Distributed Discrimination of Strain and Temperature Based on Brillouin Dynamic Grating in an Optical Fiber

Transcription

1 (213) Vol. 3, No. 4: DOI: 1.17/s Review Distributed Discrimination o Strain and emperature Based on Brillouin Dynamic Grating in an Optical Fiber Kazuo HOAE 1*, Weiwen ZOU 2, Rodrigo Kendy YAMASHIA 1, and Zuyuan HE 2 1 Department o Electrical Engineering and Inormation Systems, he University o okyo, Japan 2 State Key Laboratory Advanced Optical Communication Systems and Networks, Department o Electronic Engineering, Shanghai Jiao ong University, China * Corresponding author: Kazuo HOAE hotate@sagnac.t.u-tokyo.ac.jp Abstract: his paper reviews distributed discrimination o strain and temperature by use o an optical iber based on iber optic nerve systems. he preliminary method based on multiple resonance peaks o the Brillouin gain spectrum in a specially-designed iber is irstly introduced. he complete discrimination o strain and temperature based on the Brillouin dynamic grating in a polarization maintaining iber is extensively presented. he basic principle and two experimental schemes o distributed discrimination based on iber optic nerve systems are demonstrated. he perormance o the high discriminative accuracy (.1.3 and 5 12 ) and high spatial resolution (~1 cm) with the eective measurement points o about 5 or a standard system coniguration or about 1 or a modiied one will be highly expected in real industry applications. Keywords: Fiber optic nerve systems, Brillouin optical correlation-domain analysis, Brillouin dynamic grating, distributed discrimination, strain, temperature Citation: Kazuo HOAE, Weiwen ZOU, Rodrigo Kendy YAMASHIA, and Zuyuan HE, Distributed Discrimination o Strain and emperature Based on Brillouin Dynamic Grating in an Optical Fiber,, DOI: 1.17/s Introduction he iber optic nerve system o the Brillouin optical correlation-domain analysis (BOCDA) [1 3] or relectometry (BOCDR) [4 6] takes use o the correlation-domain continuous-wave technique, which is more advantageous than the Brillouin optical time-domain analysis (BODA) [7, 8] and relectometry (BODR) [9, 1]. Fiber optic nerve systems (BOCDA and BOCDR) have been theoretically investigated and experimentally realized to provide outstanding perormances in diagnosis o ully-distributed strain or temperature disturbances with an extremely-high spatial resolution o rom centimeters [1, 2, 4, 5] to several millimeters [3, 6] along the whole iber under test (FU). Regardless o the ability o how short segment can be diagnosed, either BOCDA/R or BODA/R via measurement o a single parameter (i.e., Brillouin requency shit, BFS) is unable to distinguish the response to strain rom the response to temperature [11 13]. In this paper, we demonstrate the research trends o distributed discrimination o strain and temperature by use o an optical iber based on iber optic nerve systems. Firstly, the preliminary method based on multiple acoustic modes in a specially designed iber is introduced. Secondly, the principle and experimental demonstration o the complete discrimination o strain and temperature based on the Brillouin dynamic grating (BDG) in a polarization maintaining are presented. Finally, the Received: 16 July 213 / Revised version: 3 August 213 he Author(s) 213. his article is published with open access at Springerlink.com

2 Kazuo HOAE et al.: Distributed Discrimination o Strain and emperature Based on Brillouin Dynamic Grating in an Optical Fiber 333 development o distributed discrimination o strain and temperature is reviewed. 2. Preliminary method 2.1 Mechanism o the Brillouin-based sensing technique Brillouin scattering is a photon-phonon interaction similar to the case o Raman scattering [14]. Brillouin scattering is due to the act that annihilation o a pump photon creates a Stokes photon and a phonon simultaneously. he phonon is the vibrational modes o atoms, which is also called a propagation density wave or an acoustic phonon. he newly-generated photon suers a down-shited requency rom the pump photon due to the Doppler shit associated with the act that the generated acoustic phonons move orward. he down-shited requency is called the Brillouin requency shit (BFS, B ) given by 2ne B Va (1) where is the optical wavelength, n e the eective reractive index, and V a the eective acoustic velocity [14]. he BFS in silica optical ibers is 1 GHz 11 GHz in the 155-nm region, as shown in Fig. 1 [13]. Fiber-C Fiber-B2 Fiber-A Fiber-D 1 8 expressed by B C C (2) where C (=.4 MHz/.5 MHz/) and C (=.9 MHz/ 1.1 MHz/ ) denote the strain and temperature coeicients. he relation given in (2) is the basic mechanism o Brillouin distributed sensing techniques [1 1], however, gives a challenge how to discriminate their responses by use o a single-length optical iber. Ex,uz (a.u.) n (%) Brillouin gain [(mw) 1 ] 1 1. LP.8 1 mode.8 L mode 1 LP 1.6 L 1 L 2 mode.6 L1 ind Index-proile L mode L 4 mode.2 reractive-index proile L L L 3 L (m) r (m) 1 otal BGS L 1 1 E x, u z (a.u.) n (%) L 2 2 L L3 3 L Frequency (GHz) 1 27MHz SMF PMF-x PMF-y Gain(a.u.) Frequency (GHz) Fig. 2 Simulation results: evaluated ield distributions o the optical LP 1 mode and our longitudinal acoustic L l modes Frequency(GHz) Fig. 1 Measured (symbolic points) and Lorenzian itted (solid curves) BGS o Fiber-A, Fiber-B2, Fiber-C, and Fiber-D with dierent dopant concentrations and dierent-diameter cores [13] JL). As depicted in Fig. 2, it is clariied that the BFS change ( B ) has a good linear dependence on strain () and/or temperature () [11 13], which can be in the SMF, total BGS (solid curve) composed o our L l modes induced the BGS (dashed curves), and experimentally measured result [18] IEEE PL). 2.2 Multiple peaks in a specially-designed optical iber A preliminary method is to use a specially-designed iber with multiple Brillouin peaks that exhibit dierent requency variations in temperature and strain [15 17]. As shown in Fig. 2

3 334 [18], there are our resonance peaks o the Brillouin gain spectrum (BGS) even in a standard single mode optical iber (SMF). his is because the acoustic wavenumber is ar larger than the optical one [19] so that our acoustic modes can simultaneously exist in an SMF. he BGS in a w-shaped high-delta optical iber with the F-doped depressed inner cladding (F-HDF) has our well-separated peaks, as shown in Fig. 3 [17]. he irst-order and the ourth-order (see Fig. 4) are more suitable or the discrimination o strain and temperature since they are attributed to two separate layers o the GeO 2 -doped core and F-doped inner cladding [19 21]. Although the preliminary method based on a specially-designed iber [15 17] is a requencybased measurement technique prior to the power-based one [22], it is still ill-conditioned and physically limited due to the correlated relationship among multiple peaks. Relative gain (db) Measured BGS Simulated BGS Frequency (GHz) Fig. 3 ypical BGS measured at 25 in the loose state (dotted curve) compared with the simulated BGS (solid curve) in the F-HDF: bottom axis measured BGS; top axis simulated BGS [17] OSA OL). Resonance requency (GHz) L 4 mode L 4 mode L 3 mode L 3 mode L 2 mode L 2 mode L 1 mode L 1 mode Strain () emperature ( ) Resonance requency (GHz) Fig. 4 Resonance requencies o dierent acoustic modes as a unction o strain and temperature: solid lines least-squares linearly its to data. [17] OSA OL) Complete discrimination o strain and temperature 3.1 Principle A Panda-type polarization-maintaining iber (PMF), which has been widely used or optical iber communications [23] or iber-optic sensors [24], is composed o two B 2 O 3 -doped-silica stress-applying parts that are inserted in a pure-silica cladding and symmetrically placed beside a GeO 2 -doped-silica core [see Figs. 5 and 5]. Due to the dierence in thermal contraction between B 2 O 3 -doped silica and pure silica, two-dimensional stress is raised and stored into the core during the drawing process in iber abrication [23]. he residual stress makes the reractive index along x-axis (n x ) slightly greater than that along y-axis (n y ). Consequently, a lightwave linearly polarized along x-axis (so called slow axis) propagates slower than that along y-axis (ast axis). he dierence between n x and n y (i.e., bireringence B = n x n y ) is small (e.g., B = ~ or the iber used here) but large enough so that lightwaves linearly-polarized along either x- or y-axis can propagate through the iber while maintaining their polarization states in despite o the external disturbance. x y Fast ast-axis axis (y) (y) Pump ( ( x ) x) Beat beat Probe ( x x) ) Readout ( y ) Readout ( y) Acoustic acoustic grating x Slow axis (x) slow-axis (x) Fig. 5 Schematic diagram o the Panda-type PMF: side view, cross section, and principle o the BDG generation and readout. As schematically shown in Fig. 5, two lightwaves, one is called pump with the optical requency x and the other called probe with the

4 Kazuo HOAE et al.: Distributed Discrimination o Strain and emperature Based on Brillouin Dynamic Grating in an Optical Fiber 335 requency x, counter-propagate through the iber; the pump and the probe are linearly polarized along x-axis, or instance. When the requency dierence between the pump and probe x x equals the BFS given by (1), SBS occurs in which an acoustic wave (phonons) propagating at the velocity V a is signiicantly generated. he acoustic wave stretches or compresses the iber core longitudinally and thus modulates periodically the reractive index. Consequently, a Bragg grating moving orward at V a is ormed. Because the acoustic wave as a longitudinal wave exhibits no dependence on transverse polarization [18], the y-polarized readout light can also be strongly diracted by the acoustic grating generated by the x-polarized pump and probe, as long as the requency y o the y-polarized readout light satisies a phase matching condition that requires y to deviate rom that o the x-polarized pump ( x ) by a bireringence-determined requency deviation [25, 26] yx y x x B/ nx. (3) he residual tensile stress ( xy ) determining the Panda-type PMF s bireringence changes with the ambient temperature ( i ): B xy k( 3 2) ( ic i) (4) where ic denotes the ictive temperature (e.g., 85 ) o silica glass, 3 ( 2 ) is the thermal coeicient o B 2 O 3 -doped-silica stress-applying parts (pure-silica cladding), and k is a constant determined by the geometrical location o stress-applying parts in the iber [27]. When temperature increases ( = i 25 > ), the residual stress is released, and thus the bireringence decreases as B B (5) ic 25 where B is the intrinsic bireringence at room temperature ( i = 25 ). In contrast, when an axial strain is applied to the iber, additional stress is generated because the stress-applying parts and the cladding contract in the lateral direction dierently due to their dierent Poisson s ratios ( 3 > 2 ) [27], and the bireringence is enlarged with applied strain as ( 3 2) B B. (6) ( 3 2)( ic 25) Consequently, the bireringence-determined requency deviation yx varies linearly with a temperature increase and the applied strain, i.e., yx yx yx C C (7) where yx is the requency deviation [see (3)] at 25 and in the loose condition, and C and C are the strain coeicient and the temperature coeicient o the requency deviation, respectively. According to (4) (6), one can deduce the strain coeicient ( 3 2) C yx (8) ( 3 2)( ic 25) and temperature coeicient 1 C yx. (9) ( ic 25) According to (8) and (9), one knows that the strain coeicient C is positive while the temperature coeicient C is negative, which means that the responses o the bireringence-determined requency deviation to strain and temperature are in opposite directions. In contrast, C v and C v in (2) are both positive values because the Young s modulus o silica is nonlinearly increased by strain and temperature [13]. By solving (2) and (7) jointly, we get the strain () and temperature () as [26] 1 C C B C C C C C C.(1) yx In physics, the two phenomena/quantities, i.e., the BFS [ (2)] and the bireringence o the iber [see (7)], are inherently independent. In mathematics, C has a sign opposite to those o other three coeicients, so that the denominator ( C v C C v C ) in (1) has the signiicant value. hereore, a complete discrimination o strain and temperature based on simultaneous measurement o the two quantities can be ensured.

5 Experimental demonstration he experimental setup is shown in Fig. 6. Part A in Fig. 6 is a typical pump-probe scheme to measure the BGS (and thus BFS B ) [7], except that a set o polarization-maintaining (PM) components (polarizer, PM isolator, PM circulator, and polarization beam splitter/combiner) are introduced to ensure the pump and probe lightwaves polarized along x-axis. he light source is a 1549-nm distributed-eedback laser diode (DFB-LD1). he intensities o the pump and the probe incident to a 31-meter-length iber under test are ampliied with erbium-doped iber ampliiers to about 14 mw and about 1 mw, respectively. he requency o the probe is down-shited rom that o the pump or 1.8 GHz to 11.1 GHz with a single-sideband electro-optic modulator (SSBM) controlled by a microwave synthesizer, so the BGS is obtained as shown in Fig. 7, which depicts the BGS o the 31-meter-length FU. he synchronous detection scheme was employed by chopping the pump with the electro-optic modulator (EOM) and detecting with a lock-in ampliier (LIA1) to ensure high precision in characterization o the BFS B. A reproducible accuracy o.1 MHz was conirmed [17]. Part A: pump-pro be B GS Part A: Pump-Probe measurement BGS Measurement [Probe] [probe] Microwave, microwave,b VB PC1 Isolator isolator 5% v DFB-LD1 5% EDFA2 po SSBM EDFA3 Polarizer PM-ISO x-pol. [Pump] [pump] EOM EDFA1 v PC2 x-pol. PBS Chopping ch pin g y-pol. VOA LIA1 PD1 PC3 DFB-LD2 DAQ M BF ramp-swept comp Ra Computer LIA2 PD2 LIA2 PD2 PM-CIR1 PM-CIR2 FU FU Part B: B: Acoustic acoustic Grating grating Readout read out Fig. 6 Coniguration o the measurement system: Part A pump-probe scheme to measure the Brillouin gain spectrum and thus the Brillouin requency shit along x-axis; Part B detection o the diraction spectrum o the acoustic grating in SBS to y-polarized readout light [26] (copyright@osa OE). Brillouin gain (a.u.) Experimental Experimantal data data Lorentzian itting itting Diracted power (a.u.) Diracted power (a.u.) ~32MHz ~32 2 Expeirmental Experimantal data Gaussian itting Lorentzian itting Pump-probe requency oset (GHz) Pump-probe requency oset (GHz) Fig. 7 Measured Brillouin gain spectrum and the diraction spectrum o the dynamic acoustic grating induced by SBS to y-polarized readout light in a 31-meter-length iber at room temperature and in the loose condition: circles denote experimental data, and the solid (dashed) curve corresponds to Gaussian (Lorentzian) itting to experimental data [26] (copyright@osa OE). Part B in Fig. 6 is or precisely measuring the BDG s diraction spectrum to y-polarized readout light (and thus the bireringence-determined requency deviation yx ). he readout light is generated rom the other laser diode DFB-LD2 (central wavelength: ~ 1549 nm), whose optical requency is ramp-swept by linearly modulating its dc injection current. hrough the polarization beam splitter/combiner ater a PM circulator, the linearly polarized readout light (intensity: ~ 63 mw) is launched into the FU with its polarization along the iber s y-axis. At irst, the acoustic phonons in SBS (and thus the dynamic acoustic grating) are maximized by ixing the microwave requency to the single-sideband modulator at the BFS B obtained above. Under this condition, the readout light is strongly diracted by the acoustic grating. he diracted light is detected synchronously by using the other lock-in ampliier (LIA2) and recorded as a unction o the ramp-swept optical requency o DFB-LD2. In this way, the diraction spectrum o the acoustic grating to y-polarized readout light is obtained. For measuring the diraction spectrum, a tunable optical band-pass ilter with the about 2-GHz bandwidth is used to eliminate the leakage o the x-polarized probe and pump. he diraction spectrum o the dynamic acoustic grating established over the

6 Kazuo HOAE et al.: Distributed Discrimination o Strain and emperature Based on Brillouin Dynamic Grating in an Optical Fiber meter-length FU measured at room temperature is plotted in Fig. 7, showing that the diraction spectrum has a proile Gaussian-like rather than Lorentzian. By least-squares Gaussian itting to the experimental data in Fig. 7, we get the bireringence-determined requency deviation yx = GHz representing the iber s bireringence B = , and the ull width at the hal magnitude (FWHM) as about 32 MHz. he repeatability test result validates that the requency deviation can be measured within a standard error o yx = 4 MHz corresponding to a reproducible B = Comparably, a simply switching the SBS measurement along x-axis and y-axis [28] can only give the bireringence accuracy around 1 5, which is insuicient or characterizing strain and temperature responses o the bireringence and or strain and temperature discrimination. he BGS and the BDG s diraction spectrum were measured and used to evaluate the B and yx when dierent strains were applied to the 31-meter-length FU. As summarized in Fig. 8, the Fig. Brillouin requency shit vb (MHz) Brillouin requency shit vb (MHz) Strain () emperature ( ) 8 Measured strain and temperature coeicients: strain dependence and temperature dependence [circles denote the experimental results or the Brillouin requency shit ( B ) in let vertical axes, and triangles correspond to the bireringence-determined requency deviation ( yx ) in right vertical axes, respectively] [26] (copyright@osa OE). two quantities shit in the same way with the applied strain but in the opposite directions with the temperature. All dependence shows the excellent linearity; thus by linear itting, we get the strain coeicient and the temperature coeicient o C v = MHz/and Cv = MHz/ or the BFS B and C = MHz/and C = MHz/ or the requency deviation yx, respectively. Using our measured result yx = GHz [see Fig. 7] together with Chiang et al. s result [27] o ( 3 2 )/[( 3 2 ) ( ic 25)] = , the coeicients in the iber are deduced approximately to be C = MHz/ and C = 54.8 MHz/ according to (8) and (9), respectively. It means that the measured coeicients or the requency deviation yx match well with our theoretical estimations. Putting above strain/temperature coeicients into (1) and taking the standard errors o our measurement system ( =.1MHz and yx MHz, respectively) into account, the accuracy o the discrimination is given as high as and. his high accuracy proves the ability o completely discriminative sensing o the strain and temperature. When a set o strain and temperature (, ) was initialized as (, 4.9 ) or (939, ) in Fig. 5, or example, two requency changes ( and yx ) were measured as (5.1 MHz, 276 MHz) or (37.1 MHz, 849 MHz). According to (3), the strain and the temperature values are calculated as ( 2.3, 4.97 ) or (942.6,.197 ), respectively. hey dier rom the initialized values with errors as low as ( 2.3,.7 ) or (3.6,.2 ). his example also veriies the complete discrimination ability o the proposed method. 4. Distributed discrimination o strain and temperature 4.1 Proo-o-concept For distributed discrimination o strain and

7 338 temperature, the localized BDG generation and readout in the PMF should be irstly proved to be eective. A correlation-based continuous-wave technique o the BOCDA system was used or random access [29], and a pulse-based time-domain technique was also employed or continuous access [3]. It was ound that the generation and readout optical waves in the BOCDA system should be synchronously requency-modulated because o the dispersion properties o all our waves (see Fig. 9) [31], including writing light waves (pump and probe), reading light wave (readout), and acoustic wave (dynamic grating as well). Consequently, the local BGS and BDG diraction grating could be correctly identiied, as schematically shown in Fig. 1 [29]. (Hz) (Hz) [Probe] [Pump/Readout] op=c op op=c op y [Readout] yx 2* yx [Probe] x [Pump] op c op op c op x 2*(2) ' x ac ac=v a ac [Acoustic] (m -1 ) x x (m 1 ) ac y y Fig. 9 Dispersion properties o all pump, probe and readout waves together with the acoustic wave (Brillouin dynamic grating) [31] (copyright@osa OE). c d m 2ne (11) m where m is the modulation requency o the lasers. However, their spatial resolutions are separately determined by the BGS linewidth ( B ) and the BDG bandwidth ( yx ) as ollows [29]: c B zb (12) 2ne m c yx zd (13) 2ne m where is the modulation amplitude o the lasers. Figure 11 depicts the experimental coniguration o the distributed generation and readout o the BGS and BDG based on the correlation-based continuous-wave technique [29]. he dierence rom Fig. 6 is the introduction o the synchronous requency modulation to the lasers. he measurement range o the distributed BGS and BDG is commonly given by the neighboring correlation peaks o the BOCDA system [1, 29]. Power (dbm) ramp-sweep GHz 18GHz6GHz A microwave, B y-polarization x-polarization -58 DFB-LD1 DFB-LD2 EDFA polarizer -68 PC PM-ISO Optical requency (Hz) 3dB SSBM DFB-LD1 PMF EOM EDFA polarizer PM-CIR iber delay PC PBS/C synchronization chop computer LIA1 PD1 VOA DAQ Bias LIA2 PD2 BF PM-CIR DFB-LD2 PC EDFA polarizer Fig. 11 Experimental coniguration or the distributed generating and measuring the BGS and DGS in a PMF (copyright@osa OL). Fig. 1 Schematic o distributed measurement o DGS: beat power spectrum distribution S(z) near the correlation peak: the local dynamic grating spectrum (DGS) when the optical requency o the readout wave is not modulated or modulated synchronously to the modulation to pump light : the top plots show the measured DGS or or or when the correlation peak is localized in a heated segment [29] (copyright@osa OL). he proo-o-concept distributed measurement ability with the 1.2-m spatial resolution was veriied as demonstrated in Fig. 12. Four heated segments cascaded along the PMF sample were prepared [see Fig. 12]. he measured BGS and BDG distribution in the irst two segments are shown in Fig. 12; the characterized B and yx with opposite responses to the increased temperature are clearly illustrated in Fig. 12.

8 Kazuo HOAE et al.: Distributed Discrimination o Strain and emperature Based on Brillouin Dynamic Grating in an Optical Fiber 339 Readout Pump VB (GHz) 2.7m 2.3m 86.3m 1.1m 14.7m Hearter Hearter Hearter Hearter Probe 1.2m 1.7m 1.3m 1.4m Distributed BGS Distributed BGS yx (GHz) summarized in Figs. 13 and 13, respectively. Reerred to the characterized coeicients in Fig. 8 and the cross-sensitivity matrix in (1), the deducted distributions o strain and temperature along the iber are depicted in Figs. 13(d) and 13(e), which clearly shows the easibility o distributed discrimination o strain and temperature. Fig. 12 Distributed measurement results: prepared PMF sample, examples o 3D distribution o measured BGS and DGS rom 2 m to 9 m when the heater is turned on, and summarization o the detected BFS B (upper) or requency deviation yx (lower) and their temperature-induced changes near the heated segments [29] (copyright@osa OL). 4.2 wo-laser-based scheme o distributed discrimination By use o the experimental setup shown in Fig. 11, the two-laser-based scheme o distributed discrimination o strain and temperature was successully demonstrated with the 1-cm spatial resolution [32]. he m = MHz determines the nominal measurement range as d m = 8.35 m according to (11). For local BGS and BDG measurement, B = 1.5 GHz and D = 1 GHz correspond to the nominal spatial resolution z B = 5 cmandz D = 8 cm [see (12) and (13)], respectively. o validate the easibility o distributed discrimination o strain and temperature, we constructed an 8-m PMF sample consisting o nine (A I) cascaded iber portions o 1 cm 16 cm in length, which is illustrated in Fig. 13. he A, C, G, and I portions were loosely laid at 25.1 or reerence, while the B, D, F, and H portions were loosely inserted into a temperature-controlled water bath with the.1- accuracy. he E portion was also inserted into the water bath and glued to a set o translation stages to load strain. he measured distribution o the changes in B and yx are Readout 2.3mm 4.3m m A B C D E F G H I Probe Pump Case1: No no strain strain applied applied on E portionon E portion Case2: Strain strain applied applied on E portion) on E portion E A B C D F H G I E 5-5 A B C D F G H I Position [m] Case1: No no strain strain applied applied on E portionon E portion 25 Strain applied on E portion Case2: strain applied on E portion (d) 2 15 B D E F H (d) 1 C 55 A G II (e) 2 E A B C D F G H I Change o yx (MHz) Change o vb (MHz) Change o yx [MHz] Change o B [MHz] ( ) [ o C] () [] Fig. 13 PMF sample comprising 9 cascaded strained and heated iber portions (dashed box) with 1 cm 16 cm in length, distribution o the change in the Brillouin requency shit, the requency deviation yx, and distributed discrimination o strain (d) and temperature (e) [32] (copyright@ieee PL). 4.3 One-laser-based scheme o distributed discrimination Superior to all BDG generation/detection schemes using multiple individual laser sources [29, 3, 32, 33], the one-laser-based new scheme [31] can generate the coherent light waves to write the BDG (pump and probe waves) and to detect BDG

9 34 (readout wave) by use o a sideband-generation technique. he new scheme overcomes the relative requency luctuations among multiple waves and ensures a much higher precision than previous schemes in the measured BGS and BDG. It can also provide a higher speed in the measurement o the BGS and BDG since the time-consuming averaging process is not necessary. Figure 14 shows the experimental setup o one-laser-based generation and detection o the BDG in a PMF. A 1549-nm distributed-eedback laser diode (DFB-LD) serves as the laser source. A 4-GHz intensity modulator (IM2) driven by a radio requency synthesizer (RF2 at RF2 ) with a proper dc bias is used to generate double sidebands with the suppressed carrier (DSB-SC). he output o IM2 is launched into a iber Bragg grating (FBG) through a circulator. he light relected rom the FBG, ater urther iltered by a tunable band-pass ilter (BF1), is used as the light source or pump and probe waves to write the BDG. Ater ampliied by an erbiumdoped iber ampliier (EDFA1), the light is divided into the pump and probe waves. he requency o the probe wave is down-shited by a single-sideband modulator (SSBM) driven by another synthesizer (RF1 at RF1 ); the pump is chopped by IM1 or lock-in detection. he pump and probe waves are linearly polarized along the x-axis (slow-axis) o the PMF, which is used as the sensing medium. On the other hand, the light passing through the FBG is used as the readout wave to interrogate the BDG. he readout wave is y-polarized and launched into the PMF in the same direction as the pump wave. As shown in Fig. 15, the yx measured in this one-laser-based scheme shows high stability and accuracy (several MHz), while that obtained by the two-laser scheme suers very high luctuation (several hundreds o MHz). Compared to the previous two-laser scheme (see Fig. 11), the one-laser-based scheme demonstrated here has the similar spatial resolution (~1 cm) and measurement range (~5 m), but has several advantages including the aster measurement speed without time-consuming averaging, simpler measurement without sophisticated synchronization, and higher accuracy. IM2 DFB-LD dc b ia s RF22 1/ I/ m IM2 y iber Fiber y cabl Cable e x x RF 1 ~3.5-km iber [probe] EDFA1 dc bias Isolator isol ator BF1 1 3dB SSBM EDFA2 PC1 x-pol l. Ci rculator PMF [Pump] [pump] IM1 EDFA3 PM-CIR1 FBG G chop Chop PC2 PBS/ C EDFA4 PC3 PC 3 PM-CIR2 VO A LI A1 PD1 [Readout] [ re ] DAQ computer LI A2 PD2 BF2 z z Computer 3.5-km iber RF1 [Probe] y x PBS/C Fig. 14 Experimental setup o one-laser-based Brillouin dynamic grating generation/detection and discrimination o strain and temperature based on BOCDA (the light rom DFB-LD is intensity-modulated to generate two sidebands with suppressed carrier) [31] (copyright@osa OE). Bireringence, yx, (GHz) One-LD scheme wo-ld scheme Measurement time Fig. 15 Comparison o the BDG measurement accuracy between the one-laser scheme and two-laser scheme [31] (copyright@osa OE). Figure 16 summarizes the experimental results when the iber is heated rom 25 to 3 or/and the strain (= 2 ) is applied both at the location o 3.1 m. It is clear that the imposed strain increases both the B and yx, while the heating has a contribution to the yx opposite to that to the B. By reerring to the strain and temperature coeicients o the B and yx [see (1)], we evaluated the distribution o the temperature and strain along the

10 Kazuo HOAE et al.: Distributed Discrimination o Strain and emperature Based on Brillouin Dynamic Grating in an Optical Fiber 341 iber. he results summarized in Figs. 16 and 16(d) match well with the setting situation. Brillouin req, vb, (GHz) Brieringence, yx, (GHz) Measured strain () Measured temperature ( ) Status1 Status2 Status3 Status1 Status2 Status Status1 Status2 Status (d) Status1 Status2 Status3 Fig. 16 Measured distribution o Brillouin requency B and bireringence-determined requency deviation yx, discriminated distribution o strain and temperature (d): solid lines (Status 1): = 25, = ; squares (Status 2): = 25, = 2 ; circles (Status 3): = 3, = 2 [31] (copyright@osa OE). 4.4 Range elongation o one-laser-based scheme Due to the nature o the correlation-domain continuous-wave technique in the BOCDA system, the spatial resolutions o the distributed BGS and BDG measurement are in trade-o relation with the measurement range [see (11) (13)]. It results in a limited range o the distributed discrimination o strain and temperature. o overcome it, the temporal gating [34] or the dual modulation method [35] is introduced into the one-laser-based discriminative sensing system. he temporal gating method can be easily implemented by use o a pulse modulation o the RF2 that drives the IM2 in Fig. 14. he pulse modulation has 1/N duty ratio and 1/N repetition requency compared to that o the LD requency modulation. he pump-probe interaction occurs only at the positions where the counter propagating probe and pump pulses overlap [36]. As a result, the range given in (11) can be elongated by N times. In the experiment, the spatial resolution was set at 5 cm; the range increased by 2 times (rom 25 m to 5 m) [36]. A iber under test o a 5-m-length iber composed o two types o the PMF (A and B) is shown in Fig. 17. he iber is emerged in a x-probe input x-probe y-readout input Strain variation () PMF FU ype A,1m xy~43.5ghz ype B,2m xy~45.5ghz ype A,2m xy~43.5ghz Water bath strain Strain variation () Water bath Strain Fig. 17 Experimental results: coniguration o the iber under test, and discriminative distributed measurements o strain and temperature [35] (copyright@ieee PL) emperature variation( ) emperature variation( )

11 342 temperature controlled water bath or temperature control or pulled with a movable stage to load strain. he distributed discrimination o strain and temperature along the iber is illustrated in Figs. 17 and 17, which clearly show the easibility o the elongated measurement range. 5. Conclusions We have demonstrated the recent research progress on the discrimination o strain and temperature by use o a single-length optical iber. Ater a brie introduction o the preliminary method based on a specially-designed optical iber, the novel method based on the BDG in a PMF or complete discrimination o strain and temperature is presented. he undamental principles as well as two dierent schemes o distributed discrimination o strain and temperature are overviewed. hanks to the inherent coherence o the writing and reading waves o the BDG, the one-laser-based scheme has more eatures o the high speed and high accuracy, which are also easily extended with the assistance o the temporal gating method or the dual modulation method. he next research eorts will ocus on the improvement in the sensing speed along the entire iber [37, 38] and the simpliication o the entire iber-optic nerve system towards the real industry application in the smart materials and smart structure. Acknowledgment his work is supported by the Grant-in-Aid or Scientiic Research (S) and the Global Center o Excellence Program (G-COE) rom the Ministry o Education, Culture, Sports, Science and echnology, Japan. Weiwen Zou is partially supported by the National Natural Science Foundation o China (Grant No ), the International Cooperation Project rom the Ministry o Science and echnology o China (Grant No. 211FDA1178), Shanghai Pujiang Program rom SCSM (Grant No. 12PJ1456), and the SMC Young Star Scientist Program o Shanghai Jiao ong University. Zuyuan He is partially supported by the National Natural Science Foundation o China (Grant No ). Open Access his article is distributed under the terms o the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and source are credited. Reerences [1] K. Hotate and. Hasegawa, Measurement o Brillouin gain spectrum distribution along an optical iber using a correlation-based technique proposal, experiment and simulation, IEICE ransactions on Electronics, vol. E83-C, no. 3, pp , 2. [2] K. Hotate and M. anaka, Correlation-based continuous-wave technique or optical iber distributed strain measurement using Brillouin scattering with cm-order spatial resolution applications to smart materials, IEICE ransactions on Electronics, vol. E84-C, no. 12, pp , 21. [3] K. Y. Song, Z. He, and K. Hotate, Distributed strain measurement with millimeter-order spatial resolution based on Brillouin optical correlation domain analysis, Optics Letters, vol. 31, no. 17, pp , 26. [4] Y. Mizuno, W. Zou, Z. He, and K. Hotate, Proposal o Brillouin optical correlation-domain relectometry (BOCDR), Optics Express, vol. 16, no. 16, pp , 28. [5] Y. Mizuno, W. Zou, Z. He, and K. Hotate, Operation o Brillouin optical correlation-domain relectometry: theoretical analysis and experimental validation, Journal o Lightwave echnology, vol. 28, no. 22, pp , 21. [6] Y. Mizuno, Z. He, and K. Hotate, One-end-access high-speed distributed strain measurement with 13-mm spatial resolution based on Brillouin optical correlation-domain relectometry, IEEE Photonics echnology Letters, vol. 21, no. 7, pp , 29. [7] X. Bao, D. J. Webb, and D. A. Jackson, 32-km distributed temperature sensor based on Brillouin loss in an optical iber, Optics Letters, vol. 18, no. 18, pp , [8] M. Nikles, L. hevenaz, and P. Robert, Simple distributed iber sensor based on Brillouin gain spectrum analysis, Optics Letters, vol. 21, no. 1, pp , [9]. Kurashima,. Horiguchi, H. Izumita, S. Furukawa, and Y. Koyamada, Brillouin optical-iber time

12 Kazuo HOAE et al.: Distributed Discrimination o Strain and emperature Based on Brillouin Dynamic Grating in an Optical Fiber 343 domain relectometry, IEICE ransactions on Electronics, vol. E76-B, no. 4, pp , [1] M. N. Alahbabi, Y.. Cho, and. P. Newson, 15-km-range distributed temperature sensor based on coherent detection o spontaneous Brillouin backscatter and in-line Raman ampliication, Journal o the Optical Society o America B, vol. 22, no. 6, pp , 25. [11]. Horiguchi,. Kurashima, and M. ateda, ensile strain dependence o Brillouin requency shit in silica optical ibers, IEEE Photonics echnology Letters, vol. 1, no. 5, pp , [12]. Kurashima,. Horiguchi, and M. ateda, hermal eects o Brillouin gain spectra in silica optical ibers, IEEE Photonics echnology Letters, vol. 2, no. 1, , 199. [13] W. Zou, Z. He, and K. Hotate, Investigation o strain- and temperature-dependences o Brillouin requency shits in GeO 2 -doped optical ibers, Journal o Lightwave echnology, vol. 26, no. 13, pp , 28. [14] G. P. Agrawal, Nonlinear iber optics (3rd), New York: Academic Press, 21. [15] C. C. Lee, P. W. Chiang, and S. Chi, Utilization o a dispersion-shited iber or simultaneous measurement o distributed strain and temperature through Brillouin requency shit, IEEE Photonics echnology Letters, vol. 13, no. 1, pp , 21. [16] L. Zou, X. Bao, and L. Chen, Brillouin scattering spectrum in photonic crystal iber with a partially germanium-doped core, Optics Letters, vol. 28, no. 21, pp , 23. [17] W. Zou, Z. He, and K. Hotate, Stimulated Brillouin scattering and its dependences on strain and temperature in a high-delta optical iber with F-doped depressed inner cladding, Optics Letters, vol. 32, no. 6, pp. 6 62, 27. [18] W. Zou, Z. He, and K. Hotate, wo-dimensional inite element modal analysis o Brillouin gain spectra in optical ibers, IEEE Photonics echnology Letters, vol. 18, no. 23, pp , 26. [19] W. Zou, Z. He, and K. Hotate, Acoustic modal analysis and control in w-shaped triple-layer optical ibers with highly-germanium-doped core and F-doped inner cladding, Optics Express, vol. 16, no. 14, pp , 28. [2] W. Zou, Z. He, and K. Hotate, Experimental study o Brillouin scattering in luorine-doped single-mode optical ibers, Optics Express, vol. 16, no. 23, pp , 28. [21] W. Zou, Z. He, and K. Hotate, Dependence o Brillouin requency shit in optical ibers on draw-induced residual elastic and inelastic strains, IEEE Photonics echnology Letters, vol. 19, no. 18, pp , 27. [22]. P. Parker, M. Farhadiroushan, V. A. Handerek, and A. J. Rogers, emperature and strain dependence o the power level and requency o spontaneous Brillouin scattering in optical ibers, Optics Letters, vol. 22, no. 11, pp , [23]. Okoshi, Single-polarization single-mode optical ibers, IEEE Photonics echnology Letters, vol. 17, no. 6, pp , [24] K. Hotate, K. Makino, Z. He, M. Ishikawa, and Y. Yoshikuni, High spatial resolution iber-optic distributed lateral-stress sensing by stepwise requency modulation o a super structure grating distributed Bragg relector laser diode, IEEE Journal o Quantum Electronics, vol. 24, no. 7, pp , 26. [25] K. Y. Song, W. Zou, Z. He, and K. Hotate, All-optical dynamic grating generation based on Brillouin scattering in polarization-maintaining iber, Optics Letters, vol. 33, no. 9, pp [26] W. Zou, Z. He, and K. Hotate, Complete discrimination o strain and temperature using Brillouin requency shit and bireringence in a polarization-maintaining iber, Optics Express, vol. 17, no. 3, pp , 29. [27] K. S. Chiang, D. Wong, and P. L. Chu, Strain-induced bireringence in a highly bireringent optical iber, Electronics Letters, vol. 26, no. 17, pp , 199. [28] X. Bao, Q. Yu, and L. Chen, Simultaneous strain and temperature measurements with polarization maintaining ibers and their error analysis by use o a distributed Brillouin loss system, Optics Letters, vol. 29, no. 12, pp , 24. [29] W. Zou, Z. He, K. Y. Song, and K. Hotate, Correlation-based distributed measurement o a dynamic grating spectrum generated in stimulated Brillouin scattering in a polarization-maintaining optical iber, Optics Letters, vol. 34, no. 7, pp , 29. [3] K. Y. Song, W. Zou, Z. He, and K. Hotate, Optical time-domain measurement o Brillouin dynamic grating spectrum in a polarization maintaining iber, Optics Letters, vol. 34, no. 9, pp , 29. [31] W. Zou, Z. He, and K. Hotate, One-laser-based generation/detection o Brillouin dynamic grating and its application to distributed discrimination o strain and temperature, Optics Express, vol. 19, no. 3, pp , 211. [32] W. Zou, Z. He, and K. Hotate, Demonstration o Brillouin distributed discrimination o strain and temperature using a polarization-maintaining optical iber, IEEE Photonics echnology Letters, vol. 22, no. 8, pp , 21. [33] Y. Dong, L. Chen, and X. Bao, High-spatial-resolution time-domain simultaneous

13 344 strain and temperature sensor using Brillouin scattering and bireringence in a polarization-maintaining iber, IEEE Photonics echnology Letters, vol. 22, no. 18, pp , 21. [34] R. K. Yamashita, W. Zou, Z. He, and K. Hotate, Measurement range elongation based on temporal gating in Brillouin optical correlation domain distributed simultaneous sensing o strain and temperature, IEEE Photonics echnology Letters, vol. 24, no. 12, pp , 212. [35] W. Zou, Z. He, and K. Hotate, Enlargement o measurement range by double requency modulations in one-laser Brillouin correlation-domain distributed discrimination system, presented at Conerence on Lasers and Electro-Optics (CLEO) (Optical Society o America, 211), Baltimore, paper ChL5, Baltimore, Maryland, United States, May 1 6, 211. [36] M. Mure, M. Imai, and S. Miura, Measurement range enlargement o Brillouin optical correlation domain analysis by pulse correlation method, in Proc. 42nd Meeting on Lightwave echnology, pp , 28. [37] W. Zou, Z. He, and K. Hotate, Realization o high-speed distributed sensing based on Brillouin optical correlation domain analysis, in Conerence on Lasers and Electro-Optics (CLEO) (Optical Society o America, 29), Baltimore, Maryland, United States, Jun. 2 4, pp. 1 2, 29. [38] W. Zou, Z. He, and K. Hotate, Distributed dynamic-strain sensing based on Brillouin optical correlation domain analysis, in he 8th Paciic Rim Conerence on Lasers and Electro-Optics (CLEO/PR) (Optical Society o America, 29), Shanghai, China, Aug. 3 Sep. 3, pp. 1 2, 29.