High Spatial Resolution Distributed Fiber Optic Technique for Strain and Temperature Measurements in Concrete Structures

International Workshop on SMART MATERIALS, STRUCTURES NDT in Canada 2013Conference & NDT for the Energy Industry October 7-10, 2013 Calgary, Alberta, CANADA High Spatial Resolution Distributed Fiber Optic Technique for Strain and Temperature Measurements in Concrete Structures Wenhai Li, Xiaoyi Bao Department of Physics, University of Ottawa, wli@uottawa.ca ABSTRACT Structural Health Monitoring (SHM) is a promising approach that provides a real time diagnosis of the state of wear or damage of an infrastructure. Distributed fiber optic sensors (DFOS) are ideally suited monitoring strain in concrete structures due to their small size, low cost, ability to be embedded internally, and multiplexing capabilities. DFOS are able to provide relevant information on large structures, such as bridges, buildings, and dams, yielding information about static and dynamic strain, temperature, and wind or water pressure. In this report, we present a new instrument based on Optical Frequency Domain Reflectometry (OFDR), which enables temperature and strain measurements with a tens of millimeter spatial resolution over hundred of meters and with a level of accuracy of 1 strain and 0.1 C. The tests were carried out using polarization maintaining sensing fiber surface-mounted to the surface of plain concrete specimens. Measurements were performed with an OFDR instrument, while mechanical loads were imposed to the specimens. Preliminary experiments seem very promising as the OFDR measurements are found comparable to values obtained with conventional sensors used in civil engineering. Moreover, the OFDR distributed sensing system makes it possible to detect and localize concrete cracks during the mechanical loading. Keywords: Optical frequency-domain Reflectometry, Fiber measurements, Rayleigh scattering, Distributed fiber optic sensors, Structural Health Monitoring INTRODUCTION Distributed Optical fiber sensors (DOFSs) have been attracting intensive research all over the world for several decades, and they have already shown superior advantage over their conventional electrical counterparts, which attributes to the distributed capability. A fully DOFS is usually operated by measuring the surrounding environment change along the length of the sensing fiber, which is very useful for Structural Health Monitoring (SHM) of large structures such as bridges, pipelines, oil wells, dams and other civil constructions[1]. Optical fibers consist in light waveguides made of silica glass; they are light, small size, insensitive to electromagnetic 1

fields, and low attenuation of the transmitted light signal. They enable to perform long range transmissions in the kilometre range. In DOFS, optical fibers act not only as the transmission medium but also as a continuous transducer, and a DOFS system can be considered as thousands of multiplexed measurement points obtained using a single interrogation unit using an optical fiber as sensor. Several DFOS techniques have been developed based on measurement of intrinsic backscatter of fibers. These include techniques based on Raman, Brillouin, and Rayleigh scattering [2-10] as well as those involving optical low-coherence reflectometry (OLCR) and frequency domain reflectometry (OFDR) [11-14]. Techniques based on Raman and Brillouin scatter measurement employ optical time domain reflectometry (OTDR), which operates by mapping position to the time of flight for a pulse to travel to and from the sensing location. OTDR technology has found wide applications for long distance distributed measurement due to its larger dynamic range, while its measurement resolution is usually in order of meter to tens of meter, which is not well suited for applications that require high resolution. OLCR technique utilizes low-coherence detection techniques to achieve super high spatial resolution (micrometer), but its measurement range is only reach the order of several meters. This limits its application only in fields of optical coherence tomography (OCT). OFDR, which uses continuous frequency modulated optical wave probing, is characteristic by high spatial resolution and large dynamic range. The coherent detection scheme of OFDR gains the high sensitivity down to -100dB and a space resolution in millimeter range can be achieved. OFDR fills the gap in measurement range between OTDR and OLCR, and it becomes very attractive for practical temperature and strain monitoring applications such as SHM and smart material, etc. This last method is considered in our report. On the other hand, a major limitation of DFOS is that DFOS is sensitive to both temperature and strain, which introduces measurement errors with sensor systems designed to monitor strain, whereas temperature variations along the sensing fiber could lead to unwanted, thermal-apparent strain readings. The conventional way to discriminate temperature and strain is to prepare two physically separated sensing fibers, one experiences both strain and temperature changes while the other experiences only temperature changes. Thus strain data can be obtained by compensating the temperature effects. The use of single sensing fiber, however, is always demanded when the minimum intrusion of embedding sensors is required. So far, a number of temperature/strain discrimination techniques using a single sensing fiber have been proposed, which include methods to examine Brillouin frequency shifts, Brillouin gain amplitudes [15-18], hybrid Raman-Brillouin gains [19-21], and the use of specialty fibers such as large effective area non-zero dispersion shifted fiber (LEAF), etc. These methods usually have large spatial resolution in order of meters. An approach to discriminate temperature/strain with higher spatial resolution of mm or submm order is a dual wavelength interrogation technique [22], in which two different wavelength fiber Bragg gratings (FBG) are written at the same location of sensing fiber. An FBG is a sensing element inscribed into a fiber by UV exposure [23, 24] and reflects a spectrum with a particular wavelength (Bragg wavelength) when a broad band incident light is inserted. The Bragg wavelength shifts in proportion to strain or temperature variations, so that observation of Bragg wavelength shifts allows strain or temperature measurements. As the strain and temperature responsivities of FBG depend on the photo-elastic and thermo-optic coefficients of optical fiber, when two different wavelength FBGs are written at the same location, these two parameters 2

exhibit different variations with wavelength, thus permits the discrimination between these two parameters. FBG can also be imprinted into a polarization maintaining (PM) fiber [25, 26], and it reflects two Bragg wavelengths corresponding to two polarization modes, fast and slow modes, due to the birefringence effect in the PM fiber. Two measurands of temperature/strain are determined by observing two Bragg wavelengths, which have different sensitivities toward temperature/strain between the two modes. It was shown that temperature/strain sensitivities for orthogonally polarized fundamental modes of PM fiber differ enough to allow discrimination of the two parameters. Methods that employ FBGs can achieve higher resolution but are often limited by the number of gratings that can be multiplexed in a single fiber. In principle, these methods are compatible with distributed sensing technique based on OFDR with the use of a PM fiber, in which every segment cell along fiber can be treated as FBG, and the applied temperature or strain effectively shifts the spectrum reflected from the segment cell. The significant advantage of the technique is that huge number of FBGs can be replaced by a single PM fiber to make distributed measurement. Froggatt et al. [27] demonstrated a simultaneous temperature/strain measurement based on this technique, and temperature/strain resolutions of 3.5 C/35μ over 35m length were reported. The measurement errors were mainly limited by position-dependent temperature and strain sensitivities induced by distributed birefringence along sensing fiber. In this report, we used the distributed autocorrelations and cross-correlations of the spectral signatures to calculate the temperature/strain coefficients of the PM fiber to form a distributed parameter matrix, which can be used to compensate measurement errors. The proposed sensor system achieved temperature and strain accuracy of 1 strain and 0.1 C simultaneously with 6.2mm spatial resolution. The high performance of the sensor system makes it very attractive for practical temperature/strain monitoring in structural health monitoring applications. DOFS BASED ON RAYLEIGH BACKSCATTERING ANALYSIS BY OFDR USING POLARIZATION MAINTAINING FIBER Fundamentals and operating principles of Rayleigh OFDR system are explained in great details in [11-14, 27-28], and two sets of Rayleigh backscatter patterns as a function of PM fiber length are needed. First, the Rayleigh scatter pattern at an ambient temperature is measured and stored as reference profile and then the second scatter pattern is measured with heat or stress applied at some section of the fiber, which is stored as measurement profile. Then two short segments of Rayleigh scatter data at the same position along the fiber are extracted from the reference and measurement profiles and transformed back into the frequency domain using a FFT transform. This data processing forms the amplitude data profile as a function of optical frequency. Usually the scatter profile from a segment is comprised of a random set of complex numbers, which represents a permanent structure in the fiber core. To determine the spectral shift between the reference and perturbed scans, a cross-correlation is performed for each fiber segment. Together with the distributed autocorrelation of the spectral signatures of PM fiber, a distributed parameter matrix of the temperature/strain coefficients is formed to compensate measurement errors. A PM fiber has two polarization modes, which are called fast and slow modes, due to the polarization dependence of the index of refraction of the fiber core. PM fiber is fabricated by 3

introducing stress in the core via a non-circular cladding cross-section, or via rods of another material included within the cladding, as depicted in Fig. 1. Fig. 1: PM fiber cross-section view. Since the refractive indices of the fast and slow modes are different, a reflector of a fiber segment will have two distinct reflection wave-numbers corresponding to the two modes. When the segment of fiber experiences a change in temperature/strain, the reflected spectrum from the two modes are shifted with respect to one another; This spectral shift between the two modes can be calibrated as the change of temperature or strain by calculating the autocorrelation of the Rayleigh scatter spectral profile. A typical autocorrelation of a fiber segment scatter is shown in Fig. 2(a). Fig. 2: Depicts show temperature and strain measurement principle. (a) Autocorrelation of spectral scattered data obtained from a segment of PM fiber when a segment fiber cell is with/without heat or stress. (b) Cross-correlations of spectral scattered data obtained from a segment of PM fiber between reference (no heat/stress) and measurement data (with heat or stress). Both a and c are proportional to temperature and strain applied to the fiber segment. In Fig. 2(a), the horizontal axis represents the frequency and the vertical axis the amplitude of autocorrelation of the Rayleigh scatter spectrum. The solid curves of three peaks are obtained through autocorrelation of reference spectrum (in the case that there is no heat or stress applied to the fiber), whereas the dotted curves present the autocorrelation of measured spectrum as the fiber segment is heated or stressed. The central peak is the zero-shift peak of the autocorrelations and two side peaks (dotted and solid) are symmetrically located at both sides of the center peaks. 4

The frequency shift between the side peak and the center peak are induced by the permanent structure of the PM fiber, which are related to the spectral shifts between its slow and fast modes. As these spectral shifts are intrinsically related to local fluctuations in the fiber backscattering coefficients due to temperature or strain variation, the change of the spectral shifts, a, can be used to measure the change of temperature/strain applied to the fiber segments. Similarly, a cross-correlation between the amplitude spectra of the measured scatter segment and the reference segment presents a frequency shift, which is also proportional to temperature and strain applied to the fiber. As indicated in Fig. 1(b), two set of cross-correlation data of each with three peaks, are obtained from a segment of PM fiber with/without temperature/strain applied. As all three peaks are shifted from center by an equal amount, the frequency spectral shift c between the two central peaks can be used to measure the change of temperature/strain applied to the fiber segments. The autocorrelation and cross-correlation shifts, a and c, induced by temperature/strain changes, T and, respectively, can be described by the following matrix equation: a c Ta Tc ( ( T ( a c ( The elements Ta(, Tc(, a( and c( are coefficients of autocorrelation and crosscorrelation spectral shifts with respect of temperature/strain, respectively, and z denotes segment cell at the particular position along the PM fiber. Here, the length of the segment is 6.5mm, or 880 data points in Rayleigh scatter time-domain signatures. In the case of Ta(, Tc(, the of autocorrelation and cross-correlation shifts to temperature in the function of location, a reference Rayleigh scatter profile is first taken and stored at ambient temperature. Then a series of measurement Rayleigh scatter data are taken after a series known temperature changes are applied to the fiber accordingly, while the fiber is in loose condition. A fiber segment cell, z of Rayleigh scatter data at same position along PM fiber are extracted from the reference and measurement profiles and transformed back into the frequency domain using FFT transform; In frequency domain, the autocorrelation and cross-correlation shifts, fa and fc, are calculated, respectively. Therefore Ta and Tc, the coefficients of autocorrelation and cross-correlation frequency shifts to temperature can be calculated by Ta = fa/ T and Tc = fc/ T, respectively. By scanning the whole length of PM fiber with same segment length z and repeating the autocorrelation and cross-correlation calculations, the localized coefficients of frequency shifts to temperature, Ta( and Tc( in the function of fiber length can be obtained. The coefficients of frequency-shift to strain along the sensing fiber, a( and c( can also be calculated with the similar manner, whereas a reference Rayleigh scatter data is first taken when the PM fiber is in loose condition (at room temperature), and a series of measurement data are taken when a series of known strains are applied to the fiber. The same fiber segment cell, z of Rayleigh scatter data at same position along PM fiber are extracted from the reference and measurement profiles, transformed back into the frequency domain to calculate the autocorrelation and cross-correlation shifts along the fiber. The strain coefficients of auto-correlation and cross-correlation can be calculated by a= fa/ and c = fc/, respectively. For a PM fiber here according to our experiment, the temperature coefficients of autocorrelation shifts, Ta( along the PM fiber is in the range of 0.34-0.45pm/ C, and the coefficients of autocorrelation shifts Tc( is about 14.8pm/ C; The strain coefficients of frequency shifts a( is from 0.0077 to 0.0093pm/ and c( is about 1.08pm/, respectively.. (1) 5

Thus the change of temperature and strain can be calculated by using (1) and taking reverse matrix operation: 1 T Ta( a( a. (2) Tc( c( c Thus c( a a( T c( Ta( a( Tc( a Ta( c. a( Tc( c( Ta( c Tc( (3) Experimental setting EXPERIMENTS RESULTS AND DISCUSSION The OFDR experimental set up is schematically shown in Fig. 3. The sweep rate of tunable laser source is =40nm/sec and the tuning range is about 60nm (from 1520nm to 1580nm). As the optical delay, in the auxiliary interferometer is 240ns (~48m), the sampling rate fs is about 1.2MHz, according to fs=τ. The sensing length of PM fiber is about 12-m. The Rayleigh scattered data are digitized using a 12-bit, four-channel DAQ card. The polarization controller (PC) in the front of PM fiber is used to adjust the state of polarization of input light coupled to PM fiber, so that maximum ratio of side-peak-to-center-peak amplitudes of autocorrelation can be obtained. The maximum ratio ranges from 0.2 to 0.4 with signal-to-noise ratio of ~17dB, which is adequate to calculate the autocorrelation and cross-correlation frequency shifts, as discussed in previous sections. Fig. 3: The measured frequency shifts according to applied stresses at various positions along the stressed fiber length. 6

Calibration of stress and thermal coefficients The coefficients of wavelength shifts to stress can be calibrated using OFDR system described above. A section of PM fiber of 0.3-m fiber section was stressed by using a translation stage, and the strains applied were from 0 to 6800 with strain increment 400, whereas the frequency shifts, a and a along the stressed fiber length were calculated through autocorrelation and cross-correlation calculations, as shown in Fig. 4. Fig. 4: The measured frequency shifts according to applied stresses at various positions along the stressed fiber length. (a) Cross-correlation frequency shifts at response of various stresses; (b) Autocorrelation frequency shifts at response of various stresses. In the case of cross-correlation as shown in Fig. 4(a), the frequency shifts in the function of fiber length are flat, and the strain coefficients of frequency shifts, c( along the fiber have a constant value of around 1.08pm. On the other hand, in the case of autocorrelation, the frequency shifts are zigzagged with random values along the stressed fiber, as shown in Fig. 4(b), and the coefficients ranges from 7.7 to 9.3pm/1000, which are much smaller(>100 times) than crosscorrelation calculations. This phenomenon can be explained according to the high birefringence of the PM fiber: First, as PM fiber is a high birefringence fiber, its beat lengths are in millimeter range as mentioned in previous sections. So that it is reasonable that the difference of the refractive index between the slow and fast polarization modes of PM fiber exhibits spatial variation along the fiber length; Second, during the PM fiber fabrication, residual stresses (like stress rods) are induced in the fiber by constructing the fiber out of materials with significantly different thermal expansion coefficients. As the fiber cools, these rods contract at a different coefficient than the pure silica glass around them, creating large, built-in stress. It is this stress that causes the core index of refraction to become polarization dependent, thus the states of polarization are maintained. Given this physical basis of birefringence in PM fiber, the magnitude of the difference in the refractive index between the slow and fast polarization modes is then considerably affected by thermal effect, and on the other hand, is only much weakly affected by applied strain. The temperature induced frequency shifts and the coefficients of the temperature frequency shifts along PM fiber can also be measured and calibrated when PM fiber section is heated, and the wavelength shifts is measured using OFDR system. The coefficients of cross-correlation frequency shifts, Tc( is around 1.48pm/ C, and the coefficients of autocorrelation frequency shifts, Ta( are not uniform along the fiber length, which are from 0.34 pm/ C to 0.46 pm/ C. 7

Strain ( ) These cross-correlation coefficients are about 3-4.5 times smaller than the autocorrelation ones, but are still much larger than strain autocorrelation ones. Concrete specimens tested under compression loading In this experiment, as illustrated in Fig. 5, a 120 25 20 cm 3 RC beam was equipped with PM fiber cables, which were embedded in the central part of the beam. After pouring and hardening of concrete, a section of the same PM fiber was bonded on the top and bottom surfaces of the beam. The beam was loaded step by step under a 4 point bending test, with a distance of 45 cm between the two central points, as shown in Fig.5. Results obtained at load levels of 3 kn to 15 kn are presented in Fig. 6(a). Distributed sensing profiles in tension and compression both show a trapezoid shape between the 4 points of the testing bench as expected. Strain peaks can obviously be seen at the middle of the profiles compatible with the crack location. For the load level of 12 and 15 kn, concrete tensile stresses exceed the ultimate strength and crack appears in the bottom side of the beam, which is under tension. Crack can be identified by visual inspection. However, for the load level of 3 and 6 kn, the crack is invisible at these stages and the strain peaks are observed in the strain profiles, as shown in Fig. 6(a). Fig. 5: PM fiber embedded within concrete and the 4-point load test set up using OFDR. 600 500 400 300 200 100 0-100 L1 L2 L3 L4 L5 L6 L7 L8 L9 L10 L11 L12 L13 L14 L15-200 1.0 1.5 2.0 2.5 3.0 3.5 4.0 Distance (m) Fig. 6: Compared results of the sensing systems obtained at load levels of 1 to 17kN, for the fiber in the bottom part of the beam subjected to tension (from 1.12 to 2.1m) and in the upper part of the beam under compression (from 2.6 to 3.65m). 8

Strain ( ) The specific shape of the strain peaks, as shown in Fig 7, is related to the mechanical response of the fiber-glue combination while transferring the crack opening from the concrete to the fiberglue by shear stress within the coating material. It is depending on the geometry and the mechanical properties of the materials of the fiber-glue combination [29]. It worth to notice that the width of the peak is about 15mm, which is much larger than the width of visible crack of 0.1 to 0.2mm. 500 400 520 3kN 6kN 9kN 12kN 15kN 300 200 100 63 (a) (b) 0 1.0 1.2 1.4 1.6 1.8 2.0 2.2 Distance (m) Fig. 7: (a) Compared results of the OFDR system obtained at load levels of 3, 6, 9, 12 and 15kN for the sensors in the bottom part of the beam subjected to tension. (b) Load levels of 15kN. Strain peaks are observed on the OFDR sensing profiles at locations compatible with the crack locations identified by visual inspection. Therefore, this measuring system can be used to detect and localize cracks in concrete. As the crack can be detected by OFDR at the early stress stage when the concrete is in its true elastic state and crack is invisible at that stage, OFDR sensor can be used as a powerful mean for crack detection and analysis. CONCLUSION A truly distributed optical fiber sensing system based on Optical Frequency Domain Reflectometry (OFDR) and analysis of the Rayleigh backscattering provides very high spatial resolution (mm) and high sensitivity to temperature and strain. The use of polarization maintaining fiber as the sensor provides a self-reference means to compensate temperature errors during strain measurement. Preliminary results are presented in this paper, including embedded and surface mounted fiber installations. These experimental results are very promising, since measurements performed with distributed sensing system are comparable to values obtained with conventional sensors used in civil engineering. The OFDR sensing system makes it possible to detect and localize cracks occurring on concrete surface during the mechanical loadings long before the crack become visible. REFERENCES 1. Glisic, B., and Inaudi, D. Fibre optic methods for structural health monitoring, Wiley, 2007. 9

2. Hartog, A., Distributed Fiber-Optic Sensors: Principles and applications, 251-301 (Optical Fiber Sensor Technology, 2000). 3. Barnoski, M. K., and Jensen, S. M., Fiber waveguides: A novel technique for investigating attenuation characteristics, Appl. Opt., vol. 15, pp. 2112, 1976. 4. Dakin, J. P., Distributed optical fiber sensors, in Distributed and Multiplexed Fiber Optic Sensors, II, pp. 76, Proc. SPIE, vol.1797, 1992. 5. Hartog, A. H., A distributed temperature sensor based on a liquid-core optical fiber, IEEE J. Lightwave Technol., vol. 1, pp. 498, 1983. 6. Dakin, J. P., et al., Distributed optical fiber Raman temperature sensor using a semiconductor light source and detector, Electron. Lett., vol. 21, pp. 569, 1985. 7. Stierlin, R., et al., Distributed fiber optic temperature sensor using single photon detection, Appl. Opt., vol. 26, pp. 1368, 1987. 8. Horiguchi, T., and Tateda, M., Optical fiber attenuation investigation using stimulated Brillouin scattering between a pulse and a continuous wave, Optics Letters, vol. 14(8), pp. 408, 1989. 9. Shimizu, K., Horiguchi, T., and Koyamada, Y., Coherent self-heterodyne Brillouin OTDR for measurement of Brillouin frequency shift distribution in optical fibers, IEEE J. Lightwave Technol., vol. 12, pp. 730, 1994. 10. Bao, X., Webb, D. J., and Jackson, D. A., Combined distributed temperature and strain sensor based on Brillouin loss in an optical fiber, Opt. Lett., vol. 19, pp. 141, 1994. 11. Glombitza, U., and Brinkmeyer, E., "Coherent frequency-domain reflectometry for characterization of single-mode integrated optical waveguides", J. Lightwave Technol., vol. 11, no. 8, pp. 1377, 1993. 12. Passy, R., Gisin, N., von der Weid, J. P., and Gilden, H. H., Experimental and theoretical investigations of coherent OFDR with semicoductor laser sources, J. Lightwave Technol., vol. 12, pp. 1622, 1994. 13. Passy,R., Gisin,N., Von der Weid, J. P., High-sensitivity-coherent optical frequencydomain reflectometry for characterization of fiber optic network components, IEEE Photon. Technol. Lett.vol. 7, pp.667, 1995. 14. Soller, B. J., Gifford, D. K., Wolfe, M. S., and Froggatt, M. E., High resolution optical frequency domain reflectometry for characterization of components and assemblies, Opt. Express., vol. 13, pp. 666, 2005. 15. Lee, C. C., Chiang, P. W., and Chi, S., Utilization of a dispersion-shifted fiber for simultaneous measurement of distributed strain and temperature through Brillouin frequency shift, IEEE Photon. Technol. Lett., vol. 13, pp.1094, 2001. 16. Zou, L., Bao, X., Afshar, S. V., and Chen, L., Dependence of the Brillouin frequency shift on strain and temperature in a photonic crystal fiber, Opt. Lett., vol.29, pp. 1485, 2004. 17. Bao, X., Yu, Q., and Chen, L., Simultaneous stain and temperature measurements with polarization-maintaining fibers and their error analysis by use of a distributed Brillouin loss system, Opt. Lett., vol. 29, pp. 1342, 2004. 18. Zou, W., He, Z., and Hotate, K., Demonstration of Brillouin sistributed discrimination of strain and temperature using a polarization-maintaining optical fiber, IEEE Photonics Technol. Lett., vol.22, No.8, pp. 526-528, 2010. 19. Alahbabi, M. N., Cho, Y. T., and Newson, T. P., Simultaneous temperature and strain measurement with combined spontaneous Raman and Brillouin scattering, Opt. Lett., vol. 30, pp. 1276, 2005. 10

20. Bolognini, G., Soto, M. A., and Pasquale, F. D., Fiber-optic distributed sensor based on hybrid Raman and Brillouin scattering employing multiwavelength Fabry-Pérot lasers, IEEE Photon. Technol. Lett., vol. 21, pp.1523, 2009. 21. Bolognini, G., and Soto, M. A., Optical pulse coding in hybrid distributed sensing based on Raman and Brillouin scattering employing Fabry-Pérot lasers, Opt. Express vol. 18, pp. 8459, 2010. 22. Xu, M. G., Archambault, J., Reekie, L., and Dakin, J. P., Discrimination between strain and temperature effects using dual-wavelength fiber grating sensors, Electron. Lett., vol. 30, pp. 1085, 1994. 23. Meltz, G., Morey, W. W., and Glenn, W. H., Formation of Bragg gratings in optical fibers by a transverse holographic method, Opt. Lett., vol. 14, pp. 823, 1989. 24. Hill, K. O., Malo, B., Bilodeau, F., and Johnson, D. C., Bragg gratings fabricated in monomode photosensitive optical fiber by UV exposure through a phase mask, Appl. Phys. Lett., vol. 62, pp. 1035, 1993. 25. Sudo, M., Nakai, M., Himeno, K., Suzaki, S., Wada, A., Yamauchi, R., Simultaneous measurements of temperature and strain using PANDA fiber grating, Proceedings OFS-12, 1997, pp. 170-173, Williamsburg, VA, USA. 26. Wada, D., Murayama, H., Igawa, H., Kageyama, K., Uzawa, K., and Omichi, K., Simultaneous distributed measurement of strain and temperature by polarization maintaining fiber Bragg grating based on optical frequency domain reflectometry, Smart Mater. Struct. vol. 20, 2011. 27. Froggatt, M., Gifford, D., Kreger, S., Wolfe, M., and Soller, B., Distributed strain and temperature discrimination in unaltered polarization maintaining fiber, in Optical Fiber Sensors, OSA Technical Digest (CD), paper ThC5, Optical Society of America, 2006. 28. Froggatt, M. E., Gifford, D. K.. Kreger, S., Wolfe, M., and Soller, B. J., Characterization of polarization-maintaining fiber using high-sensitivity optical-frequency-domain reflectometry, IEEE J. Lightwave Technol., vol. 24, No.11, pp. 4149, 1994. 29. Duck, G., Renaud, G., and Measures, R., The mechanical load transfer into a distributed optical fiber sensor due to a linear strain gradient: embedded and surface bonded cases, Smart Mater. Struct., 8, 175 181, 1999. 11