More Info at Open Access Database www.ndt.net/?id=18374 NON-DESTRUCTIVE EVALUATION OF WOOD MOISTURE CONTENT USING GPR TECHNIQUE EFFECT OF FIBER DIRECTION AND WOOD TYPE Tien Chinh. MAI 1, Zoubir Mehdi SBARTAΪ 1, Frédéric BOS 1 1 I2M Dept GCE-Laboratory, University of Bordeaux, France; e-mail: zm.sbartai@i2m.u-bordeaux1.fr Abstract This paper presents some laboratory measurements in the aim of studying the sensitivity of GPR electromagnetic waves to moisture variation in wood material. Dielectric relative permittivity was measured using reflected wave recorded by GSSI SIR 3000 system connected to 1.5 GHz antennas on several wood samples of Spruce and Pine. The results of this study show the good relations between relative permittivity and moisture content of the different wood samples. Moreover, due to the dependence of wood permittivity to moisture, GPR features in time domains present some correlations with moisture content of wood material. The effect of wood direction is significant at moisture content higher than 20% by mass. The effect of wood type between Pine and Spruce can be neglected in the relation between permittivity and volumetric water content. Keywords GPR, permittivity, moisture, timber, fiber direction, NDT. 1. Introduction Wood is a biomaterial, which is constituted of cellulose (up to 45 %), lignin (25-30 %) and hemi cellulose (20-30 %). The most common causes of wood deterioration are biological due to organism attacks (fungi and insects). Moisture content of wood is recognized as the most critical condition for the biological attack development. For example, it is known that the minimum value of moisture content for wood degrading fungi is about 17 % by mass and the optimum value varies between 30 % and 70 %. These moisture values depend on the fungi and wood type. Non-destructive evaluation of moisture content of wood material is then critical for the preservation of timber structures. Ground penetrating Radar (GPR) is a non-destructive technique based on the transmitting and receiving of electromagnetic (EM) waves. The most advantage of this technique is their capacity of investigating a large surface of a structure in a short time. It has been developed firstly for soil investigations and currently is frequently used for masonry arch bridges, concrete bridges, and road inspections [1,2]. The technique is generally used for defects detection and thickness evaluation (rebar depth, asphalt pavement, concrete slabs...). Currently, researches investigate GPR capacity to moisture evaluation of construction materials such as concrete [3] and masonry [4]. However, limited works are available regarding the field of timber structures inspection. In this field of GPR application, Muller [5] reported that this technique was well suited for the inspection of timber bridges for piping and rotting defects. The author used high frequency ground coupled dipole antenna at 1.2 GHz. The tests were carried out on girder from a bridge after its demolition. The results show clearly on radargrams figures, cracks and rotten wood. These observations are due essentially to the reflections on defects leading to the modification of the permittivity of the material between damaged and non-damaged wood. In the field of historical structures inspection, Lualdi et al [6] have used GPR technique to detect timber beams and to evaluate the type and dimension of their connection to bearing wall by 3D reconstruction imaging. They also implemented some GPR investigations for detecting small beams supporting structure in timber floors. The investigations were implemented using GPR system connected to 1 GHz antenna by recording parallel profiles orthogonal to the direction
of the timber beams. The authors reported that 3D GPR reconstruction is a powerful technique for the detection and dimension evaluation of hidden structure of timber floors. The presented studies in the literature highlight the advantage of using GPR technique at high frequency for defect detection and geometry characterization of timber structure. However, limited studies exist regarding GPR application for timber structures evaluation and typically the possibility of timber characterization such as moisture detection and evaluation for the preservation of timber structures. Recently Martinez et al [7] and Maï et al [8] implement experimental programs on wood specimens and highlight the potential of GPR for wood evaluation. This paper presents some results of research works carried out at the University of Bordeaux. GPR technique by using SIR 3000 system and 1.5 GHz dipole antennas were used in order to assess the capacity of this technique for moisture evaluation in timber structures. Spruce and Pin wood type were tested on several samples conditioned at different moisture contents and three fiber directions (longitudinal, transversal, and radial). 2. GPR Tests on wood material GPR equipment is composed of an EM waves pulse generator SIR 3000 from GSII, and a couple of transmitting (T) and receiving (R) antennas. The antennas are a ground-coupled bow-tie dipoles optimized for a frequency of 1.5 GHz. The measurement is based on the radiation of EM energy from T to R antenna. GPR tests were carried out on samples of 20 x 18 x 8cm of Spruce and Pine wood. For each wood type, three samples were tested and three signals were recorded for each sample. The measurement consisted in placing the coupled antennas on the surface of the sample in three directions (longitudinal, transversal, and tangential) with respect to electrical field direction (Fig. 1). The samples were tested at several moisture contents from oven dried to a moisture content of 50% by mass. a) Longitudinal direction b) Transversal direction c) Radial or tangential Figure 1. The different configurations of GPR measurements on wood samples. Signals analysis consisted first in filtering the signals using FIR filter from 300 MHz to 3000 MHz, the amplitude and arrival time of the reflected wave from the bottom of the sample are extracted and analyzed. The analysis consisted in calculating the permittivity (dielectric constant) using the arrival time of the air wave (directed wave recorded in air at fixed offset of 59mm) and the reflected wave by the following relation: ε! = V 2 0 ( V )...(1)
V is calculated by : V = d r Δt + d ER V 0...(2) Where V 0 is the velocity of EM in air, d ER is the fixed offset (59 mm), Δt is the delay between air wave and reflected wave. dr is the propagated distance of the reflected wave in the material that can be estimated by the following relation: were e is the thickness of the sample d r =! # " d ER 2 $ &+ e 2 %...(3) 3. Results and discussion Typical GPR signals recorded on wood samples and in air are presented in figure 2. The signal represented by blue dotted line is a typical signal recorded in air. It is the direct wave transmitted directly and received at an offset of 59mm. This signal is taken as a reference for time zero calibration. The signal in red line is a typical signal recorded on wood specimens. It is composed of two signals. The first pick represents the transmitted signal (direct wave which is a mix of air wave and the direct wave propagated in the wood material). The second part of the signal is the reflection on the bottom of the sample (reflected wave). This figure shows that the signal recorded on the wood sample is attenuated and the arrival time is increased. Air wave (direct wave recorded in air) Direct wave Normalized amplitude Reflection from the bottom Time (ns) Figure 2. Typical signals recorded on wood samples for longitudinal and transversal directions. The effect of moisture on GPR signal is presented on figure 2 for different moisture content (11.6, 11.8, 20.7, and 35.9 %). From this figure, it is clearly that the signal presents a significant high attenuation and an increase of the arrival time of the reflected signal. These results implies that the GPR waves loss energy and decrease the velocity for higher moisture content due to the increase of polarization effects because water have higher dielectric constant of about 80 comparing to that of dry wood (approximately equal to 2).
Figure 2 shows also a significant effect on the direct wave. Increasing moisture content decreases the amplitude of the direct wave and increases the arrival time. This typical wave can be of great interest in the case of timber structure evaluation when the reflected wave is not clearly identified (high attenuation or thick wood elements) or in the case of unknown thickness for the calculation of the propagated distance. Several papers discuss the use of the direct wave for moisture evaluation as concrete [3]. Currently, this typical signal is analyzed for studying the effect of moisture and fiber directions. However, in the case of this study only the reflected wave is analyzed. Normalized%amplitude% Time%(ns)% Figure 3. Effect of moisture on GPR waveform - Typical signals recorded for moisture content of 11.6, 11.8, 20.7, and 35.9 %. The relation between the dielectric constant and moisture content for Spruce and Pin is presented in figure 4a and figure 4b, respectively. It can be seen similar behavior for both types of wood, which consist of increasing of dielectric constant with respect to moisture from 2% to 35%. However the behavior presents two slopes with a change at approximately 20% of moisture. This point can be attributed to the point of fiber saturation. Below this point, water in the material is completely bound; the permittivity is then very low. In contrast, when increasing moisture higher than this point, the volume of free water increases that explain the increase of the slope. The effect of the saturation point is more clearly observed in the longitudinal direction due probably to the fact that the polarization of water molecule is higher compare to transversal and radial directions. This figure shows also the effect of the fiber direction on the permittivity of the tested wood types. The permittivity measured in the longitudinal direction with respect to the E field is higher than the transversal and radial directions. This is because the polarization effects are higher in the longitudinal direction as reported by [9-11] and Maï et al. [8] from weak perturbation method measurements on wood specimens at 1.2 GHz that is very close to GPR frequency used in the presented study. This fiber direction effect is not significantly observed at low moisture content until 20% where the water is bound. Increasing moisture content increases the difference in permittivity between measurements at different fiber directions.
a) b) Spruce) Pin) Dielectric)constant) moisture)content)by)mass)(%)) moisture)content)by)mass)(%)) Figure 4. Relationship between dielectric constant and moisture content for longitudinal, transversal, and radial directions. a) for Spruce wood, b) for Pin wood Figure 5 presents the relationship between dielectric constante and volumetric water content for the to tested types of wood (Spruce and Pin) for the three directions (longitudinal, transversal, and radial). The mening of using volumetric water content rather than water content by mass is to avoid the effect of the difference in density between both wood types. The measured density at 12% of moisture (air dried condition) is about 430 kg/m3 for Soruce and 550 kg/m3 for Pin. For the same moisture content by mass the total mass and volum of water in Pin is higher compare to Spruce due to the fact that the Pin density is higher. When represnting the water content as the volume of water by percentage of the volume of the specimens (20 x 18 x 8 cm) on figure 5, very similar behavior can be seen for Spruce and Pin. From this figure, it is clare that in the case of Spruce and Pin, the dielectric constant is independent of the wood type. It can be seen also that the radial and transversal directions are verry close and the longitudinal direction is higer. This result have been also reported by some authors [9-11] for the cases of dielectric properties measurement using microwave technique. Dielectric'constant' Volumetric'water'content'(%)' Pin( Longitudinal' Spruce'(' Longitudinal' Pin( Transversal' Spruce(' Transversal' Pin( Radial' Spruce('Radial'! Figure 5. Relationship between dielectric constant and volumetric water content for Spruce and Pin in the three directions (L, T, R)
4. Conclusion The results show high sensitivity of GPR wave propagation to moisture variation in wood at 1.5 GHz frequency. GPR measurements indicate that the permittivity measured at 1.5 GHz is affected by moisture, fiber direction with a low effect of wood oven-dry density (difference between Spruce and Pine). The results show double slopes linear dependence of the permittivity to moisture content with a change of the slope at the fiber saturation point. The effect of moisture content at E field oriented at longitudinal direction is shown to be more significant than E field oriented at transversal and radial directions. Radial and transversal directions have very close permittivity. It has been shown that this measurement approach can be an efficient method to define the permittivity behavior of wood material and to detect moisture variation of timber structures. References [1] Bungey, JH. (2004), Sub-surface radar testing of concrete: a review, Construction and Building Materials, 18, 1-8. [2] McCann, DM., Forde, M. (2001), Review of NDT methods in the assessment of concreteand masonry structures, NDT&E International, 34, 71 84. [3] Sbartaï, ZM., Laurens, S., Balayssac, J.-P., Ballivy, G., and Arliguie, G. (2006), Effect of concrete moisture on radar signal amplitude. American Concrete Institute Materials, 103, 6, 419-426. [4] Maierhofer, C., Wöstmann, J., Trela, C., and Röllig, M. (2008), Investigation of moisture content and distribution with radar and active thermography. RILEM International Conference, SACOMATIS, Varenna, Italy. [5] Muller, W. Timber girder inspection using ground penetrating radar, NDT-CE international conference. [6] Lualdi, M., Zanzi, L., Binda, L. (2003). (2003), Acquisition and processing requirements for high quality 3 D reconstructions from GPR investigations NDT-CE international conference. [7] Martínez-Sala R, Rodríguez-Abad I, Diez Barra R, Capuz-Lladró R. Assessment of the dielectric anisotropy in timber using the nondestructive GPR technique. Construction and Building Materials, Volume 38, January 2013, Pages 903-911. [8] Tien Chinh Maï, Stephen Razafindratsima, Zoubir Mehdi Sbartaï, François Demontoux, Frédéric Bos. Non-destructive evaluation of moisture content of wood material at GPR Frequency, Construction and Building Materials, 77 (2015) 213 217. [9] Gary, S., Schajer, F., Bahar, O. (2006), Measurement of wood grain angle, moisture content and density using microwaves. Holz als Roh- und Werkstoff, 64, 483 490. [10]Johansson, J., Hagman, O., Fjellner, BA. (2003), Predicting moisture content and density distribution of Scots pine by microwave scanning of sawn timber, J Wood Sci, 49, 312 316. [11]Torgovnikov, GI. (1990), Dielectric properties of wood and wood based materials, Springer, Berlin Heidelberg, New York,196pp.