ARTICLE IN PRESS
Optik 119 (2008) 535–539 www.elsevier.de/ijleo
A novel ﬁber Bragg grating high-temperature sensor$ Yage Zhana,, Shaolin Xuea, Qinyu Yanga, Shiqing Xiangb, Hong Heb, Rude Zhub a
Department of Applied Physics, College of Science, Donghua University, Shanghai 201620, China Shanghai Institute of Optics and Fine Mechanics, Chinese Academy of Sciences, Shanghai 201800, China
Received 25 May 2006; received in revised form 21 February 2007; accepted 25 February 2007
Abstract A novel ﬁber Bragg grating (FBG) sensor for the measurement of high temperature is proposed and experimentally demonstrated. The interrogation system of the sensor system is simple, low cost but effective. The sensor head is comprised of one FBG and two metal rods. The lengths of the rods are different from each other. The coefﬁcients of thermal expansion of the rods are also different from each other. The FBG will be strained by the sensor head when the temperature to be measured changes. The temperature is measured basis of the wavelength shifts of the FBG induced by strain. A dynamic range of 0–800 1C and a resolution of 1 1C have been obtained by the sensor system. The experiment results agree with theoretical analyses. r 2007 Elsevier GmbH. All rights reserved. Keywords: Sensor; Fiber Bragg grating; High temperature
1. Introduction Reliable high-temperature sensors are important and indispensable in some ﬁelds, such as in some structure health monitoring and material processing, electrical transformer, petroleum pipeline and so on [1,2]. Traditional electrical high-temperature sensors have some disadvantages, including low reliability, large temperature ﬂuctuation and latent danger of ﬁre accident. Optical ﬁber Bragg grating (FBG) sensors have numerous advantages over traditional electrical sensors, such as immunity to electromagnetic interference, higher stability and sensitivity, more easiness of multiplex, being competent for application in harsh environments, ‘‘smart structures’’ and on-site measurements [3,4]. FBG $ Project supported by Science and Technology Committee of Shanghai (Grant No. 011661081). Corresponding author. E-mail address: [email protected]
0030-4026/$ - see front matter r 2007 Elsevier GmbH. All rights reserved. doi:10.1016/j.ijleo.2007.02.010
sensors are the most appropriate sensors for monitoring applications in the ﬁelds mentioned above. But common FBG sensors cannot used directly as high-temperatures sensor because they will be decayed when its temperature higher than 200 1C and will be destroyed when its temperature higher than 350 1C [5,6]. Until now, only a very few kind of technologies on FBG high-temperature measurement have been researched [7,8]. Brambilla et al. have researched the high-temperature measurement characteristics of FBGs that with special dopants (such as Sn and/or Na2O). They discovered that these FBGs exhibit unusual oscillations in reﬂectivity . These methods are not suited for high-temperature measurement. This paper proposes a novel kind of FBG hightemperature sensor. The novel sensor is very suited for high-temperature object, especially for high-temperature object in usual temperature atmosphere. The experimental results and the characteristic of the sensor system are also described. The sensor is based on a novel FBG
ARTICLE IN PRESS 536
Y. Zhan et al. / Optik 119 (2008) 535–539
sensor head and a ﬁber long period grating (LPG) as a linear edge ﬁlter for interrogation. The novel sensor head has been designed, prepared and used in hightemperature measurement experiments successfully. A dynamic range of 0–800 1C and a resolution of 1 1C have been experimentally achieved. Experimental results agree with theoretical analyses.
to the other end. For briefness, the length change of L is given by DL1 ¼
l 1j DT 1j a1j ðj ¼ 1; 2; . . . ; nÞ,
l 2j DT 2j a2j ðj ¼ 1; 2; . . . ; nÞ,
n X j¼1
2. Theoretical analyses
DL ¼ DL1 DL2 ,
2.1. Principle of sensor head Sensor head is very crucial in sensor system. But common FBG cannot be used as high-temperature sensor head directly. We have designed a novel high-temperature FBG sensor head. The sensor head is mainly comprised of a FBG and two metal rods, as shown in Fig. 1. The two metal rods have different length and different coefﬁcient of thermal-expansion (CTE). The lengths of the two metal rods are L1 and L2, respectively. The CTEs of the two metal rods are a1 and a2, respectively. The rods are ﬁxed into one adiabatic plate. In order that there is not transverse thermal radiation, the two metal rods have been protected by adiabatic cylinder1 and adiabatic cylinder2, respectively. The left ends of the two metal rods connect two adiabatic rods, respectively. The FBG is pre-strained and glued to the end surface of the adiabatic rods on points A and B. The FBG is protected by the adiabatic cylinder3 in order that the FBG is not be modulated by the environmental temperature and the thermal radiation of the adiabatic plate. The sensing ends (see also in Fig. 1) touch the object whose temperature to be measured. When temperature to be measured is changed, the two metal rods will have different elongation, which will make L change (the distance between the two adiabatic rods) and the FBG be strained. The temperature is measured basis of wavelength shifts of the FBG. The adiabatic cylinders are effective. The transverse thermal radiation of the metal rods is negligible. When the rods are in heat balance, the temperature of each metal rod reduces linearly from whose sensing end
A. Adiabatic rod1
Adiabatic cylinder3 Metal rod2(L2 α 2) B . Adiabatic plate Adiabatic rod2 L
Object Adiabatic cylinder1 Metal rod1(L1 α 1) Adiabatic cylinder2
Fig. 1. Schematic diagram of sensor head structure.
where DL1 and DL2 are the elongations of the two metal rods, respectively. DL is the length change of L, namely the elongation of FBG section of ﬁber. lij, DTij and aij (i ¼ 1, 2) are the length, average temperature and average CTE of the jth subsection of the metal rod. The corresponding wavelength shift DlB of the FBG is expressed by [3,10] DlB ¼ lB 1 pe ¼ lB 1 pe DL1 DL2 DL ¼ lB 1 pe , ð4Þ L L where pe ¼ ð1=ÞðDneff =neff Þ ¼ ðn2eff =2Þ½p12 nðp11 þ p12 Þ is the effective photo-elastic coefﬁcient of the glass ﬁber with Possion ratio n. P11 and P12 are the photo-elastic coefﬁcients of ﬁber. neff is the effective refractive index of the guide mode in the ﬁber. For a typical fused silica ﬁber, pe ¼ 0.22. The two metal rods of the sensor head are made from an H62 brass rod and a 45# carbon steel rod, respectively. The CTEs of the two metal rods are a1 and a2, respectively. a1 and a2 have been measured and determined numerically by a1 ¼ 15:78250 þ 0:02796 T 2:4085 105 T 2 106 , a2 ¼ 10:99550 þ 0:00994 T 5:5421 105 T 2 106 . ð5Þ In the same temperature range, a1 is larger than a2. The lengths of the two rods are L1 and L2, respectively. The curve of the wavelength change of the FBG have been theoretically simulated with suppositions of both L1 ¼ 20 cm, L2 ¼ 18 cm and L1 ¼ 18 cm, L2 ¼ 20 cm in the range of 0–500 1C. The simulation results are shown in Fig. 2. Similarly, the simulation results in the range of and 0–1000 1C are shown in Fig. 3. If L1 ¼ 20 cm and L2 ¼ 18 cm, the peak wavelength of the FBG shifts almost linearly with temperature in the range of 0–800 1C. When the temperature ascends from 0 to 800 1C, it shifts 6.80 nm. Generally, 6.8 nm wavelength shift will not induce the FBG worse or broken. The sensitivity of the sensor system is enhanced when the metal rod with larger CTE is longer than the metal rod with smaller CTE, which can be conﬁrmed by that
ARTICLE IN PRESS Y. Zhan et al. / Optik 119 (2008) 535–539
Fig. 2. The temperature–wavelength response of the sensor FBG in the range of 0–500 1C.
1550 1555 Wavelength (nm)
Fig. 4. Schematic spectra of the FBG temperature sensor and the LPG employed as edge ﬁlter. Coupler
Interrogation system BBS: Broadband source; IMG: Index matched gel
Fig. 5. Schematic diagram of the sensor system.
Fig. 3. The temperature–wavelength response of the sensor FBG in the range of 0–1000 1C.
the slope of curve (a) is larger than the slope of curve (b) in Figs. 2 and 3. So all the experiments are implemented in the conditions of L1 ¼ 20 cm and L2 ¼ 18 cm.
2.2. Interrogation principle Wavelength interrogation technology is very important for FBG sensor system. In our high-temperature sensor system, an LPG is used as a linear response edge ﬁlter to convert wavelength into intensity encoded information for interrogation. The principle of using an LPG to interrogate an FBG temperature sensor is based on the temperature related optical intensity measurement. Fig. 4 shows the schematic reﬂection and transmission spectra of the FBG and the LPG used
in the experiments. The LPG is used as a linear response edge ﬁlter because the useful spectrum region of the LPG is shown to be nearly linear over a sufﬁciently wide range . If an interrogation system is arranged according as the way shown in Fig. 8, light from the broadband source (BBS) will be modulated by the LPG and then illuminates the FBG via a 2 2 coupler. After being LPG modulated, the light has a section of available linear spectrum. The reﬂected light from the FBG is detected by the photo-detectors (PD) and will change with the Bragg wavelength shift of FBG. Therefore, the ﬁltering mechanism of the LPG yields a linear relationship between the wavelength shift of FBG and the PD detected light intensity.
3. Experiments and results Fig. 5 shows the schematic diagram of the experimental setup. The sensor head is made in accordance with Fig. 1. A brass (H62) rod is used as the longer metal rod with a larger CTE and a carbon steel (45#) rod is used as the shorter metal rod with smaller CTE.
ARTICLE IN PRESS 538
Y. Zhan et al. / Optik 119 (2008) 535–539
Light from the BBS illuminates the FBG through a coupler. The lengths of the brass rod and the carbon steel rod are 20 and 18 cm, respectively. The CTE of the two metal rods are a1 and a2, respectively, same as function (5). The Bragg wavelength of the FBG is 1549.96 nm after it is glued on the adiabatic rods. The reﬂected light from the FBG is detected by the wavelength interrogation system through the same coupler. The other end of the coupler immerses in index matching gel (IMG). In experiments, the temperature of the sensing end is controlled by a stove. The temperature of the stove can be modulated by step of 0.1 1C in the range of 0–500 1C with an accuracy of 0.2 1C. Three series of experiments have been down.
The curve in Fig. 7 is accordant with curve (a) in Fig. 2 in the range of 0–500 1C. It can be deduced that the sensor can measure the temperature in the range of 0–800 1C. The FBG had 4.25 nm wavelength shifts when the temperature changed from 0 to 500 1C. The theoretical value is 4.31 nm. The relative error is 1.4%. All the experimental data in Fig. 7 can be ﬁtted by a slight second-order polynomial function. The function can be expressed as
3.1. Primary experiments
l ¼ 1549:2677 þ 0:0097 T.
l ¼ 1549:9006 þ 0:0060 T þ ð5:2369 106 Þ T 2 , (6) where l is the wavelength of FBG and T the temperature to be measured. There is a linear response when the temperature to be measured is higher than 100 1C. The function can be expressed as (7) 4
To prove elementary performance of the sensor system, ﬁrst series of experiment has been done. An optical spectrum analyzer (OSA) has been used for interrogating the wavelength of the sensor FBG. The experimental setup is shown in Fig. 6. Limited by the characteristics of present stove, experiments are implemented in the temperature range of 0–500 1C. The experimental result is shown in Fig. 7. Coupler
Fig. 6. Schematic diagram of the experimental setup (I).
1555 Data of the experiment Fitted curve of the data
The error of the slope value is 1.2330 10 and the standard deviation of the ﬁt is 0.0458. Proﬁted from the good demodulation system, a resolution of 1 1C is obtained.
3.2. Farther experiments In order to make the whole sensor simple, low cost and effective, an LPG has been used as a linear ﬁlter for interrogation . The experimental setup is shown in Fig. 8. Second and third series of experiments have been done with the LPG interrogation technology. In second series of experiments, the output power of the BBS was set at three different work points for the three sub-series of experiments to explore the stability and repeatability of the sensor system. The results are shown in Fig. 9. In third series of experiments, the usual interferences (such as heat convection in surroundings) are attached, to explore the sensor system’s ability of anti-interference. The results are shown in Fig. 10. From Fig. 9, it is obvious that the sensor system has good stability and repeatability. From Fig. 10, it is certiﬁed that the sensor system has better anti-interference ability when the temperature to be measured is
200 300 400 Temperature (°C)
Fig. 7. The results of the experiment.
DAC & Signal processing
Fig. 8. Schematic diagram of the experimental setup (II).
ARTICLE IN PRESS Y. Zhan et al. / Optik 119 (2008) 535–539
be obtained by the sensor system. The experiment results well agree with the theoretical analysis.
Relative intensity PD output
Power3 (all with the same BBS)
17.0 16.8 16.6 16.4 16.2 0
150 200 Temperature/°C
Fig. 9. Experimental results of the stability and repeatability of the FBG high-temperature sensor. 17.6
Relative intensity PD output
Result with external distubance (power4) Result without external distubance (power5)
17.4 17.2 17.0 16.8 16.6 16.4 0
100 150 Temperature/°C
Fig. 10. Experimental results of the anti-interference ability of the FBG high-temperature sensor.
higher than 100 1C. The sensor system has weaker antiinterference ability when the temperature is lower than 100 1C. So our work group is making efforts to improve the anti-interference ability of the sensor system.
4. Conclusion In conclusion, a novel FBG sensor structure for measurement of high temperature is studied. The sensor is simply, low cost and easy to be implemented. A dynamic range of 0–800 1C and a resolution of 1 1C can
 J. Leng, A. Asundi, Structural health monitoring of smart composite materials by using EFPI and FBG sensors, Sensor Actuators A 103 (2003) 330–340.  Y. Zhao, Y. Liao, Compensation technology for a novel reﬂex optical ﬁber temperature sensor used under offshore oil well, Opt. Commun. 215 (2003) 11–16.  A. Kersey, M. Davis, H. Patrick, M. Leblanc, K. Koo, Fiber grating sensors, J. Lightwave Technol. 15 (1997) 1442–1463.  B. Lee, Review of present status of optical ﬁber sensors, Opt. Fiber Technol. 9 (2003) 57–59.  Y. Shen, S. Pal, J. Mandal, T. Sun, K. Grattan, S. Wade, S. Collins, G. Baxter, B. Dussardier, G. Monnom, Investigation of the photosensitivity, temperature sustainability and ﬂuorescence characteristics of several Erdoped photosensitive ﬁbers, Opt. Commun. 237 (2004) 301–308.  S. Baker, H. Rourke, V. Baker, D. Goodchild, Thermal decay of ﬁber Bragg gratings written in boron and germanium co-doped silica ﬁber, J. Lightwave Technol. 15 (1997) 1470–1477.  J. Canning, K. Sommer, M. Englund, Fibre gratings for high temperature sensor applications, Meas. Sci. Technol. 12 (2001) 824–828.  T. Morse, Y. He, F. Luo, An optical ﬁber sensor for the measurement of elevated temperatures, IEICE Trans. Electron. E83-C (3) (2000) 298–302.  G. Brambilla, V. Pruneri, L. Reekie, C. Contardi, D. Milanese, M. Ferraris, Bragg gratings in ternary SiO2: SnO2:Na2O optical glass ﬁbers, Opti. Lett. 25 (16) (2000) 1153–1155.  Y. Rao, In-ﬁbre Bragg grating sensors, Meas. Sci. Technol. 8 (1997) 355–375.  Y. Zhan, H. Cai, R. Qu, S. Xiang, Z. Fang, X. Wang, Fiber Bragg grating temperature sensor for multiplexed measurement with high resolution, Opt. Eng. 43 (2004) 2358–2361. Yage Zhan was born in 1977, in China. She received the B.S. degree in Physics from the HeNan University in 2000, and the M.S. degree in Optical Engineering form Shanghai Institute of Optics and Fine Mechanics, CAS, in 2002, respectively. She is now working for her Ph.D. Her research interests include ﬁber grating sensors and demodulation technologies for ﬁber grating sensors.