Microwave nondestructive testing of Malaysian timber for grading applications

Ghodgaonkar, Deepak K.1, Majid, Wan Mahmood B. W. A.2 and Husin, Hilmi B.3 ABSTRACT Microwave nondestructive testing methods are fast, contactless, accurate and continuous techniques for evaluation of moisture content, slope-of-grain, density of knots and specific gravity of the timber. MNDT techniques using free-space microwave measurement system involve measurement of reflection and transmission measurements in free-space. Materials such as timber which are lossy and anisotropic causes a linearly polarized electromagnetic wave to be depolarized (i. e. elliptically polarized) upon transmission through the material. The free-space microwave measurement (FSMM) system can be used to measure polarization angle and axial ratio of elliptically polarized wave transmitted by the timber specimen. The slope-of-grain which is defined as an angle between the grain direction and the direction of incident electric field, is closely related to the polarization angle. This polarization angle and axial will be calculated from measured complex transmission coefficients when transmitting and receiving antenna feeds are co-polarized and when these feeds are at an angle (which is 56.3). An experimental procedure is developed for the measurement of SOG using co-polarized and polarized transmission measurements using FSMM system. We are reporting SOG measurements for Malaysian timber (light red merant and yallow meranti) specimens using this method. INTRODUCTION Microwave nondestructive testing (MNDT) of materials is an important science which involve development of sensors/probes, methods and calibration techniques for detection of flaws, cracks, defects, voids, inhomogeinities, moisture content (MC), etc. by means of microwaves (Zoughi, 1990). MNDT techniques (such as ground probing radar, free-space microwave techniques) are increasing being used for quality control and condition assessment of concrete structures (Al-Qadi and others, 1991, Ghodgaonkar et al, 1999A). Also, for stress grading of timber and evaluation of moisture content, slope-of-grain and density of knots of timber, MNDT techniques are increasingly being used (Bostrom, L, 1994, Leicester and Seath, 1996). MNDT methods are fast, contactless, accurate and continuous techniques for evaluation of moisture content, slope-of-grain (SOG), density of knots and specific gravity of the timber (Yen, Y. H. 1981). Timber strength grading is very important for the structural safety of timber construction. This grading will be easier if the grain direction is measured under production conditions. Recently, Shen et al. have described a method for the measurement of wood grain angle using microwaves. The term microwaves refer to alternating current signals/electromagnetic waves with frequencies between 300 MHz (3 8 11 10 Hz) and 300 GHz (3 lO Hz). Since the penetration of microwaves in good conducting materials is very small, MNDT techniques are mainly used for nonmetallic materials. The spatial resolution of these techniques depends on the wavelength of the wave. For the commonly used microwave band of 3-100 GHz, wavelength varies from 100 mm to 3mm. These techniques have advantages over other NDE methods (such as radiography, ultrasonics, eddy current) regarding low cost, good penetration in nonmetallic materials, good resolution and contactless feature of the microwave sensor (antenna). Currently, microwave nondestructive testing can be best classified as a specialized technique (except for moisture gaging) as compared with other nondestructive test methods. But, modem electronics and computer processing will improve its potential for industrial applications (Botsco, 1986). For MNDT techniques, the measured parameters are reflection coefficients, transmission coefficients, dielectric constants, loss factors, complex permeabilities as a function of frequency (microwaves) and temperature. These measured
1 2

Associate Professor, Faculty of Electrical Engineering, MARA University of Technology, 40450 Shah Alam, Malaysia Professor, Faculty of Civil Engineering, MARA University of Technology, 40450 Shah Alam, Malaysia 3 Final Year B. Eng. (Hons.) Student, Faculty of Electrical Engineering, MARA University of Technology, 40450 Shah Alam, Malaysia

parameters can be related to material parameters of interest (e.g., flaws, inhomogeinities, moisture content, etc.) by suitable modeling and calibration. There are two classes of MNDT methods which are 1) free-space methods operating in the far-field region employing spot-focusing hom lens antennas and 2) waveguide methods operating in the near-field region which employ open-ended coaxial lines, rectangular waveguide, microstrip lines and cavity resonators as probes. However, in the waveguide methods, it is necessary for the composite material to be in close contact with the probe. So, these methods are not contactless. In this study, free-space methods are used which are nondestructive as well as contactless. We have developed a free-space microwave measurement (FSMM) system which can measure electromagnetic properties (complex permittivity, complex permeability, reflection coefficients, etc) for evaluation of composite materials. The main advantage of this FSMM system is that with suitable modifications, it is possible to make precise, accurate and reproducible MNDT measurements on materials under high or low temperature conditions and complex electromagnetic environmental conditions (e. g., DC biasing fields, ionizing radiation, etc) due to contactless feature of free-space measurements. Another significant advantage of free-space methods is that the measurements can be made when incident, reflected and transmitted signals are circularly/elliptically polarized electromagnetic waves. Composite materials such as timber which are lossy and anisotropic causes, a linearly polarized electromagnetic field to be depolarized (i. e. elliptically polarized) upon transmission through the material. So, free-space methods are specially suited for MNDT of materials. R. M. Redheffer (Redheffer, 1966) was the first researcher to suggest a simple free space method for measurement of dielectric constant from the measured phase of transmission coefficient. He reported that free-space methods are nondestructive and contactless techniques which are specially suited for dielectric measurement of materials. Harold L. Bassett (Bassett, 1971) was the first researcher to measure complex permiittivities in free-space using spot-focusing antennas at a frequency of 9.4 GHz. He measured complex permiittivities of fused silica as a function of temperature from ambient to 2500'C. In the last ten years, a number of freespace methods were developed for measurement of electromagnetic properties using FSMM system (Ghodgaonkar et al, 1989, 1990 and Umari et al, 1991). This measurement system consists of a pair of spot-focusing hom lens antennas, mode transitions, coaxial cables and a vector network analyzer (VNA). The inaccuracies in free-space measurements are due to two main sources of errors. 1) Diffraction effects at the edges of the material specimen/sample. 2) Multiple reflection between hom lens antennas and mode transitions via the surface of the sample. The spot-focusing antennas are used for minimizing diffraction effects and free-space LRL (line, reflect, line) calibration method implemented on VNA eliminates errors due to multiple reflections. The time domain gating or smoothing feature of VNA is used to reduce post calibration errors in reflection and transmission measurements. Previously, we reported (Ghodgaonkar et al, 1998 and Ghodgaonkar et al, 1999B) complex permittivities of Malaysian timber specimens (15.2 cm 15.2 cm 2.5-2.8 cm). They were calculated from measured values of reflection coefficients S11 and transmission coefficients S21 using FSMM system. Then, moisture content of the timber specimen was deduced by applying dielectric mixture theory to the dielectric constant values for dry and wet specimens. In this paper, FSMM system is used to measure polarization angle and axial ratio of elliptically polarized wave transmitted by the timber specimen. The slope-of-grain which is defined as an angle between the grain direction and the direction of incident electric field, is closely related to the polarization angle. This polarization angle was calculated from measured complex transmission coefficients when transmitting and receiving antenna feeds were co-polarized and when these feeds are at an angle (which is 56.3). An experimental procedure is developed for the measurement of SOG using co-polarized and polarized transmission measurements. We are reporting SOG measurements for Malaysian timber (light red meranti and yallow meranti) specimens using this method. FREE-SPACE MICROWAVE MEASUREMENT SYSTEM Fig. 1 gives a schematic diagram of the free-space microwave measurement system. A pair of spot-focusing horn lens antennas have been mounted on a large table (1.83 m 1.83 m) of 2.54 cm thick wood. These antennas (model no. 857012X-950/C) were manufactured by Alpha Industries, Woburn, MA (USA). These antennas have two-equal planoconvex lenses mounted back to back in a conical horn antenna. One plano-convex lens gives an electromagnetic plane wave and the other plano-convex lens focuses the electromagnetic radiation at the focus. For these antennas, the ratio of focal distance to antenna diameter (F/D) of the lens is equal to one and D is approximately 30.5 cm.

A specially fabricated sample holder is mounted at the common focal plane for holding planar samples. The sample is sandwiched between two perspex plates (one plate is fixed and the other is moveable). The transmit and receive horns are mounted on a carriage and the distance between them can be changed with an accuracy of 25.4 m. by using a dial indicator. From measured radiation patterns supplied by the manufacturer, the 3 dB and 10 dB E-plane beamwidths can be calculated. These beamwidths will vary in proportion to wavelength in free space (). The 3 dB and 10 dB beamwidths are approximately and 1.9 , respectively. The depth of focus for these horn lens antennas is approximately 10 . Because of spot focusing action of antennas at the focus, the diffraction effects are negligible if the minimum transverse dimension of the sample is three times the 3-dB E-plane beamwidth (which is approximately 3 ). This measurement set up covers a frequency range of 7.5-14.0 GHz. But, the same setup can be used in the frequency range of 7.5 -40 GHz by appropriate change of mode transitions. The focused antennas are connected to the two ports of the Wiltron 37269B vector network analyzer by using precision coaxial cables, rectangular-to-circular waveguide adapters and coaxial-to-rectangular waveguide adapters. The receive antenna can be rotated from co-polarized position in steps of 10 between - 70 to + 70. The polarization of transmit and receive antennas depend on polarization of the wave in rectangular waveguide used in coaxial-to-rectangular waveguide adapter. Vector network analyzer measures amplitude and phase of reflected or transmitted signal in transmission media such as coaxial line, rectangular/circular waveguide, microstrip line and free-space. A complete VNA system consists of a fast sweeping synthesized signal source, auto-reversing S-parameter test set, display unit and a controlling computer. This network analyzer is used to make accurate reflection and transmission (S-parameters) measurements in free-space using line-reflect-line calibration model. Because of multiple reflections between coaxial-to-rectangular waveguide adapters, rectangular-to-circular waveguide transitions and hom lens antennas, there is a need to calibrate the measurement system in free-space for S-parameter measurements. We have implemented free-space LRL calibration technique (Ghodgaonkar and others, 1989, 1990). This calibration technique along with smoothing or time domain gating feature of the network analyzer, can eliminate effects of multiple reflections. It is known that LRL calibration technique can produce the highest quality calibration available. Also, it is easier to realize LRL calibration standards in free-space as compared with open, short and matched termination standards used in coaxial and waveguide media. So, LRL calibration is the best calibration technique for the free-space medium. Free-space LRL calibration is implemented in free-space by establishing three standards. The reference planes for port 1 and port 2 are located at the focal planes of transmit and receive antennas. The through standard is realized by keeping the distance between two antennas equal to twice the focal distance. It means that there is a common focal plane for the through standard. The line standard is achieved by separating the focal planes of the two antennas. The distance between focal planes is approximately a quarter wavelength at mid-band, The reflect standards for port 1 (transmit hom) and port 2 (receive hom) are obtained by placing a metal plate (15.24 cm 15.24 cm 2.1 mm) on sample holder at the reference plane. LRL calibration kit for coaxial line of the vector network analyzer is modified by defining LRL standards regarding wave impedances and line lengths. Because of the characteristics of spot-focusing hom lens antennas, the electromagnetic fields in the neighborhood of common focal plane are plane wave in character, so the use of modified coaxial calibration kit is justified. The error model for LRL calibration includes error terms for directivity, isolation, source impedance match, load impedance match, transmission frequency response and reflection frequency response. This error model has 12 error coefficients which are evaluated from measured data for the LRL standards. By performing the LRL calibration using the free-space calibration standards, two-port free-space LRL calibration is obtained. For measurement of complex reflection coefficient (S11), and complex transmission coefficient (S21) of timber specimens, the reference planes corresponding to transmit and receive antennas were located at the front and back face of the sample, respectively. The residual postcalibration errors can be further reduced by using time domain gating or smoothing function of VNA. It is observed that magnitude and phase of S11 are within 0.2 dB and 1 of the theoretical value of 0 dB and 180 for the metal plate. For the through connection, the measured magnitude and phase of S21 are within 0.05 dB and 0.2 of the theoretical values of 0 dB and 0

SLOPE-OF-GRAIN (SOG) MEASUREMENTS SOG is defined as an angle between the grain direction and the direction of the incident electric field. For Malaysian timber specimens (15.24 cm 15.24 2.6 cm), we have measured axial ratio (AR) and polarization angle () of the elliptically polarized transmitted signal. SOG is closely related to the polarization angle (Yen, 1981). FSMM system is calibrated using LRL calibration technique in the frequency range of 8-12.5 GHz. Then, the receive antenna is rotated by an angle (set at 60) in the clock-wise direction with respect to the transmit antenna.

E =S S - S cos () E = sin ()
co X 21 co 21 21 Y

S S S S

= | E | + |E | = |E | - | E | = - 2 Real (E E ) = 2 Imag. (E E ) S = 0.5 tan-1 S

2 2 X Y 2 2 X Y X * Y X * Y 2 1 1 3 0

S AR = tan 0.5 sin S S = Measured S when transmit and receive antennas are co - polarized. S = Measured S when receive antenna feed is at an angle E = Complex conjucate of E
co 21 21 21 21 * Y Y

The angle can be measured accurately from |S21| measurements when there is no specimen in the sample holder. The measured value of is 56.3 which corresponds to |S21| = - 5.11 dB. Then S21 is measured in the frequency range of 812.5 GHz for co-polarized configuration and when the receive antenna is at an angle . For calculation of polarization angle and axial ratio of the received elliptically-polarized wave, we calculate Stokes parameters S0, S1, S2 and S3 (Hollinger et al, 1992 and Cornbleet, 1994) using equations given above.EX and EY are the two orthogonal components of the received elliptically-polarized wave. Table 1 gives and AR values for light red meranti and yallow meranti specimens when they the direction of electric field is approximately aligned with the grain direction and when specimens are rotated counter clock-wise by an angle which is approximately 30 with respect to transmit antenna.

Table 1

and AR value for a light red meranti and yallow meranti specimens.
Specimen (degrees) 1.35 -25.4 -1.45 -26.5 Light Red Meranti, aligned with grain direction Light Red Meranti, 30 grain direction Yallow Meranti, aligned with grain direction Yallow Meranti, 30 grain direction

AR 95.8 -2.0 -7.122 -1.86

AR values should lie between 1 (for circular polarization) and (infinity for linear polarization). From table 1, it is observed that the polarization angle values are close to the expected values (0 when specimen is aligned and 30 when specimen is rotated). Also, AR values varies in an expected pattern because of higher values when grain direction is aligned and lower values when it is at an angle of 30. The polarization angle is a good indicator of SOG if the
specimen is sufficiently thick and moisture content is sufficiently high (Yen, 1981). Otherwise it is necessary to find a depolarization index which is a ratio of and SOG. But, in this method, preliminary results indicate that SOG values are very close to . The measurement of may give SOG values. CONCLUSIONS Microwave nondestructive testing using FSMM system can give contactless and continuous monitoring of SOG for timber strength grading. By measuring co-polarized and polarized transmission measurements, polarization angle and axial ratio of the transmitted elliptically polarized wave can be measured. SOG values appear to be closely related to the polarization angle. We will be reporting more detailed measurements relationship of SOG and polarization angles for different timber specimens. REFERENCES 1. 2. Al-Qadi, I. L., Ghodgaonkar, D. K., Varadan, V. K. and Varadan, V. K. 1991. Effect of moisture on asphaltic concrete, IEEE Transactions on Geoscience and Remote Sensing, Vol. 29, pp. 710-717. Bassett, H. L. 1971, A free-space focused microwave system to determine the complex permittivity of materials to temperature exceeding 2000C, The Review of Scientific Instruments, Vol. 42, pp. 200-204. Bostrom, L. 1994. A comparison between four different timber strength grading machines" Proceeding of 9th International Symposium on Nondestructive Testing of Wood, Forest Products Society, Madison,Wl, pp. 157-167. Botsco, R. J., Cribbs, R. W., King, R. J. and McMaster, R. C. 1986, Microwave methods and applications in nondestructive testing, In Nondestructive Testing Handbook, Pual McIntire, Ed. Second Edition, Vol. 4, Section 18, American Society of Nondestructive Testing. Cornbleet, S. 1994, Microwave and Geometrical Optics, Academic Press, New York, pp. 452-456. Ghodgaonkar, D. K., Majid, Wan Mahmood B. W. A. and Majid, Rosnoizam B. A. 1999A. Accurate measurement of electromagnetic properties of concrete for nondestructive evaluation at microwave frequencies. Creating with Concrete, University of Dundee International Congress, Dundee, Scotland (UK), September 8-10, 1999, pp. Ghodgaonkar, D. K., Majid, Wan Mahmood B. W. A. and Rahim, Razimah B. A., 1999B. Accurate measurement of dielectric constants and loss tangents for nondestructive evaluation of Malaysian timber at microwave frequencies," International Conference on Timber Engineering, PTEC 99, Rotorua (New Zealand), March 15-20, 1999, pp. 87-92.

3. 4.

5. 6.

7.

8.

Ghodgaonkar, D. K., Majid, Wan Mahmood B. W. A. and Limat, Hamidah B. H. 1998. Microwave nondestructive testing of light red meranti timber. World Conference on Timber Technology, Swiss Federal Institute of Technology, Lausanne (Switzerland), August 17-20, 1998, pp. 333-339. Ghodgaonkar, D. K., Varadan, V. V. and Varadan, V. K. 1990. A free-space measurement of complex permittivity and complex permeability of magnetic materials at microwave frequencies, " IEEE Transactions on Instrumentation and Measurements, Vol. 39, pp. 398-394.

9.

10. Ghogaonkar, D. K., Varadan, V. V. and Varadan, V. K. 1989. Free-space method for measurement of dielectric constants and loss tangents at microwave frequencies. IEEE Transactions on Instrumentation and measurement,, Vol. 38, pp. 789-793. 11. Hollinger, R. D., Varadan, V. V., Ghodgaonkar, D. K. and Varadan, V. K. 1992. Experimental characterization of isotropic chiral composites in circular waveguides", Radio Science, Vol. 27, pp. 161-168. 12. Leicester, R. H. and. Seath, C. A. 1996. Applications of microwave scanners for stress grading", Proceeding of Fourth International Wood Engineering Conference, New Orleans, USA, pp. 2-435 to 2-440. 13. Redheffer, R. M. 1966, The measurement of dielectric constant. In Techniques of Microwave Measurements, C. G. Montgomery, Ed., Vol. 2, Dover, pp. 591-657. 14. Shen, J., Schajer, G. and Parkar, R. 1994. Theory and practice in measuring wood grain angle using microwaves", IEEE Transactions on Instrumentation and Measurements, Vol. 43, pp. 803-809. 15. Umari, M. H., Ghogaonkar, D. K., Varadan, V. V. and Varadan, V. K. 1991. A free-space bistatic calibration technique for the measurement of parallel and perpendicular reflection coefficients of planar samples. IEEE Transactions on Instrumentation and Measurement, Vol. 40, pp. 19-24. 16. Yen, Y. H. 1981. Microwave electromagnetic nondestructive testing of wood in real time. Ph. D. dissertation. University of Wisconson-Madison. 17. Zoughi, R. 1990. Microwave nondestructive testing:Theories and applications. in International Advances in Nondestructive Testing, Vol. 15, Gordon and Breach Publishers, pp. 255-288.

Computer

Microwave Vector Network Analyzer Wiltron 37269B

Printer

Coaxial Cables Spot-focusing horn lens antennas Antenna Mounts Sample Holder

Table

Fig. 1 Free-space microwave measurement system for microwave nondestructive testing of composite materials.