3D MICROWAVE IMAGING TECHNOLOGY USING ANTENNA ARRAY FOR DAMAGE ASSESSMENT OF CONCRETE STRUCTURE

Transcription

1 16th ASCE Engineering Mechanics Conference July 16-18, 23, University of Washington, Seattle 3D MICROWAVE IMAGING TECHNOLOGY USING ANTENNA ARRAY FOR DAMAGE ASSESSMENT OF CONCRETE STRUCTURE Yoo Jin Kim 1, Associate Member ASCE, Luis Jofre 2, Franco De Flaviis 3, and Maria Q. Feng 4, Associate Member ASCE ABSTRACT Various nondestructive evaluation (NDE techniques have been studied to locate steel rebars or dowels, and to detect invisible damage such as voids and cracks inside concrete and debonding between rebars and concrete caused by corrosions and earthquakes. In this study, 3-dimensional (3D microwave imaging technology was developed to detect such damage and to locate steel rebars or dowels. Extending their earlier analytical work on 2D image reconstruction, the authors have developed an imaging algorithm with visualization software for reconstructing a 3D volume image. This paper presents a numerical focusing procedure, which allows the recovery of a 3D object from its scattered field. A multi-frequency technique was applied in order to improve the quality of the reconstructed image. The effectiveness of the numerical focusing and the multi-frequency techniques was demonstrated through numerical simulations. Furthermore, innovative design of the antenna array is described, in which a hybrid approach for network distribution was used. Keywords: Microwave, Nondestructive Evaluation (NDE, 3D Imaging Technology, Damage Detection INTRODUCTION The use of microwave for investigating the integrity of structural elements is becoming established as a nondestructive evaluation method in civil engineering, especially for detection of invisible damage and objects such as steel rebars or dowels. In the previous works (Feng et al., 22, Kim et al., 22, the authors developed microwave surface imaging technology and two-dimensional (2D sub-surface imaging technology for the assessment of bonding condition of FRP-jacketed concrete structures. In these studies, the locations and areas of the voids/debonds between the FRP jacket and the concrete were successfully assessed and 2D sub-surface reconstructed image showed the main features of the objects such as air void and steel inside concrete. 1 Dept. of Civil & Environmental Eng., University of California, Irvine, CA yook@uci.edu 2 Dept. of Signal Theory & Comm., Technical University of Catalonia, Barcelona, Spain. luis.jofre@upc.es 3 Dept. of Electrical & Computer Eng., University of California, Irvine, CA franco@uci.edu 4 Dept. of Civil & Environmental Eng., University of California, Irvine, CA mfeng@uci.edu

2 In this study, three-dimensional microwave imaging technology was developed to detect such damage and to locate steel rebars or dowels. Extending the earlier analytical and experimental works on 2D image reconstruction, the authors have developed a 3D image reconstruction algorithm with visualization program and designed an antenna array module, which will be a part of whole antenna array. This paper presents the analytical procedure to reconstruct 3D volume image from the scattered field using numerical bi-focusing operator. A multi-frequency technique was applied in order to improve the quality of the reconstructed image. The effectiveness of the use of bi-focusing operator and multi-frequency technique was demonstrated through numerical simulation. Finally, the design of 4 4 antenna array module, which is a part of whole antenna array, was introduced with modified multiplexing switching technique and feeding network. 3D IMAGE RECONSTRUCTION ALGORITHM AND SIGNAL PROCESSING Analytical Formulation As illustrated in Fig. 1, the system consists of N k N l antenna elements forming a transmitting array and N m N n antenna elements forming a receiving array. In the measurement, an N t N r measurement matrix, where N t is the total numbers of transmitting antenna elements and N r is the total numbers of receiving antenna elements, can be obtained as follows: for every selected transmitting element, the receiving array is scanned obtaining an N r -measurement column, then the procedure is repeated for the rest of the N t transmitting elements. Following the electromagnetic compensation principle, the illumination of an object (defects such as an air void induces an equivalent electric current distribution, and this distribution makes an electromagnetic image of the object (Gironés et al., The reconstruction algorithm forms an image by means of the synthesis of two focused arrays (transmitting and receiving arrays. All the signals from the antenna elements of the arrays are weighted by a focusing operator so as to be focused on a unique point. This can be achieved by a numerical treatment of the measurement matrix, explained as follows. The focusing operator can be obtained by taking an inverse of the field induced by a current point source. It is well known that the electric fields of the point electric source are proportional to the Green s function, e -jkr /r. Therefore, the incident field at r i (x i,y i,z i when focusing from every transmitting point r Tkl (x Tkl,y Tkl,z Tkl on the reconstructing point r f (x f,y f,z f can be expressed as Nl Nk jke rtkl ri e E ( x, y, z = I ( x, y, z r i i i i Tkl f f f l= 1 k= 1 Tkl r i (1 jke rtkl rf where I Tkl (x f,y f,z f, the focusing operator, is given by I ( x, y, z = r r. Tkl f f f Tkl f e Scattered field measured at r Rmn (x Rmn,y Rmn,z Rmn of a defect placed at r (x,y,z is e Es( xrmn, yrmn, zrmn = Ei( x, y, z Iobj r jk e r Rmn Rmn r r (2 where I obj is a constant for every object containing its electromagnetic macroscopic characteristics. When focusing back the received field at r Rmn (x Rmn,y Rmn,z Rmn on the interest point r f (x f,y f,z f, 2

3 electromagnetic image of Ef (xf,yf,zf at rf (xf,yf,zf can be expressed as E f (x f, y f, z f = Nn Nm Es ( xrmn, yrmn, zrmn I Rmn ( x f, y f, z f n =1 m =1 (3 where IRmn(xf,yf,zf, the focusing operator, is given by I Rmn ( x f, y f, z f = rrmn r f e jke rrmn r f. Finally, all the processes can be grouped as follows: No jk r r N l e e Rmn, p I Rmn ( x f, y f, z f I obj, p rrmn r, p l =1 n =1 m =1 p =1 Nn Nm E f (x f, y f, z f = Nk ITkl ( x f, y f, z f k =1 jk r r e e Tkl, p rtkl r, p (4 In addition to the reconstruction algorithm, the post-process program for displaying the measurement and assessment results was developed. The post-process program makes the user to identify the feature of an object more easily. The program, first, displays cut images at several vertical locations, and then the user can pick one of the cut images to perform a detail investigation. Also, the post-process program is integrated with the image reconstruction algorithm, which makes it possible to carry out the on-site measurement and assessment within several minutes at most. Figure 2 shows an example screen of the post-process program, in which one of the cut images are magnified in another pop-up window. (a Several cut images along z-axis (b Magnified image at z=mm level Figure 2. Example screen of post-process program Resolution of the System The image reconstruction algorithm developed in the previous section was applied to the case of two (transmitting and receiving arrays each consisting of 128 (8 16 antennas, with the frequency of 5.2 GHz, while the wavelength in concrete, λe, is 25.6 mm. Numerical simulations using analytical measurement data were conducted in order to verify the resolution capability of the system. Two point-like objects were placed with the distance of 25. mm 3

4 along each axis at the center of the reconstructing volume. As shown in Fig. 3, the results demonstrated that the system, due to the use of bi-focusing (focusing both the transmitting and receiving arrays, is able to achieve a resolution in the order of the wavelength in the dielectric medium, which is 25. mm at 5.2 GHz in concrete. (a X-direction (b Y-direction (c Z-direction Figure 3. Resolution of the system (25mm at 5.2GHz Multi-Frequency Technique In order to improve the quality of reconstructed image by reducing the background noises, a narrow band multi-frequency technique was applied. In this technique, microwaves with several frequencies near the designed illuminating frequency were sent from and received by the antenna arrays. By averaging the images reconstructed from the waves of different frequencies, background noise including the steel interference can be averaged out to improve the signal-to-noise ratio. This technique has been used with some success for similar problems (Pierri et al., 2. According to the frequency sweeping, both incident/scattered field and numerical focusing operator become a function of frequency: Ei( xi, yi, zi Ei( xi, yi, zi, ken Es( xrmn, yrmn, zrmn Es( xrmn, yrmn, zrmn, ken Ef ( xf, yf, zf Ef ( xf, yf, zf, ken I ( x, y, z I ( x, y, z, k Tkl, Rmn f f f Tkl, Rmn f f f where k en, the effective wavenumber at the N th frequency, is given by en N r en (5 k = k ε (6 Then, the electromagnetic image of E f (x f,y f,z f at r f (x f,y f,z f can be expressed as 4

5 I ( x, y, z, k Rmn f f f en NFreq Nn Nm No jken rrmn r, p Nl Nk jken rtkl r, p N= 1 n= 1m= 1 Iobj, p ITkl ( xf, yf, zf, ken 1 Rmn, p 1 1 Tkl, p p= r r r r l= k= Ef ( xf, yf, zf = e e where NFreq is the number of frequency used. The multi-frequency technique was applied to reconstruct the images of two point-like objects separated 25.mm, which is the wavelength (λ e in concrete at 5.2GHz. Four frequencies of 4.6, 4.8, 5., and 5.2GHz were used in this case. As shown in Fig. 3, the background noise was successfully reduced with the images of the point like objects more clearly reconstructed, when using multi-frequencies. ( Y (mm -2 Y (mm X (mm X (mm (a Using single frequency (b Using multi frequencies Figure 3: Improvement of image quality by the use of multi-frequency technique. DESIGN AND FABRICATION OF SLOT ANTENNA ARRAY Before making the whole antenna array, 4 4 antenna array was designed as a module of the whole antenna array in order to investigate the switching and the feed network. The design of the 4 4 antenna array can be directly expanded to the whole array, or can be a sub-module of the whole array, i.e. the whole array consists of sixteen 4 4 modules. The illumination frequency on a working medium of concrete (ε r = 5.3 was chosen as 5.2GHz, which is a trade-off between the resolution and the penetration depth, and the type of single antenna element was determined as a microstrip-fed single antenna. The structure of slot antenna is illustrated in Fig. 4(a. The simulation and measurement results, as plotted in Fig. 4(b, are slightly different, but the radiation at the illumination frequency of 5.2GHz is acceptable. There were two main considerations in the design of large antenna array: switching each antenna element and feed network to deliver the microwave from the feed to the antenna element. 5

6 In order to switch on/off active/inactive antenna elements, multiplexing technique with diode switch was used. However, the signal cannot be fully transferred from the port to the end of each antenna due to the losses at each diode (3dB/diode loss, which results 21dB accumulated losses for 128 elements. In this study, multiplexing technique with some modifications was used: instead of using diode switch at every junction, only 1 diode switch was used at the entrance of every element avoiding the loss at diodes, and the feeding microstrip (or slot was modified to satisfy the matching condition at the port. Feed Slot Substrate Foam Reflector Magnitude (db Measurement Simulation (HFSS -2 (a Structure of slot antenna Frequency (GHz (b Simulation and measurement results Figure 4. Microstrip-fed slot antenna Another consideration lies on feed network to deliver the microwave signal with minimizing loss and keeping bandwidth. The problem is the loss of bandwidth mainly due to the T-junctions of the network. To avoid this problem, the microstrip-slot transition, which can give us a broad bandwidth, is invoked (Aikawa et al., Microstrip-to-slot feed improves the bandwidth of the antenna than all microstrip feed. The dimension of the microstrip-slot transition is optimized using Agilent s Momentum software. As shown in Fig. 5, the geometry for the 4 4 antenna array was optimized with two ports (port 2 indicates activated antenna element with large bandwidth from 4.6GHz to 5.8GHz under 1dB. Port 2 Slot Port 1 Microstrip Magnitude (db Frequency (GHz (a Geometry of feed network (b Reflection (S 11 at port 1 Figure 5. Optimized microstrip-slot transition 6

7 Using the modified multiplexing technique for switching and microstrip-slot transition as a feed network, the design of 4 4 antenna array was finalized and fabricated as shown in Fig. 6. The authors are currently working on the measurement of the module, and the whole antenna array will be designed based on the design of the 4 4 module. (a Feed network (b Slot antenna and feed Figure 6. Photos of fabricated antenna array CONCLUSION In this study, 3D microwave imaging technology was developed in order to detect invisible damage and to identify the location of steel rebars or dowels. The numerical simulation results showed that the resolution of the system was obtained in the order of the wavelength in a medium by using the bi-focusing operator. The design of 4 4 antenna module optimized with the modified multiplexing technique and microstrip-slot transitions gave a large bandwidth and good radiation performance of the antenna array at the illuminating frequency of 5.2GHz. ACKNOWLEDGMENTS This work was supported by California Department of Transportation under Award No. 65A14. REFERENCES Feng, M. Q., De Flaviis, F., and Kim, Y. J. (22, "Use of Microwaves for Damage Detection of Fiber Reinforced Polymer-Wrapped Concrete Structures", Journal of Engineering Mechanics, ASCE, 128(2, Kim, Y. J., Jofre, L., De Flaviis, F., and Feng, M. Q. (22, "Microwave Sub-Surface Imaging Technology for Damage Detection of Concrete Structures", Proceedings of 15th ASCE Engineering Mechanics Division Conference, , New York, NY, June 2-5. Gironés, I. M., Jofre, L, Ferrando, M., De Los Reyes, M, and Bolomey, J. Ch. (1987, "Microwave Imaging with Crossed Linear Arrays", IEE Proceedings, 134(3, , June. Aikawa, M., and Ogawa, H. (1989, "Double-Sided MIC s and Their Applications", IEEE Trans. on Microwave Theory and Techniques, 37(2, , February. Pierri, R., Brancaccio, A., and De Blasio, F. (2, "Multifrequency Dielectric Profile Inversion for a Cylindrically Stratified Medium", IEEE Trans. Geoscience and Remote sensing, 38(4,