Report
Temperature sensor based on dual fiber Bragg gratings
Authors:
Henrik Ekestam Jim Larsson
hekestam@kth.se jimlars@kth.se
Supervisor:
Michael Fokine
SA104X
Degree Project in Engineering Physics Department of Applied Physics KTH Royal Institute of Technology
May 2015
- i - Abstract
The objective of the project was to examine if it was possible to develop a low-cost temperature sensor using dual fiber Bragg gratings (FBGs). The intention was to use one FBG as a reference and let the other FBG function as the sensor. The study shows that it is possible to characterize the temperature sensitivity of each FBG and use the reference to sweep over the applicable spectrum to find the Bragg-wavelength of the sensor. This could be done measuring only the total intensity instead of intensity per wavelength with an optical spectrum analyzer, and with high accuracy. Thus, it seems feasible to construct a low cost and accurate optical temperature sensor based on dual FBGs.
Abstract in Swedish
Målet med projektet var att undersöka möjligheten att utveckla en lågkostnads-temperatursensor baserad på dubbla fiber Bragg-gitter (FBG). Syftet var att använda ett FBG som referens och låta det andra FBG:t fungera som sensor. Studien visar att det är möjligt att karaktärisera
temperaturkänsligheten hos de båda FBG:erna och använda referensen för att svepa över den relevanta delen av spektrumet för att identifiera sensorns Bragg-våglängd. Det går att åstadkomma genom att mäta bara den totala intensiteten istället för intensitet per våglängsenhet med en optisk spektrumanalysator, och det går att göra med hög noggrannhet. Således ter det sig troligt att det till låg kostnad går att konstruera en precis optisk temperatursensor baserad på dubbla fiber Bragg-gitter.
- ii - Contents
Abstract... i
Abstract in Swedish ... i
Contents ... ii
Abbreviations ... iii
List of figures ... iv
1. Introduction ... 1
1.1 Quantifying our surroundings ... 1
1.2 Optical fiber sensors ... 1
1.3 Current FBG-based optical sensors ... 1
1.4 Aim of the project ... 2
2. Theory... 3
2.1 Bragg’s law ... 3
2.2 Wavelength shift of Bragg maxima ... 3
2.3 Production of fiber Bragg gratings ... 4
2.4 Optical spectrum analyzer ... 5
2.5 FBG reflection and transmission ... 5
2.6 FBG design parameters ... 6
3. Experimental Setup ... 7
3.1 Illustration of the system ... 7
3.2 Description of the system ... 8
3.3 Expected response of sensor FBG reflectance ... 8
3.4 Molding of reference FBG ... 10
3.5 Calibration ... 11
3.6 Controlling the system ... 11
3.7 Usage of Optical spectrum analyzer ... 12
4. Results ... 13
4.1 FBG Spectra ... 13
4.2 Calibration curves: reference FBG ... 15
4.3 Calibration curve: sensor FBG ... 16
4.4 Reflection of Sensor FBG during scan ... 17
5. Discussion ... 18
5.1 The individual FBGs ... 18
5.2 The combined system ... 18
6. Conclusion ... 20
6.1 Suggestions for further research ... 20
7. References ... 21
Appendix A: Hardware used in the system ... 22
Appendix B: H-Bridge circuit diagram ... 23
Appendix C: Code ... 24
C.1 Arduino code ... 24
C.2 Python processing code ... 25
- iii - Abbreviations
OFS Optical Fiber Sensor FBG Fiber Bragg Grating OSA Optical Spectrum Analyzer LED Light Emitting Diode FWHM Full Width – Half Maximum MOSFET Metal–Oxide–Semiconductor
Field-Effect Transistor
- iv - List of figures
Figure Description
1 Interferometric fabrication of a FBG. ... 4
2 FBG reflection and transmission. ... 5
3 Bragg-wavelength, reflectance and FWHM for a reflection peak. ... 6
4 Illustration of the proposed sensor system. ... 7
5 Expected response of sensor FBG reflectance. ... 9
6 Reference FBG contained in a molded tin-lead metal alloy. ... 10
7 Schematic of the mold seen from two directions. ... 10
8 Arrangement of fiber and cast at the molding process. ... 10
9 Example of temperature cycle used for calibrating the FBGs. ... 11
10 Reflection from reference FBG for different temperatures. ... 13
11 Transmission of reference FBG. ... 13
12 Reflection sensor FBG for different temperatures. ... 14
13 Bragg-wavelength of reference FBG versus temperature. ... 15
14 Linear regression for the different regions in figure 13. ... 15
15 Wavelength of reference FBG versus temperature during two different cooling-heating. .... 16
16 Bragg-wavelength versus temperature for sensor FBG. ... 16
17 Reflection spectra of sensor FBG for different temperatures of the reference FBG. ... 17
18 Reflected intensity of the sensor FBG versus the temperature of the reference. ... 17
19 Circuit diagram of the constructed H-Bridge. ... 23
20 Explanation of symbol used for MOSFET. ... 23
- 1 - 1. Introduction 1.1 Quantifying our surroundings
An eternal question within the subject of physics is: How do we measure and quantify attributes of objects in our surroundings? The classical way of achieving such measurements is to construct an instrument that is put in contact with the object with a property to be measured, and study how this contact between the object and the instrument influences the instrument. For example, an ordinary thermometer where the length of a liquid column changes in response to the heat exchange introduced by the contact with the measured object.
In modern environments, one often desires to use the results of the measurement together with a computer to perform computations and analyze the acquired data. A common way to achieve this is to construct an electronic sensor where the quantity that is to be measured influences either the voltage, current or resistance of the sensor in a known way. Thus, using wide-spread methods to get a
computer to collect data on the voltage, current or resistance, one has acquired a method to collect and digitally interpret information about the desired quantity.
1.2 Optical fiber sensors
A common need in industrial environments is the requirement to measure some quantity, for example temperature, voltage, physical strain or refractive index. Typically, an electronic sensor is deployed to conduct the measurement, but as an alternative sensors based on fiber optics may be used, with several advantages:
- Fiber cables consist of non-conductive material and are therefore generally not affected by electromagnetic interference.
- The fiber sensor is a passive device without a need for electricity - Fiber cables are chemically inert.
- The cable has low weight.
- The fiber is small in size and can thus easily be embedded in structures.
- Maintenance requirements are typically small.
Thus, there exist situations in which the usage of optical fiber sensors (OFS) would be favorable to electronic sensors and even environments where the only viable possibility is to deploy optical fiber sensors.
1.3 Current FBG-based optical sensors
The deployment of currently available optical fiber sensors based on fiber Bragg gratings (FBG) is cost prohibitive. The main expense in the setup stems from the need to identify the change in reflected wavelength by the sensor. The cost-introducing element with FBG-sensors is that the change in wave length typically is very small, characteristically with a magnitude measured in picometers (10-12 m) [1, 2]. Thus, the optical spectrum analyzer (OSA) deployed at the end of the fiber needs to be advanced enough to distinguish such changes, with the effect that the needed equipment is very expensive. If an FBG-sensor without the need for an OSA could be engineered, the resulting cost- reduction would introduce opportunities to perform cheaper and more accurate measurements of a desired quantity.
- 2 - 1.4 Aim of the project
The aim of this project is to examine if it is possible to by simple means construct a low cost temperature sensor based on dual FBGs without the need for an optical spectrum analyzer. The project builds upon earlier research showing that the temperature sensitivity of a FBG can be increased up to five-fold by introducing a tin-lead mold around the FBG [3]. The objective is to examine a system of two FBGs, one reference and one sensor, embed the reference in a metal alloy and determine if such a system could be used to construct a low cost optical temperature sensor.
- 3 - 2. Theory 2.1 Bragg’s law
Bragg-diffraction occurs when electromagnetic radiation is reflected by a material with a periodical structure. Bragg's law applies as a condition for constructive interference
2𝑑 sin 𝜃 = 𝑛𝜆 ( 1 )
where d is the periodicity of the interference pattern, θ is the scattering angle, n is an integer and λ is the wavelength of the radiation.
For a FBG, it approximately holds that 𝜃 = 90° ⇒ sin(𝜃) = 1. For the first diffraction maxima, n=1, equation 1 then gives
λ=2d. ( 2 )
Equation 1, however, is only meaningful for vacuum, and needs to be adapted for glass fiber. The distance d in equation 2 is thus replaced by the effective refractive index of the FBG times the spacing of the grating, Λ:
𝜆𝑏= 2𝑛𝑒𝑓𝑓Λ ( 3 )
Equation 3 is the condition for the wavelength of the first Bragg maxima.
2.2 Wavelength shift of Bragg maxima
The Bragg wavelength is as given by equation 3 above dependent on the refractive index and the periodicity of the FBG. Both of these values can be displaced by longitudinal forces as well as by changes in temperature. The relative sensitivity of Bragg wavelength as strain and temperature change is given by equation 4:
Δ𝜆𝑏
𝜆𝑏 = (1 − 𝜌𝑒)𝜖𝑧+ (𝛼 + 𝜂)Δ𝑇 ( 4 )
Where 𝛼 is the thermal expansion coefficient of the fiber, 𝜂 is the thermo-optic coefficient representing the shift in refractive index with changes in temperature, 𝜌𝑒 is the photo-elastic
coefficient representing shift in refractive index due to strain and 𝜖𝑧 is the longitudinal strain over the grating. [1, 2]
Typical values for germanium doped optical fiber cores are [1, 2]:
𝜌𝑒= 0.22, 𝛼 = 0.55 × 10−6 ℃−1, 𝜂 = 8.6 × 10−6 ℃−1
- 4 -
Thus, theoretical values of the sensitivity of the Bragg wavelength in the range of 1550 nm due to temperature shifts and strain are, as given by equation 4:
Δ𝜆𝐵
Δ𝑇 = 14.18 𝑝𝑚 ℃⁄ ( 5 )
And:
Δ𝜆𝐵
Δ𝜖𝑧 = 1.2 𝑝𝑚 ℃⁄ ( 6 )
2.3 Production of fiber Bragg gratings
Fiber based on doped silica contains defects in the glass structure [1, 2]. When the fiber is irradiated with certain wavelengths of UV-light these defects absorb some of the radiation resulting in a permanent increase in the refractive index locally in the fiber, depending on the amount of absorbed radiation. If the irradiation along with the fiber occurs with periodically varying intensity, an FBG is produced with the characteristics of equation 3.
There exists three main methods for the manufacturing of a FBG: Interferometric, phase mask and point by point [2]. The interferometric method is based on the principle of constructive and destructive interference between light waves. A laser beam is passed through a prism and is thus divided into two beams, as shown in figure 1 below. The two beams are then reflected by two mirrors towards the fiber at an angle 𝜑 such that an interference pattern forms when the beams meet within the fiber. The parts of the fiber illuminated by the interference pattern will see an increase in refractive index, with the highest increase located at the maxima of the interference pattern. By adjusting the position and angle of the mirrors and the rest of the system, different interference patterns can be achieved.
Figure 1. Interferometric fabrication of a FBG.
- 5 -
Another method of manufacturing a FBG is by applying a phase mask just over the fiber and then illuminate the mask with a laser beam. The light from the laser passes through the phase mask and undergoes constructive and destructive interference depending on the pattern of the phase mask. This method is sometimes combined with the interferometric method, replacing the prism with a phase mask, with the advantage that the movable mirrors enables the possibility of adjusting the properties of the FBG without changing the phase mask [1].
The third method of manufacturing a FBG is point by point illumination [2]. A focused laser is moved along the fiber and the intensity is varied to create the desired pattern. The main advantage of this method is that it enables the possibility of manufacturing a FBG with very complicated changes in refractive index. A disadvantage is that the method relies heavily on the precision of the translational movement of the laser.
2.4 Optical spectrum analyzer
An optical spectrum analyzer (OSA) is a sensor that measures the wavelength distribution of incoming radiation and the respective intensity [1]. An OSA typically works by separating the incoming light by wavelength. To be able to distinguish small differences in wavelength the OSA needs to contain high quality instruments making it hard to produce inexpensively.
2.5 FBG reflection and transmission
Figure 2 below shows the expected reflection and transmission curves respectively, when the emission from a uniform light source is sent into a FBG [2]. The in signal is transmitted unchanged, except for a narrow band of wavelengths centered around the Bragg-wavelength, 𝜆𝐵. The wavelengths missing from the transmission curve is instead reflected back into the fiber and will thereafter travel in the opposite direction compared to the light source. Thereby a separation of wavelengths have been formed were a signal within a desired band of wavelengths can be created.
Figure 2. A) Uniform in signal. B) Signal transmitted by FBG. C) Signal reflected by FBG
A) B) C)
- 6 - 2.6 FBG design parameters
Several design parameters need to be considered when writing FBGs: Bragg-wavelength, Reflectance and the Full Width Half Maximum (FWHM) value, see figure 3 below. The Bragg-wavelength, 𝜆𝐵, determines where the center of the reflected radiation seen in figure 2C is situated. The reflectance determines how much of the radiation is reflected at that point, which may be a value lower than 100 %. The FWHM-value measures how broad the reflection peak is at the height of half-maximum intensity.
Figure 3. Bragg-wavelength, reflectance and FWHM for a reflection peak.
- 7 -
3. Experimental Setup
The system was built using two FBGs, one was used as a reference and one as a sensor. The
temperature of the reference was controlled by the system, while the temperature of the sensor FBG was controlled by the object to be measured. The following sections describes the setup and the individual parts.
3.1 Illustration of the system
Figure 4 below contains a schematic diagram of the proposed system. A legend for the numbers contained within figure 4 is provided after the figure.
Figure 4. Illustration of the proposed sensor system.
1. Light source 2. Fiber coupler 3. Tuning FBG 4. Peltier element
5. Arduino microcontroller 6. H-bridge
7. Sensing FBG 8. Light reference
9. Reflection intensity sensor 10. 3 dB coupler
11. Power source
12. Reference temperature sensor
- 8 - 3.2 Description of the system
This section contains a general description of the system. For a description of the individual parts, refer to the coming sections.
Figure 4 in section 3.1 above contains a schematic of the proposed system. At (1) light is produced and directed towards the optical fiber. At (2) the light is passed into the fiber in which the first FBG is fabricated. The reference FBG (3) is an ordinary FBG embedded in metal, refer to section 3.4, and surrounded by Peltier-elements (4). The Peltier-elements gives the opportunity to regulate the temperature of the reference FBG, which as given by equation 4 leads to a shift in the Bragg-
wavelength and thereby a change in the transmitted signal as described in section 2.7. The metal that the FBG is encapsulated in increases the temperature sensitivity about five-fold compared to a naked FBG [3].
The transmitted signal from the reference FBG is then fed into a 3 dB coupler at (10). Every time a signal passes through this junction, 50 % of the intensity is directed in the two forward directions respectively. That means that a signal with half the intensity of the reference FBG transmission will be sent into the sensor FBG (7) and the other half will be directed into the light reference direction at (8). Unlike the reference FBG, the sensor FBG is not embedded into metal. The sensor FBG will be in contact with the object which temperature is to be measured, and will therefore have the same
temperature as the object.
The reflected signal from the sensor FBG is then once again fed through the 3 dB coupler (10) and once more the intensity is halved in each direction. The signal is then transported into the reflection intensity sensor (9) which measures the summed intensity over all applicable wavelengths.
The system, and specifically the tuning of the Peltier-elements, is handled by an Arduino micro- controller (5). Since the Arduino is not able to provide more current than about 40 mA [4], an external power source (11) is used to drive the Peltier-elements. Since the Arduino according to specifications has to control the current to the Peltiers, and also be able to reverse polarity, a standard H-bridge (6) built on power-MOSFET transistors is placed between the power source and the Peltier-elements and connected to the Arduino. Temperature data on the reference FBG is provided to the Arduino by a K- Type thermocouple connected to an amplifier (12) and then to the Arduino.
3.3 Expected response of sensor FBG reflectance
As described in the previous section, light is fed into the fiber and directed towards the reference FBG, which will reflect some wavelengths around its Bragg-wavelength while the rest of the signal is transmitted practically unchanged. If light with a spectrum like in figure 5A below enters the system, then the expected transmission of the reference FBG is as shown in figure 5B, with a prominent dip around the Bragg-wavelength of the reference. When the dip from the transmission of the reference FBG does not overlap with the Bragg wavelength of the sensor FBG, a signal like in figure 5C will be fed into the reflection intensity sensor. If the dip however does overlap with the Bragg-wavelength of the sensor, a signal like in figure 5D will instead be sent into the reflection intensity sensor. Since the dip from the reference FBG can be tuned by the current into the Peltier-elements, which controls the temperature of the reference FBG, the system is able to scan through the spectrum and search for the Bragg-wavelength of the sensor FBG. Since the Bragg-wavelength of the sensor FBG is controlled by
- 9 -
the external temperature that is to be measured, the Bragg-wavelength found by scanning the spectrum can be used together with a calibration curve to calculate the external temperature.
Figure 5. A) Light emitted by light source. B) Light transmitted by the reference FBG. C) Light reflected by sensor FBG when the dip from the reference FBG does not overlap with the expected reflected wavelength from the sensor FBG. D) Light reflected by sensor FBG when the dip from the reference FBG does overlap with the expected reflected wavelength.
A) B)
C) )
A)
D)
- 10 - 3.4 Molding of reference FBG
The reference FBG was embedded in a tin-lead metal alloy. By molding an alloy around the fiber, a better connection between the Peltier-elements and the reference FBG may be achieved as well as increased thermal conductivity compared to the free air that would otherwise surround the fiber. It has also been shown [3] that putting a mold around the FBG may lead to up to a fivefold increase in thermal sensitivity of the grating compared to a naked FBG, due to the thermal expansion of the alloy.
The molding process began with the fiber containing the FBG being fixated with magnets to minimize the possibility of translational movement during the process. A molding cast were produced by bending thin wolfram (W) foil into the desired shape for the mold shown in figure 7 above. The molding cast were aligned with the fiber and likewise fixated, see figure 8. A soldering iron were used to melt the tin-lead alloy which were dripped into the cast until full. Since some of the metal solidified during the dripping process the mold were reheated using the soldering iron to ensure a homogenous structure of the mold. With the metal still melt, a small tube with inner diameter the size of the thermocouple were inserted into the mold to function as the temperature sensor connector shown in figure 6 and 7 above. The metal were then allowed to solidify before the molding cast was carefully removed.
Figure 7. Schematic of the mold seen from two directions.
Figure 6. Reference FBG contained in a molded tin- lead metal alloy.
Figure 8. Arrangement of fiber and cast at the molding process.
- 11 - 3.5 Calibration
The theoretical predictions in equation 5 and 6 above make a good approximation to the respective sensitivities for fibers not contained in metal alloy, but due to slight differences in the FBGs due to fabrication, and the reference FBG being retained in metal, more precise values are obtained through calibration of the grating. Calibration was done by putting the grating though a known cycle of changes in the quantity that is to be measured, which for this project means that the gratings was tested for their sensitivity during shifts in temperature, and record the corresponding changes in Bragg-wavelength. An example of such a temperature cycle is presented in figure 9 below, where the changes in temperature were introduced by the Arduino and the Peltier-elements.
3.6 Controlling the system
The system was controlled by an Arduino micro controller connected to a computer for data analysis.
The Arduino regulated the temperature of the reference FBG by controlling the polarity of the current sent through the Peltier-elements. Since the Arduino are not able to supply more than 40 mA of current [4], the system had to be connected to an external power supply providing the needed current.
The power supply and the Arduino was connected via a standard power-MOSFET H-bridge, thereby introducing the ability to the Arduino to control if any current flows to the Peltier-elements at all, and also the polarity of the current, thus controlling if the reference FBG was heated or cooled. The amount of current and applied voltage was controlled manually on the power supply and only changed in between experiments.
The temperature of the reference FBG was measured by a standard K-Type thermocouple, inserted in the temperature sensor connector shown in figure 6 and 7 above. The thermocouple was connected to an amplifier and then to the Arduino, which were programmed to make a temperature measurement by averaging 300 values for increased precision. The thermocouple combined with the H-Bridge provides the Arduino with the ability to regulate the temperature of the reference FBG as desired. The regulation was done by setting an upper and a lower limit on desired temperature. During the heating part of the cycle, the Arduino provided the Peltier-elements with maximum current set on the power supply at all times, and did so until the upper limit on desired temperature was reached. The same was done during the cooling process, but with reversed polarity. When changing between cooling and
Figure 9. Example of temperature cycle used for calibrating the FBGs.
- 12 -
heating, the Arduino turned off power completely for five seconds in order to not short-circuit the H- Bridge and not introduce unnecessary mechanical stress on the Peltier-elements. An example of the resulting temperature cycle is presented in figure 9 above.
3.7 Usage of Optical spectrum analyzer
An OSA were used to characterize the reference and sensor FBG respectively, before introducing them to the system. Ten measurements were averaged to reduce the amount of dynamic noise while the static noise background were measured and subtracted.
- 13 - Figure 11. Transmission of reference FBG.
4. Results 4.1 FBG Spectra
Table 1 and figure 10 below presents acquired spectra, Bragg-wavelengths and FWHM-values for the reference FBG at three different temperatures. Figure 11 shows a spectrum of the transmission of the reference FBG. Figure 12 presents spectra of the reflectance from the sensor FBG at three different temperatures and table 2 presents corresponding data of Bragg-Wavelengths and FWHM-values.
Table 1. Bragg-wavelength and Full Width Half Maximum values for reference FBG.
Temperature [°C] 𝜆𝑏 [nm] FWHM [nm]
5.09 1539.37 0.333
32.82 1541.04 0.333
105.03 1542.21 0.333
Figure 10. Reflection from reference FBG for different temperatures.
- 14 -
Table 2. Bragg-wavelength and Full Width Half Maximum values for sensor FBG.
Temperature [°C] 𝜆𝑏 [nm] FWHM [nm]
9.17 1541.55 1.17
44.33 1541.89 1.17
80.67 1542.22 1.17
Figure 12. Reflection sensor FBG for different temperatures.
- 15 -
Figure 13. Bragg-wavelength of reference FBG versus temperature in the interval 5 °C to 105 °C during four cycles of heating and cooling.
4.2 Calibration curves: reference FBG
Figure 13 below presents Bragg-wavelength of the reference FBG as a function of temperature over four heating-cooling cycles. Figure 14 contains linear regressions for four different regions in figure 13. Figure 15 presents Bragg-wavelength versus temperature for two different heating-cooling cycles with the same upper and lower temperature limit.
Figure 14. Linear regression for the different regions in figure 13.
A) Linear regression for the interval 12 °C to 53 °C during warming. B) Linear regression for the interval 55 °C to 105 °C during warming. C) Linear regression for the interval 8 °C to 70 °C during cooling. D) Linear regression for the interval 80 °C to 102 °C during cooling.
A) B)
C) D)
- 16 - 4.3 Calibration curve: sensor FBG
Figure 16 presents Bragg-wavelength versus temperature for the sensor FBG.
Figure 16. Bragg-wavelength versus temperature for sensor FBG.
Figure 15. Wavelength of reference FBG versus temperature during two different cooling-heating cycles over an equal number of periods.
A) B)
- 17 - 4.4 Reflection of Sensor FBG during scan
Figure 17 contains a series of spectra showing how the reflection of the sensor FBG changes when the sensor FBG’s temperature is held constant and the reference wavelength is swept past the sensor wavelength. Figure 18 shows how the integrated intensity over all applicable wavelengths changes as a function of reference temperature, when the temperature of the sensor is held constant.
Figure 18. Reflected intensity of the sensor FBG versus the temperature of the reference.
Figure 17. Reflection spectra of sensor FBG held at constant room temperature for different temperatures of the reference FBG.
A) Tref = 18.93 °C, B) Tref = 46.54 °C, C) Tref = 52.93 °C, D) Tref = 56.63 °C, E) Tref = 57.95 °C F) Tref = 59.78 °C
A) B) C)
F) E)
D)
- 18 - 5. Discussion 5.1 The individual FBGs
Figure 11 shows that the transmission curve of the reference FBG contains a dip corresponding to the Bragg-wavelength, as expected. Figure 10 and 12 shows that in accordance with theory, the Bragg- wavelength is displaced by temperature shifts, with increased temperatures leading to increased Bragg-wavelengths. For the sensor FBG the displacement is approximately linear, as seen in table 2 and figure 12, as well as in the calibration curve for the sensor, figure 16. A linear approximation for the temperature sensitivity is as given by figure 16 16.62 pm/°C, in good agreement with the
theoretical value of about 14 given in equation 5. The existence of “plateaus” in figure 16 indicates that displacements is within the same order of magnitude as the resolution of the OSA.
However, for the reference FBG the displacement is not linear. Figure 13 shows that the heating- cooling process suffer from hysteresis, that is, the achieved values are different for the same temperature at different parts of the cycle. This can, at least partly, be explained by delays in the temperature measurement. Figure 13 and 14 depicts a behavior of the reference FBG where the temperature sensitivity falls dramatically after 55 °C during heating and 70 °C during cooling. At lower temperatures, the sensitivities are 41.7 pm/°C and 30.8 pm/°C respectively, which is much higher than for the naked sensor FBG and shows that the metal alloy gives the desired effect. At temperatures higher than 55 °C and 70 °C respectively, the sensitivity drops, during cooling it drops to 16.1 pm/°C, same as for the naked sensor fiber. During the heating process the sensitivity drops to 7.1 pm/°C, less than half the non-metal-contained fiber. This indicates that the metal loses its grip onto the fiber and expands without stretching the fiber, and thereby possibly isolates the fiber from the heat-source, leading to lower temperature sensitivities than even the naked sensor FBG. All these effects need to be accounted for when calibrating the reference FBG.
Figure 15 shows that the amount of hysteresis is dependent on how long the period of the heating- cooling process is, for the same temperature limits. A slower process leads to less hysteresis while at the same time limiting how many measurements the system can make over a given time period. The hysteresis can thus be handled by slowing down the process or by trying to account for the
phenomenon at time of calibration.
5.2 The combined system
Figure 17 and 18 depicts the behavior of the combined system when the transmission of the reference FBG is used as in signal for the sensor FBG. Figure 17 shows that the system behaves as predicted in section 3.4, that is, when the reflection spectra of the reference spectra does not overlap with the reflection from the sensor, the sensor’s output is unaffected by the existence of the reference.
However, when the reference’s Bragg-wavelength is swept past the Bragg-wavelength of the sensor the existence of the reference shows as a dip in the output from the sensor.
Figure 18 shows how the total intensity of the reflection from the sensor changes when the reference is swept over the spectrum. As hoped, the intensity shows a clear decline around specific values for the temperature of the reference. Combined with the calibration curves for both the reference and the sensor, it should be possible to determine the temperature of the sensor by studying the change in total reflected intensity as the reference temperature is changed.
- 19 -
The bottom of the parabola in figure 18 is quite broad, which is directly dependent on the FWHM- value of the sensor FBG. A smaller value should lead to a smaller region with reduced intensity.
Another reason for the difficulty of identifying the very lowest point in figure 18 is the shape of the sensor FBG reflection, seen in figure 12. Two to three small and closely situated peaks exists in the spectrum, which will lead to complex behavior at the bottom of the well in the curve, as observed. A less wide and more distinct reflection peak for the sensor might thus increase the sensitivity of the measurements.
- 20 - 6. Conclusion
The objective of this project was to examine if it was feasible to use dual fiber Bragg gratings (FBG) in order to build an inexpensive and accurate temperature sensor. This project has shown that each individual FBG can be characterized with regards to temperature sensitivity. It has also been shown that embedding a FBG in a tin-lead alloy increases the temperature sensitivity. A system was built using one FBG as a reference and the other as the actual temperature sensor. By sweeping the reference over the applicable spectrum, it was possible to measure the total reflected intensity and identify a distinct decline in intensity when the wavelength of the reference FBG passed the wavelength of the sensor FBG. Thus, using the measurements of total intensity combined with calibration data for the FBGs it should be possible to identify the temperature of the sensor FBG, and thereby determine the temperature of the object to be measured.
6.1 Suggestions for further research
In order to build an independent measurement system based on the method presented in this report a method for producing and guiding light into fiber that is inexpensive and compact is needed. Possible candidates for such a system includes the use of an ordinary light emitting diode to produce light and a lens to direct the light into the fiber.
Further research should also include other ways to control the reference signal apart from temperature, since the temperature cycling is rather slow as shown in figure 15, and thereby limits the amount of measurements that can be taken in a given time interval.
Different Bragg-wavelengths should also be examined, for example visible red light at around 650 nm. This combined with a light emitting diode (LED) as the light source could prove to be very inexpensive and easy to measure, since red light LEDs are very widespread and commonly used.
- 21 - 7. References
[1] Current Trends in Short- and Long-period Fiber Gratings; Werneck, M. Allil, R, Ribeiro, B., Nazaré F.;2013; ISBN 978-953-51-1131-3
[2] Optical Fiber Sensor Technology: Advanced Applications – Bragg Gratings and Distributed sensors; Grattan K.T.V., Meggit B.T.; Kluwer Academic Publishers; 2000; ISBN 0-7923-7946-2 [3] Fiber Bragg Gratings in Temperature and Strain Sensors; Häggmark I.; 2014; KTH
[4] Arduino Uno: data sheet
- 22 -
Appendix A: Hardware used in the system Arduino Uno
Adafruit Thermocouple Amplifier, product id: 1778 Adafruit thermocouple type K, product id: 270 Peltier-element QuickCool QC-17-1.0-3.9M
Own-built H-Bridge based on IOR power-MOSFET IRLU8721PbF
- 23 -
Appendix B: H-Bridge circuit diagram
Figure 19. Circuit diagram of the constructed H-Bridge.
Figure 20. Explanation of symbol used for MOSFET in figure 19.
- 24 - Appendix C: Code
C.1 Arduino code
const int inPin = 0; // temperature pin float starttemp;
const int pluspin = 8; //heating pin const int minpin = 5; //cooling pin
void setup() {
Serial.begin(9600);
starttemp = 44; //center of temp interval pinMode(pluspin,OUTPUT);
digitalWrite(pluspin,LOW);
pinMode(minpin,OUTPUT);
digitalWrite(minpin,LOW);
delay(1000);
}
void loop() {
digitalWrite(pluspin,LOW);
digitalWrite(minpin,LOW);
float temp = getTemp();
float lim = 36; //difference from //starttemp to upper //and lower temp limit
while (temp - starttemp < lim) {
digitalWrite(minpin,LOW);
digitalWrite(pluspin,HIGH);
temp = getTemp();
}
digitalWrite(pluspin,LOW);
getTemp();
getTemp();
getTemp();
getTemp();
getTemp();
while (starttemp - temp < lim) {
digitalWrite(pluspin,LOW);
digitalWrite(minpin,HIGH);
temp = getTemp();
}
digitalWrite(minpin,LOW);
getTemp();
getTemp();
getTemp();
getTemp();
getTemp();
}
float getTemp() {
float value = getRawTemp(inPin);
Serial.print(value); Serial.print(" > ");
float volts = (value *5.0) /1024.0;
Serial.print(volts); Serial.print(" > ");
float celsius = (volts -1.25) / 0.005;
Serial.print(celsius);
Serial.println(" degrees Celsius, ");
delay(59); // the process should take exactly //one second
return celsius;
}
float getRawTemp(int Pin) {
int times = 300; //number of averages float tempArray[times];
for (int i = 0; i < times; i++) {
tempArray[i] = analogRead(inPin);
delay(3);
}
float sum = 0;
for (int i = 0; i < times; i++) {
sum += tempArray[i];
}
float mean = sum/times;
return mean; }
- 25 - C.2 Python processing code
import matplotlib matplotlib.use('TkAgg')
from matplotlib.backends.backend_tkagg import FigureCanvasTkAgg import matplotlib.pyplot as plt
import tkinter as Tk
import serial
import serial.tools.list_ports from time import sleep import time
import sys
class Grafik:
def __init__(self):
self.port = list(serial.tools.list_ports.comports())[0][0] #Select the Arduino self.ser = serial.Serial(self.port, 9600,timeout=0.1) #Connect to the Arduino
self.root = Tk.Tk()
self.root.wm_title("Thermometer")
self.root.protocol('WM_DELETE_WINDOW',self.end) try:
self.root.iconbitmap('Thermometer.ico') except:
pass
self.f = plt.figure(figsize=(6,6), dpi=100) #Create the figure self.t = 0
self.x = [] #List with number of measurements self.y = [] #List with measurements
self.tlista = [] #List with times each measurement were taken self.starttid = time.time()+1.0
self.a = self.f.add_subplot(111) self.line, = self.a.plot(self.x,self.y) self.a.set_title('\n')
self.a.set_xlabel('Time (s)')
self.a.set_ylabel('Temperature (°C)')
# a tk.DrawingArea
self.canvas = FigureCanvasTkAgg(self.f, master=self.root) self.canvas.show()
self.canvas.get_tk_widget().pack(side=Tk.TOP, fill=Tk.BOTH, expand=1)
- 26 -
self.canvas._tkcanvas.pack(side=Tk.TOP, fill=Tk.BOTH, expand=1) self.root.after(3500,self.plot)
if not 'idlelib' in sys.modules:
Tk.mainloop()
def plot(self):
self.append()
self.line.set_data(self.tlista,self.y) ax = self.canvas.figure.axes[0]
ax.set_xlim(min(self.tlista), max(self.tlista)) ax.set_ylim(min(self.y)-1, max(self.y)+1)
self.a.set_title('T = ' + str(self.y[len(self.y)-1]) + ' °C, t = '+ str(int(self.tlista[len(self.x)-1])) + ' s\n')
self.canvas.draw()
self.root.after(100,self.plot)
def append(self):
tot = self.ser.readlines() #Read data from the Arduino for i in tot:
temp = float(i.decode('ascii').split('>')[2].strip().split(' ')[0]) self.t += 1
self.x.append(self.t) self.y.append(temp)
self.tlista.append(time.time()-self.starttid) print(i.decode('ascii').strip())
def export(self,namn=None):
#Exports the data as a txt file and the plot as png if namn == None:
namn = time.strftime('Thermometer\Thermometer__%Y-%m-%d_%H.%M.%S.png') print(namn)
tnamn = namn.replace('png','txt') print(tnamn)
plt.savefig(str(namn))
with open(tnamn,'w',encoding='utf8') as tfil:
tfil.write('Measurement #\tTime\tTemperature\n') for i in range(len(self.y)):
tfil.write('\t'.join((str(self.x[i]),str(self.tlista[i]),str(self.y[i])))+'\n')
def end(self):
self.export() self.root.destroy()
if __name__ == '__main__':
g = Grafik()
