix micro-cracks and micro-delaminations, even though there is no apparent change in material compliance. A critical level is finally reached at which time the properties of the composite begin to fall and compliance change is evident. In this study the apparent compliance change and the type of damage accumulation is investigated. To measure the compliance change, a test unit was developed that uses a dynamic load measuring system. The load cell measures the load throughout each impact pulse and the compliance and energy absorbed by the specimen is then related to the recorded curve. Initially no change in the impact pulse was apparent; however, after a finite number of cycles the peak load and area under each impact pulse drop, indicating an increase in compliance. Unfortunately, the impact load does not provide information on the form and degree of damage. Thus, millimeter wave nondestructive investigation is used, in conjunction with impact fatigue tests, to examine microstructural aspects of damage initiation and growth. The millimeter wave scanning technique results in detectable damage growth throughout the impact fatigue test. Damage size and growth patterns specific to composites are obvious, and after significant damage can be related to the observable macro damage. Continued development of these investigative techniques promises to enhance the ability of detecting defects and damage growth in fiber reinforced composite materials as well as improving the understanding of impact fatigue initiation in these complex materials.

Preliminary results are presented for a millimeter-wavelength free-space imaging system used to detect and characterize defects in layered dielectric composite slabs. Such systems throughout the microwave spectrum have shown great potential as alternative nondestructive evaluation tools for on-line, in-situ examination of low-loss dielectric materials such as plastics, ceramics, and various dielectric composites. Results of a fixed-frequency W-band imaging system operating in either through-transmission or reflection mode are presented here. Incorporation of focused-beam antennas allows high-resolution measurement of small variations within the sample under test. The results are based on measurement of the relative amplitude and phase of the reflected or transmitted wave in monostatic or bistatic configurations, respectively. With proper calibration, the measured parameters can be used to estimate dielectric property variations within the material media. A theoretical simulation for plane wave propagation in a multilayered media is used to interpret the measurement results.

This paper describes a series of experiments and calibration techniques aimed at a better understanding of observed anomalous pulsed microwave composite propagation phenomena. It is demonstrated that early wavefront contributions, associated with the precursor of pulsed microwaves propagating in air-waveguide structures, and observed slightly above the noise level of the system, exhibit small propagation times corresponding to the phase velocity of the pulsed microwave in the dispersive medium. On the other hand, lack of detection sensitivity, can lead to the observation of unphysical large time delays, and be interpreted based on the intuition of the observer, or in the worst of the cases, as noise.

Open planar resonators like rectangular and triangular microstrip resonators and slot resonators can be used to determine the dielectric constants of lossy or lossless dielectrics at microwave and millimeter-wave frequencies. These resonators can also be used to measure the thickness of dielectric slabs. In all cases, the accuracy of the method depends on the theoretical information available for the inversion of the measured data. In this paper, a general accurate pseudo-numerical technique for the analysis of single and multi-layer open planar resonators loaded with dielectric slabs is given. The technique can be used to obtain inversion curves. In the case of the microstrip resonator, coupling to the signal source is assumed to be through a coaxial probe or through a microstrip line, whereas in the case of the slot resonator, a microstrip line feeds the resonator. The technique is applied to specific cases of single and double layer microstrip resonators of rectangular and triangular shape in order to compute the shifts in the resonant frequency and changes in the input impedance when these resonators are subjected to various lossy and lossless dielectric materials of different thicknesses. Sample results are presented and discussed. In particular, it is shown that the double layer resonator offers some advantages in the measurement of the shift of the resonant frequency as one of their two resonant frequencies (i.e., the upper resonant frequency) seems to be almost independent of the dielectric loading, and hence can be used as the frequency reference.

A mathematical model is described for a new microwave nondestructive testing method for surface crack detection in metals. The crack detection method uses an open-ended waveguide to scan the surface of a metal. The method's foundation is the observation that the standing wave experiences a phase shift when a crack is present within the waveguide aperture. The model exploits the small width of the crack to reduce a Helmholtz equation on a complicated geometry to three simpler problems that are solved using the method of matched asymptotic expansions. Simplifications are possible when the crack position and cross-section are independent of distance along the crack. It is shown that the crack behaves like a resonant cavity as the crack enters the waveguide aperture.

Thickness measurements of cellulose acetate thin films (25 - 76 microns) placed directly above a conducting plate are made with W-band millimeter-waves (75 - 110 GHz). The surface wave excited by a transmitter near the sample surface is picked up by a separate receiver. The phase of the received signal was then related to the film thickness. The radiation from the transmitter is modeled as that produced by a vertical electric dipole over a two-layer medium. A contour integration is performed to derive the theoretical dependence of the phase shift on the dielectric thickness. A quadratic dependence was expected theoretically and observed experimentally. It was also found that achieving thickness resolution to several microns appears feasible. This technique displayed relatively high tolerance to variations in placement of the receiver and the dielectric samples.

Microwave nondestructive testing of lossy dielectric materials involves one of two situations: either the test sample is illuminated and a signal representing the material condition is obtained, or the material is introduced into a cavity and the measurement is done on cavity parameters. In the first case, the major problem is calculating and identifying modes in the cavity and avoiding spurious solutions. The second requires solution in the open domain and is treated here by coupled boundary element/finite element approach. The two methods are used: one uses a rather standard edge element solution and scanning of frequency to detect resonance. The second formulation is based on curvilinear edge elements which avoids spurious modes. The edge finite elements allow accurate continuity of tangential components of the electric or magnetic field and therefore do not introduce parasitic eigenvalues into the system of equations. Examples of scattering by lossy dielectrics and of loaded cavity resonators are given. These include scattering cross section of bodies and propagation in composite materials.

Millimeter wave reflectometry offers a noncontact means of measuring the thicknesses of individual layers within multi-layer dielectric coatings on metallic substrates. Complex reflection coefficients of both TM and TE/TM waves incident on coated materials are measured over a range of frequencies and grazing angles of incidence. The sensitivity of the apparatus is such that single-layer coatings differing by 0.2 microns can be distinguished.

Microwave reflection properties of four cement paste samples with various water-cement (w/c) ratios were measured daily for 28 days using microwave frequencies of 5, 9, and 13 GHz. The dielectric properties of these samples, and hence their reflection coefficients, were measured daily and shown to decrease as a function of increasing w/c ratio. This is as a direct result of curing (no chemical interaction or hydration). The presence of curing as indicated by this result indicates that microwaves could be used to monitor the amount of curing in a concrete member. The variation in the reflection coefficient of these samples as a function of w/c ratio followed a trend similar to the variation of compressive strength as a function of w/c ratio. Subsequently, a correlation between the measured compressive strength and reflection coefficient of these blocks was obtained. The early results indicated that lower frequencies are more sensitive to compressive strength variations. However, further investigations showed that there may be a frequency around 5 GHz which is the optimum measurement frequency. This result can be used to directly and nondestructively estimate the compressive strength of a cement paste and mortar blocks.

An intuitive deconvolution algorithm known s CLEAN is considered as a means of improving images associated with microwave nondestructive evaluation (NDE). Flawed metallic surfaces are scanned using a continuous wave (cw), X-band synthetic aperture radar (SAR), and images are produced using a synthetic-aperture focusing algorithm. The results are compared with those obtained by further processing the data using the CLEAN algorithm. The utility of adopting such a scheme to improve flaw detection is discussed.

An electromagnetic model for finite surface crack detection is presented. An open-ended waveguide sensor is used to scan a cracked metal surface. The crack is modelled as a large waveguide feeding a smaller one. The characteristics of the standing wave set up in the waveguide is altered when the waveguide is terminated by an imperfect short circuit load (crack surface). This is due to the generation of higher order modes as a result of the presence of the crack. Strategic probing of the standing wave properties is used to indicate the presence of a crack. A Fourier boundary matching technique is used to satisfy the boundary conditions at the waveguide aperture and the crack boundaries. The behavior of crack characteristic signals as a function of crack size and location within the sensor aperture is also studied. A finite fatigue crack is finally detected at 38 GHz to demonstrate the practical feasibility of this technique.

Two different microwave techniques for detecting fatigue or stress surface cracks in metals have been introduced. In both cases, an open-ended rectangular waveguide is used to scan over the metal surface under examination. Once a crack has been successfully detected, for both fracture analysis and repair considerations, sizing becomes the next important step. The models established for surface crack detection and their associated computer codes can be used to establish a useful crack sizing technique. The following describes a sizing technique for this purpose.

The energy-decay behavior of pulsed microwaves launched in air has been investigated with different microwave optics geometries. Experimental results, related to different geometries indicate that within the range of our measurements, the detected power referred to the peak intensity of pulsed microwave, decays much slower than those counterparts predicted by the Friis transmission equation as well as by the radar equation. In addition, a novel standing wave measuring technique is presented, where the concept of the `apparent wavelength' is experimentally demonstrated. These observed phenomena may give rise to potential applications in bioengineering, communications and control, where maintenance of the initial pulse amplitude and/or energy profile over long distances is important.

Scalar linearized inverse scattering has recently found a unified treatment within the framework of diffraction tomography in either frequency or angular diversity. The linear inverse scattering theory can be extended to electromagnetic vector fields to include complete polarization information. Its essential feature is the formulation of a vector Porter-Bojarski integral equation to be inverted by dyadic algebra. Algorithms are discussed for frequency diversity within linearized approximations for perfectly conducting and weak scattering objects, respectively. Particularly, a vector Fourier diffraction slice theorem has been obtained. These algorithms are checked against synthetic data obtained with a FDTD-code (MAFIA) to prove whether they offer advantages over non-polarimetric tomography. Hence, the FDTD-code is utilized to obtain synthetic data for a variety of scattering geometries to demonstrate the performance of vector diffraction tomography.

We describe a step-frequency microwave radar imaging system that is suitable for nondestructive evaluation (NDE) and ground-penetrating radar (GPR) applications. The system includes a computer-automated microwave measurement apparatus along with nonlinear inverse scattering imaging algorithms. Through the use of an inverse Fourier transform, the SFR data is transformed into a synthetic time-domain pulse, and imaging algorithms are applied to the time-domain data. A calibration procedure involving the use of calibration targets is described in order to remove pulse distortions due to the effective aperture and dispersive nature of the antennas, as well as those distortions due to connectors, transmission lines, directional couplers and amplifiers. Reconstructions of various metallic and dielectric scattering objects including metallic rods, glass rods and plastic PVC pipes from real measurement data collected in our laboratory are shown.

Optimum binary phase codes of length L are characterized by an autocorrelation function R((tau) ) with uniform sidelobes of level 1/L with respect to the main lobe. These optimum binary codes are called Barker codes. Binary phase codes that exhibit minimum peak sidelobes above 1/L are called suboptimum. A genetic algorithm is implemented to conduct the search for optimum and suboptimum binary codes of a given length L. In this approach, several different fitness functions are considered. These fitness functions are based on sidelobe level (SLL) and generalized entropy measures. To verify that these are reasonable fitness functions, they are first applied to sequence lengths for which optimum codes are known to exist. It is shown that if L is such that a Barker code exists, and S is a generalized entropy measure, then the Barker codes are the only ones that give the minimum value for S. It is also shown that the proposed binary phase code search is efficient for large values of L.

An advanced Ground Penetrating Radar (GPR) system has the potential for efficiently and reliably providing high resolution images for inspecting concrete civil structures for defects and damage assessment. To achieve the required performance, improvements in radar hardware, and development and adaptation of advanced 2- and 3-dimensional synthetic aperture imaging techniques are needed. Recent and continuing advancement in computer and computer-related technology areas have made it possible to consider more complex and capable systems for a variety of imaging applications not previously conceived. We developed conceptual designs, analyzed system requirements, and performed experiments, modeling, and image reconstructions to study the feasibility of improving GPR technology for non- destructive evaluation of bridge decks and other high-value concrete structures. An overview and summary of practical system concepts and requirements, are presented.

An analysis of the one-, two-, and three-dimensional electrical characteristics of structural cement and concrete is presented. This work connects experimental efforts in characterizing cement and concrete in the frequency and time domains with the Finite Difference Time Domain (FDTD) modeling efforts of these substances. These efforts include electromagnetic (EM) modeling of simple lossless homogeneous materials with aggregate and targets and the modeling dispersive and lossy materials with aggregate and complex target geometries for Ground Penetrating Imaging Radar (GPIR). Two- and three-dimensional FDTD codes (developed at LLNL) were used for the modeling efforts. The purpose of the experimental and modeling efforts is to gain knowledge about the electrical properties of concrete typically used in the construction industry for bridges and other load bearing structures. The goal is to optimize the performance of a high-sample-rate impulse radar and data acquisition system and to design an antenna system to match the characteristics of this material. Results show agreement to within 2 dB of the amplitudes of the experimental and modeled data while the frequency peaks correlate to within 10% -- the differences being due to the unknown exact nature of the aggregate placement.

In this paper we present results from a three-dimensional image reconstruction algorithm for impulse radar operating in monostatic pulse-echo mode. The application of interest to us is the nondestructive evaluation of civil structures such as bridge decks. We use a multi-frequency diffraction tomography imaging technique in which coherent backward propagations of the received reflected wavefield form a spatial image of the scattering interfaces within the region of interest. This imaging technique provides high-resolution range and azimuthal visualization of the subsurface region. We incorporate the ability to image in planarly layered conductive media and apply the algorithm to experimental data from an offset radar system in which the radar antenna is not directly coupled to the surface of the region. We present a rendering in three-dimensions of the resulting image data which provides high-detail visualization.

This paper discusses a three-dimensional synthetic aperture imaging technique based on time- domain focusing of pulse-echo radar data. We describe the basic image formation process, important data processing issues, and compensation for planar variations in the media. We present a high-resolution volumetric image reconstruction of a concrete test slab and show that we are able to identify steel reinforcing bars in the image. We conclude with a brief comparison of this imaging method with a technique based on diffraction tomography.

The objective of this paper is to present the fact that phase velocity of electromagnetic wave propagation is observable in time domain. Time domain observability of phase velocity in a rectangular waveguide was first reported by Giakos and Ishii in 1991. The validity of their observation was questioned and disputed. The questions were answered by the observers but more confirmation is desirable. This paper presents case history of time domain observed superluminal phase velocity of other authors in the past. The latest work of this author and his colleagues to confirm his earlier observation is also presented. The earliest documented observation of superluminal phase velocity can be found in Snell's law (Snell 1621). According to published time domain data of pico-second electromagnetic pulse propagation of the following cases, superluminal speed propagation of the leading edge of the pulse was observed. In a dispersive transmission system, if observed, differential speed in frequency domain dv/df is negative, the detector is responding to the phase velocity. If dv/df is positive, the detector is responding to the group velocity. This paper presents cases of dv/df negative.

To remotely image objects in a lossy inhomogeneous material requires advance knowledge on the geometry and electrical properties of the host material, as well as the positions of the transmitter and receiver relative to the host structure. Separate measurements with different apparatus are often required for obtaining this essential information. In this paper, we present a systematic approach to simultaneously retrieve information regarding both the target and the host material with a single synthetic aperture measurement. A variety of techniques have been invoked in this composite approach to extract the location, orientation, and mean propagation constant of the host structure. These parameters are subsequently used in the image focusing algorithm. It is shown that a substantial improvement in the image quality over a conventional pulse-echo system can be achieved.

This paper presents a model for computing the scattered field of a three dimensional scatterer embedded in multilayered media. The embedded object is modeled using a volume integral equation approach which allows the handling of general inhomogeneous objects. THe multilayered media is modeled using a transmission line analogy. The scattered field is obtained by obtaining the equivalent current sources in the discretization elements of the object. The convolution-form of the volume integral formulation allows the use of the conjugate gradient FFT computation method which is essential for reducing the memory storage requirements. Formulations are presented for scattering from an object in free space, a scatterer on top of a perfectly conducting surface and a scatterer embedded in multilayered media. The Green's function of the integral equation has two parts: a free space primary part and a secondary part due to the layered media. The secondary Green's function is in the form of Sommerfeld's integration which commonly has a fast oscillating integrand. Different techniques are presented to overcome this difficulty. This includes using an asymptotic solution using the stationary phase method and real axis integration. Results are presented for scattering from a sphere as an example of the object.

High temperature and high particulate density environments exist in the hopper, silo, gasifier, and combustor vessels used in advanced high efficiency coal-fired power plants. The surface and volume above a fluidized bed coal reactor contain a great deal of information about the conditions of combustion taking place within the fluidized combustion bed. The contour mapping system being developed will sense the condition of the bed surface using propagation of radio waves through the volume existing immediately above the combustion bed. This volume contains a distribution of gases and particles. The particles have a wide range of size, velocities, and distributions. In addition, the metal combustion vessel walls will create a multipath environment for the radio wave sensing signal. The paper summarizes research associated with the phased-array radar system being assembled for obtaining the surface contour of the contents in these vessels. Preliminary theoretical calculations from a simulated surface are presented. The phased array antenna system receives signals from both the near field and far field existing within the vessels.

P-31 magnetic resonance spectroscopy (MRS) and near-infra spectroscopy (NIRS) have been used to characterize the dynamic aspects of human muscle contraction. P-31 MRS is a non- invasive method for measuring ATP, phosphocreatine (PCr), and pH in exercising muscles and thereby provides information regarding oxidative and glycolytic capacities for generating high energy phosphate compounds. NIRS evaluates kinetic changes in oxygen levels in muscles during exercise and recovery. These two methods provide unique quantitative data for studies of normal muscle contraction and for more complex investigations of muscle diseases. Non-invasive MRS and NIRS examinations are readily repeatable and yield important data for longitudinal patient evaluation and therapeutic management.

The physical concept and the mathematical formalism of the wave permittivity are introduced and analyzed through the Maxwell stress tensor theory. The concept of the wave permittivity may be proved extremely useful in order to characterize propagation parameters, in terms of stored energy associated with delayed wave formation phenomena such as partial standing waves, which in turn give rise to rapid precursors observable in the time domain. Such phenomena has been observed in air, in the radiating zone, or indispersive media such as waveguides. Engineering applications related to the concept of the wave permittivity are outlined.

A method of temperature profiling in biological objects is described. In this method, thermal radiation from the object is measured as the brightness temperature by a multi-channel, topically five, 1 - 4 GHz radiometer via a contact-type, dielectric loaded waveguide antenna with a 15 X 20 mm² aperture. Coupling between the antenna and the object is analyzed under 1-dimensional approximation. The radiometric data are analyzed by a combined method of model fitting and Monte Carlo techniques to retrieve a temperature profile in tissue and its 2 (sigma) -intervals along the axis of the antenna view field. The method has been tested by numerical simulation and agar phantom experiments to demonstrate its capability of temperature profiling in muscle with 2(sigma) <EQ 1.5 K over a depth range of 0 - 4 cm from the body surface using a radiometer with brightness temperature resolution of 0.05 - 0.07 K.

The accurate knowledge of complex dielectric behavior of biological tissues at microwave frequency is of great importance in the biomedical field for its diagnostic and therapeutic uses. This information is also essential for studying the mechanism underlying the absorption of electromagnetic energy and also the emission by tissues. In the present work, in vitro complex dielectric permittivity of liver and muscle were measured in the frequency range 0.6 GHz to 1.2 GHz using a network analyzer. An open ended coaxial cable method was used and the complex dielectric permittivity was measured from the actual terminal admittance of the sample. The system was calibrated with three standard samples. The measurements were conducted at room temperature. The calculated values of (Epsilon) ' varied from 53 to 48 for liver and from 49 to 46 for muscle. The corresponding (Epsilon) ' varied from 29 to 38 and 30 to 34 respectively. The relaxation frequency and degree of dispersion were also calculated and estimation of bound water content made in each case.

Experimental and theoretical analysis on pulsed radio wave propagation measurements, obtained through a series of time domain experiments in open space as well as in a waveguide, indicate that the precursor or the very early wavefront onset, slightly above the noise level, propagates with a speed equal to the phase velocity. Since it is observable, it may be useful for radio messages. Basic principles and experimental support are presented.

An exact and compact method is determined to image the perfectly conducting bodies with one to one correspondence by high frequency scattered field data. We considered the surface waves traveling on three dimensional curved surfaces with edges by the extensions in topological approaches and constructed the rules of topological diffraction. The approach addresses the well-posed inversion methods. We obtained the scattered field via surface currents and included multiple diffractions and near field calculations. We extended the generalizations to construct the law of topological inversion. The results handle the surface traveling wave, so they carry implementations in exact and compact inversion models and in RCS computations of complex and bigger objects.

One of the major problems in sub-surface radar is the compromise between resolution and penetration depth. Stepped frequency continuous wave radars (SFCW) have improved the penetration depth of sub-surface radars by achieving greater sensitivity, instantaneous dynamic range, and better spectral control than conventional pulsed systems. However, the resolution in SFCW radar systems is limited by the fast Fourier transform (FFT) processing required to extract depth (i.e., range) information from the vector frequency data. This paper investigates the potential use of the extended prony method for range extraction in a SFCW ground penetrating radar (GPR). This method fits data to a complex exponential model, without placing the restriction on the data that the targets are constrained to definite range bins. This allows for the resolution of the system to go beyond the limitations set by the bandwidth of the waveform. Simulations are presented to examine the effects of the signal-to-noise ratio (SNR) on performance, when applying the extended prony method to a simple GPR model. At all times the results are compared with the standard FFT processing. A prototype radar system has been constructed at the University of Cape Town using standard laboratory equipment, a computer and additional digital and rf circuitry. The antennas used are two ridged wideband horns. Targets were buried in a sandpit and measurements were taken over a 2 GHz bandwidth with a center frequency of 3 GHz. Comparisons ware made between the FFT and the extended prony method for different portions of the system bandwidth, showing the extended prony method can achieve high resolution using a reduced bandwidth.