structural health monitoring based on fibre optics...
TRANSCRIPT
STRUCTURAL HEALTH MONITORING BASED ON FIBRE OPTICS
CURVATURE SENSOR
PNG WEN HAO
A project report submitted in partial fulfilment of the
requirements for the award of the degree of
Bachelor (Hons.) of Physics
Lee Kong Chian Faculty of Engineering and Science
Universiti Tunku Abdul Rahman
May 2016
ii
DECLARATION
I hereby declare that this project report is based on my original work except for
citations and quotations which have been duly acknowledged. I also declare that it has
not been previously and concurrently submitted for any other degree or award at
UTAR or other institutions.
Signature : _________________________
Name : _________________________
ID No. : _________________________
Date : _________________________
iii
APPROVAL FOR SUBMISSION
I certify that this project report entitled STRUCTURAL HEALTH
MONITORING BASED ON FIBRE OPTICS CURVATURE SENSOR
was prepared by PNG WEN HAO has met the required standard for submission in
partial fulfilment of the requirements for the award of Bachelor of Physics (Hons.) at
Universiti Tunku Abdul Rahman.
Approved by,
Signature : _________________________
Supervisor : Mr. Lin Horng Sheng
Date : _________________________
iv
The copyright of this report belongs to the author under the terms of the
copyright Act 1987 as qualified by Intellectual Property Policy of University Tunku
Abdul Rahman. Due acknowledgement shall always be made of the use of any
material contained in, or derived from, this report.
© 2016, Png Wen Hao. All right reserved.
v
ACKNOWLEDGEMENTS
I would like to thank everyone who had contributed to the successful completion of
this project. I would like to express my gratitude to my research supervisor, Mr. Lin
Horng Sheng for his invaluable advice, guidance and his enormous patience
throughout the development of the research.
vi
STRUCTURAL HEALTH MONITORING BASED ON FIBRE OPTICS
CURVATURE SENSOR
ABSTRACT
A packaged fibre-based in-line Mach-Zehnder Interferometer sensor (Hereinafter,
referred as packaged MZI sensors) with highest sensitivity up to is -3.27𝜇Wm-1 is
fabricated. Packaging is introduced to the sensor to protect the sensor under harsh
condition of the real sensing environment. The packaged MZI sensor is capable to
detect minimum curvature of 0.25km-1 and maximum curvature radius up to 4km,
which is highly sensitive in monitoring the structural health. The packaged sensor is
characterised based on the imposed curvature at various wavelengths. Operating
wavelengths of 1310nm, 1490nm and 1550nm are used in curvature calibration to
observe the sensitivity at different wavelength. More importantly, the effect of the
packaging thickness to the curvature sensitivity is justified in this project.
vii
TABLE OF CONTENTS
DECLARATION ii
APPROVAL FOR SUBMISSION iii
ACKNOWLEDGEMENTS v
ABSTRACT vi
TABLE OF CONTENTS vii
LIST OF TABLES ix
LIST OF FIGURES x
LIST OF SYMBOLS / ABBREVIATIONS xiii
LIST OF APPENDICES xiv
CHAPTER
1 INTRODUCTION 1
Structural Health Monitoring 1
Optical Fibre Sensors in SHM 2
Installation of Optical Fibre Sensor 2
Sensing principles of Optical Fibre Sensors 2
Aims and Objectives 3
Thesis Outline 3
2 LITERATURE REVIEW 5
Types of Optical Fibre Sensor in SHM Application 5
2.1.1 Long Period Fibre Grating Sensor 5
2.1.2 Fibre Bragg Grating Sensor 7
2.1.3 Mach-Zehnder Interferometer Sensor 9
Comparison of Optical Fibre Sensors 13
viii
2.2.1 Cross Sensitivity in Sensors 13
2.2.2 Fragility of Fabricated Segment in Sensors 14
Sensor System Design 15
2.3.1 Fabrication Criteria 15
2.3.2 Packaging of Sensors 16
3 METHODOLOGY 18
Fabrication of Sensor 19
Packaging of Sensor 19
Curvature Calibration of Sensor 20
3.3.1 Calibration Based on Various Wavelengths 23
Calibration Based on Different Packaging Thicknesses 23
Back-tracing of the Curvature 24
4 RESULTS AND DISCUSSION 25
Curvature Sensitivity Based on Various Wavelength 25
4.1.1 Wavelength Dependent Property of MZI Sensor 28
4.1.2 Polarization Dependent Property of MZI Sensor 29
Curvature Sensitivity Based on Different Packaging
Thicknesses 31
4.2.1 Comparison of the Pristine Thickness, Thickness
A and B. 33
4.2.2 Disparity in Trend of Optical Power Slope 36
Back-tracing of Curvature 37
5 CONCLUSION AND RECOMMENDATIONS 39
Conclusion 39
Future Works 39
APPENDICES 46
ix
LIST OF TABLES
TABLE TITLE PAGE
2.1 Comparison Table of Optical Fibre Sensor. 13
4.1 Comparison Table of Sensor in Three Different
Thicknesses. 35
4.2 Optimal Detectable Radius and Curvature 35
x
LIST OF FIGURES
FIGURE TITLE PAGE
2.1 Schematic Diagram of an In-fibre LPG Sensor 6
2.2 Coupling of a Fundamental Mode to a Cladding
Modes in LPG 6
2.3 Transmission Spectra of LPG of Length; (a) ~1cm
and (b) ~3cm (Yin, et al., 2008) 7
2.4 Schematic Diagram of an In-fibre FBG Sensor 8
2.5 Wavelength-selective Reflectional Filter of FBG 8
2.6 Schematic Diagram of an In-line Tapered MZI
Sensor 9
2.7 Splitting of Fundamental Core Mode in First Taper
Site 10
2.8 Transmission Spectra of MZIs under Different
Interferometer Lengths: (a) 20 mm, (b) 30 mm, (c)
36 mm and (d) 40 mm (Li, et al., 2011) 12
2.9 Spectral Shift Due to Temperature Changes (Raji,
et al., 2016) 14
2.10 Schematic Diagram of the Fabricated MZI Sensor 15
2.11 Linearity Region of Spectrum Profile for Different
Interferometric Length, L. (Li, et al., 2011) 16
3.1 Outline of the Methodology 18
3.2 (a) First (b) and Second Fibre Tapers of MZI
Sensor. 19
3.3 MZI Sensor with Interferometric Length of 5cm 19
3.4 Schematic Diagram of Uniform Width MZI
Package Design. 20
xi
3.5 Photo of a Packaged MZI Sensor 20
3.6 (a) Schematic Diagram of Placement of MZI
Sensor on Steel Bar; (b) Experimental Setup of MZI
Sensor on Steel Bar. 21
3.7 Experimental Setup of Bending Test 21
3.8 Schematic Diagram of Steel Bar Curvature 22
3.9 Trigonometric Diagram of Steel Bar Curvature 22
3.10 Configuration of Sensor with (a) Thickness A and
(b) Thickness B 24
4.1 Raw Result of Curvature Calibration of Sensor with
Pristine Thickness at 1550nm. ( × indicates the
loading slope and △ indicates the unloading slope.) 25
4.2 Model of the Linearity and Non-linearity Region in
the Responding Spectra 26
4.3 Curvature Calibration of Sensor with Pristine
Thickness in Responding to Wavelength 1310,
1490, and 1550nm. 27
4.4 Comparison of Flexural Moduli with Theoretical
Value within the Operating Region. 27
4.5 Disparity in Power Trend of MZI Sensor in
Responding to Variation of Wavelength 28
4.6 Disparity in Sensitivities of MZI Sensor in
Reponding to Variation of Wavelength 29
4.7 Power Changes and Spectral Shifting in
Responding to Variation of Polarization States. 30
4.8 Offset Disparity in Responding to the Variation of
Polarization State 30
4.9 Distance from the Sensor in (a) Thickness A and (b)
Thickness B 31
4.10 The Curvature Segment of the Sensor Packaging 32
4.11 Curvature Calibration in Responding to
Wavelength 1310, 1490, and 1550nm for Sensor
with (a)Pristine Thickness, (b) Thickness A and (c)
Thickness B 34
xii
4.12 Comparison of Flexural Moduli with Theoretical
Value within the Operating Region for Sensor with
(a) Pristine Thickness, (b) Thickness A and (c)
Thickness B 34
4.13 Curvature Calibration of Sensor with Various
Thicknesses at Wavelength 1310nm 34
4.14 Curvature Calibration of Sensor with Various
Thicknesses at Wavelength 1490nm 35
4.15 Curvature Calibration of Sensor with Various
Thicknesses at Wavelength 1550nm 35
4.16 Loading Spectra in Arbitrary State 1 and 2 37
4.17 Correlation between the Back-traced Curvature and
the Characterised Curvature (with error bar of 14%)
38
5.1 Schematic Diagram of MZI Multiplexing
Technique using TDM 40
5.2 Schematic Diagram of MZI Multiplexing
Technique using SDM 40
xiii
LIST OF SYMBOLS / ABBREVIATIONS
SHM Structural Health Monitoring
MZI Mach-Zehnder Interferometer
LPG Long Period Grating
FBG Fibre Bragg Grating
WDM Wavelength Division Multiplexing
TDM Time Division Multiplexing
LVDT Linear Variable Differential Transformer
xiv
LIST OF APPENDICES
APPENDIX TITLE PAGE
A Derivation of the Strain Equation 46
B The Back-tracing and Characterised Data. 48
1
1 INTRODUCTION
Structural Health Monitoring
Over the decades, material science has been established to ensures that most of the
engineering structures such as civil infrastructures, shipping, aero and aerospace
structures meet the minimum safety standard. By introducing in various types of in-
service or pre-service non-destructive tests (NDT), structural failures are managed to
be detected.
Structural Health Monitoring (SHM) in aids of the NDT methods are widely
implemented in damage detection of the large scale civil structure. Maintainability of
civil structures such as bridges, buildings, dams, vessels and platforms can be
improved by implementing various types of sensors into the system. Traditional
vibration sensors such as magneto-electric, piezoelectric, and current sensor are
commonly used in SHM. The waves generated within the structure will be reflected
when meet the discontinuities, therefore structural health information such as damage
and crack can be localised. However the damping property of most materials causes
attenuation in the waves during propagation, which may lead to signal weakening
effect and incur error in underrating the severe damage. (Yang & Hu, 2008)
2
Optical Fibre Sensors in SHM
Optical fibre sensors preponderate the conventional electronics sensors in many ways.
Firstly, optical fibre sensors are nonconductive to electromagnetic interference which
operate independently and safely under conductive environment. Besides,
predominant sensitivity of optical fibre sensor in detecting a tiny scale deformation
(strain and bending) of the material way before it’s fracture happens, overrides others
in term of failure inspection. The pre-failure detection of optical fibre sensor enables
residents to take evacuation action immediately, before the catastrophe happens.
Installation of Optical Fibre Sensor
In general, sensing system of the optical fibre sensor can be classified into two:
localised and distributed sensing system. Localised sensing system is a single point
sensor which functions to detect and feedback the impact in form of analysable
information. Whereas, a distributed fibre sensor is possibly made up of multiple
combination arrays of localised sensors, which can cater to multiple points and
parametric sensing. Multiplexing such as Wavelength Division Multiplexing and Time
Division Multiplexing is required in this sensing system to distinguish the output of
data arrays (Yin, et al., 2008). In the real application of SHM, distributed sensing
system is commonly used for multiple points detection within a large scale civil
structures, by installing the sensor arrays in a distributive manner over the loaded part
of building such as struts, beams and girders.
Sensing principles of Optical Fibre Sensors
The most commonly implemented sensing principles of optical fibre sensor are
interferometry, grating and scattering. Interferometric sensing principle splits the
coherent source into two signals and recouples them in order to retrieves information
form the interference signal. Interferometric sensors such as Mach-Zehnder,
Michelson, Fabry-Perot and Sagnac, are applying the interferometric sensing principle.
3
The physical detection including strain, temperature, pressure and refractive index can
be done majorly through investigating the measurand such as power fluctuation and
phase shifts (Lee, et al., 2012).
Grating sensing principle is commonly implemented in long period fibre grating
and short period fibre grating, by varying the period grating knowingly. Application
of grating as a wavelength-selective filter or mode-dependent splitter, is governed by
periodical modulation of the refractive index along the fibre. Parameters like
attenuation and shifting in transmission and reflection spectrum can be used to detect
the structural deformations such as mechanical strain and temperature expansion.
Scattering sensing principle such as Brillouin and Rayleigh scatterings are used by
small portion of research in sensing the structural strain (Thévenaz, 2010; Mizuno, et
al., 2015). However, this kind of sensing principle is more widely used in pipeline
leakage detection (Daniele, et al., 2007).
Aims and Objectives
The project aims to design a packaged MZI sensor and to characterise its curvature
sensitivity based on various wavelengths. Besides, the effect of packaging thickness
to its sensitivity also will be characterised in this project.
Thesis Outline
Chapter 1 gives a brief introduction to SHM, types of optical fibre sensor, sensing
methods and structures. Chapter 2 reviews several types of optical fibre sensors in
SHM, and the respective sensing principle. A suitable sensor was designated in this
section; the relevant sensors system designs is discussed. Chapter 3 focuses on the
methodologies of this project, which includes fabrications, packaging and calibrations
of MZI sensor. Chapter 4 investigates the power intensity changes in responding to the
curvature of bending results from increase loading, to characterise the sensor based on
various wavelengths and packaging thickness. Results of respective calibrations is
4
further analysed and discussed in the section thereof. Chapter 5 concludes all the
completed works and discusses some recommendations for the future topics.
5
2 LITERATURE REVIEW
Types of Optical Fibre Sensor in SHM Application
Structural health can be monitored through fibre-based flexural strain and curvature
sensor. Optical fibre sensors have different sensing principles such as interferometry,
grating and scattering. Long period fibre grating (Sharma, et al., 2012) , Fibre Bragg
grating (Mokhtar, et al., 2011) and tapered MZI (Wen, et al., 2014) are the sensor that
dominantly researched. In the part hereof, the literature review based on the structural
system and sensing principle of the three sensors will be elaborated in detailed.
2.1.1 Long Period Fibre Grating Sensor
Long-period fibre grating (LPG) was first proposed by Vengsarkar et al. in 1996, is a
grating device designed by photo-induced periodic modulations of the refractive index
along the core of a single-mode fibre (SMF). The grating period normally ranges from
100 m to 1mm. LPG can be fabricated using UV irradiation, CO2 laser, and infrared
femtosecond laser pulses to creates a permanent refractive index interference pattern
in the optical fibre (Kondo, et al., 1999; Zhu, et al., 2007).
Integrating the effect of grating period and variation in refractive index, phases
matched coupling is induced from the fundamental core mode (LP01) to the higher
cladding modes (LP0m, m=1,2,3,4…) at a specific resonance wavelength. Thereby it
generates a serial attenuation dips in the transmission spectrum (Puneet & Himani,
6
2015). The phases matching condition between the core mode and cladding modes for
LPG is govern by eq. 2.1.
𝜆𝑟𝑒𝑠 = (𝑛𝑐𝑜,𝑒𝑓𝑓01 − 𝑛𝑐𝑙,𝑒𝑓𝑓
𝑚 )Λ
(2.1)
Where 𝜆𝑟𝑒𝑠 is the resonance wavelength, 𝑛𝑐𝑜,𝑒𝑓𝑓01 is the effective refractive index
of the core mode and 𝑛𝑐𝑙,𝑒𝑓𝑓𝑚 is the effective index of the mth cladding mode. Λ is the
grating period.
Figure 2.1: Schematic Diagram of an In-fibre LPG Sensor
Figure 2.2: Coupling of a Fundamental Mode to a Cladding Modes in LPG
Attenuation in LPG is high due to the propagation of light within the lossy
cladding ambience. The changing of core and cladding properties, more essentially the
refractive indices and the grating period, Λ will affects the attenuation spectrum of
LPG (Yin, et al., 2008).
7
As the coupled mode propagates along the cladding, it is highly sensitive to the
change of the ambient refractive index, which in turn varies the propagating constant
of cladding mode that causes attenuation dips and phase shift. Besides, elongation of
the grating period will introduce the similar effect onto the sensor (the effect can be
observed in Figure 1.3; (a) and (b)) Therefore, variations such as temperature,
curvature, strain and external refractive indices can be observed by detecting the
changes of the two parameters.
(a) (b)
Figure 2.3: Transmission Spectra of LPG of Length; (a) ~1 cm and (b) ~3 cm (Yin, et
al., 2008)
2.1.2 Fibre Bragg Grating Sensor
Similar to LPG, Fibre Bragg Grating (FBG) is designed by varying the refractive index
periodically along the core of a single-mode fibre. In general, the grating period is
smaller than LPG by more than two orders of magnitude. Fabrication methods such as
UV irradiation and electron beam interference lithography are applied to induce a
permanent refractive-index change in FBG (Qiu, et al., 1999).
Unlike LPG, in FBG, only fundamental core mode survives in the transmission
signal, where no cladding modes survive within the cladding. The short period grating
acts as a wavelength selective reflection filter, to reflects the signal fall within certain
wavelength, as depicted in Figure 2.5. The reflected wavelength, namely Bragg
wavelength, is highly dependent to the elongation of grating imposed mechanically
8
and thermally to the fibre (Eric, et al., 2011). Therefore, FBG is in general
characterised by the Bragg wavelength. Any elongation in the FBG will increases the
grating period, and hence results shifting in Bragg wavelength (as can see in eq. 2.2).
Therefore, by measuring the shifting in Bragg wavelength, structural strain down to
scale of micron strain (𝜇𝜀) can be detected.
𝜆𝐵 = 2𝑛𝑒𝑓𝑓Λ (2.2)
Where 𝑛𝑒𝑓𝑓 is the effective refractive index of the fundamental mode, Λ is the
grating period.
Figure 2.4: Schematic Diagram of an In-fibre FBG Sensor
Figure 2.5: Wavelength-selective Reflectional Filter of FBG
9
2.1.3 Mach-Zehnder Interferometer Sensor
MZI sensor in typical, in-line MZI can be fabricated by laser ablation (Morales &
Lieber, 1998), fibre pulling (Clohessy, et al., 2005) or direct draw from bulk materials
(Xing, et al., 2008), causes necking at the taper sites. The fabrication techniques of
taper are much more easy compared to grating. Among the existing techniques, flame
heating technique is the most versatile fabrication method, which produces the fibre
taper with relatively good physical properties (Harun, et al., 2010).
𝑉 =2𝜋
𝜆𝑎√𝑛𝑐𝑜
2 − 𝑛𝑐𝑙2
(2.3)
Owing to necking (reduction in radius) of fibre core, a, drop in V-number
induces scattering loss at core-cladding interface. Therein, the input signal
(fundamental mode) in the core splits partially into the cladding at the first taper site.
The cladding modes propagate within the cladding along the interferometric length,
and recouple back to the fundamental mode at the second taper (Li, et al., 2011). The
mode splitting of MZI fibre taper is depicted in Figure 2.6.
Figure 2.6: Schematic Diagram of an In-line Tapered MZI Sensor
10
Figure 2.7: Splitting of Fundamental Core Mode in First Taper Site
When the cladding modes propagating along the interferometric length, it is
highly sensitive to the change of curvature, where the bending will result a tremendous
loss in signal. Adiabaticity of the MZI fibre which tells the degree of loss, is dependent
on the dimension of taper such as, the local taper length-scale, 𝑧𝑡 and the taper angle,
Ω(𝑧). Trigonometrically, taper angle can be related to the local taper length-scale by
equating:
tan Ω(𝑧) =𝑑𝜌
𝑑𝑧
(2.4)
where 𝜌 is the core radius.
Taking into consideration the small angle approximation, thus the taper angle,
Ω(𝑧) can be further expressed as:
Ω(𝑧) =𝑑𝜌(𝑧)
𝑑𝑧≈
𝜌(𝑧)
𝑧𝑡(𝑧)
(2.5)
To illustrate the Adiabaticity of fibre taper in a more complete manner, herein to
introduce the coupling length between the fundamental and cladding mode, 𝑧𝑏:
11
𝑧𝑏(𝑧) =2𝜋
𝛽1(𝑧) − 𝛽2(𝑧)
(2.6)
where 𝛽1(𝑧) and 𝛽2(𝑧) are the propagating constant of core and cladding respectively.
The fibre taper is considered adiabatic if 𝑧𝑡 ≫ 𝑧𝑏 , which the loss of cladding
propagation is negligible. In most application of the MZI sensor in detecting strain,
non-adiabatic taper turns out to be more preferable in term of the sensitivity, as a more
obvious curvature effect can be observed through a significant attenuation.
As governed by eq. 2.8 and 2.9, different degree of curvature causes elongation
in the interferometric length, L, and in turns consequent to shift in Free Spectral Range
(FSR). The changing in FSR is inversely proportional to the variation of
interferometric length. As can see in Figure 2.8, the shortening in FSR is observed in
the transmission spectra as the interferometric length is varied from 20 mm to 40 mm.
𝐼 = 𝐼1(𝜆) + 𝐼2(𝜆) + 2√𝐼1(𝜆)𝐼2(𝜆) cos [2𝜋∆𝑛𝑒𝑓𝑓𝐿
𝜆]
(2.7)
𝐹𝑆𝑅 ≈𝜆2
∆𝑛𝑒𝑓𝑓𝐿
(2.8)
12
Figure 2.8: Transmission Spectra of MZIs under Different Interferometer Lengths:
(a) 20 mm, (b) 30 mm, (c) 36 mm and (d) 40 mm (Li, et al., 2011)
13
Comparison of Optical Fibre Sensors
Tally up the criterion of the aforementioned optical fibre sensors, a table of comparison
comprised of the respective sensitivities, fabrication methods, advantages and
limitations is summarised as in table 2.1.
Table 2.1: Comparison Table of Optical Fibre Sensor
Sensor Sensitivity Fabrication method Fabrication
complexity
Fabrication cost
LPG
(Yin, et
al., 2008)
0.1-5.6 pm/μ𝜀
67-154 pm/°C
UV irradiation,
CO2 laser
Infrared
femtosecond
laser pulses
High High
FBG
(Yin, et
al., 2008)
1.2 pm/µε
14-25 pm/°C
UV irradiation
Electron beam
interference
lithography
High High
In-line
tapered
MZI (Sun, et
al., 2010)
1.07 pm/µε
50.0 pm/mm
11.7 pm/°C
Laser ablation
Fibre pulling
Direct draw
from bulk
materials
Flame heating
Low Low
2.2.1 Cross Sensitivity in Sensors
Cross sensitivity is a common issue for most of the existing optical fibre sensors. Defer
to the intention of detecting the sole structural change, issue arises when the cross
sensitivity of sensor responds to the ambient temperature during the detection. A
general comparison on cross-sensitivities of standard LPG with FBG shows that, the
strain sensitivities of LPG is vying the FBG. However, the profound thermal sensitivity
in LPG makes itself a less preferable strain sensor. (Yin, et al., 2008). Particular
technique is proposed by many researches to compensate the temperature effect from
14
the strain and curvature detection. The mutual-compensating transfer function matrix
is the technique commonly used for most of the sensors. (Mokhtar, et al., 2012; Raji,
et al., 2016)
Tapered MZI sensor has a vying thermal sensitivity with LPG and FBG, and is
highly sensitive to micro-bending. Figure 2.10 shows the spectral shifting in
responding to the temperature variation (from 16℃ to 50℃), it is shown that the sensor
experiences power attenuation and decreasing in FSR (causes spectral shifting) when
the temperature is higher. The temperature effect contributed to the MZI sensor can
also be compensated using the mutual-compensating transfer function matrix, which
as reported in a journal by Raji, et. al.
Figure 2.9: Spectral Shift Due to Temperature Changes (Raji, et al., 2016)
2.2.2 Fragility of Fabricated Segment in Sensors
Fragility is the major drawback of both fibre gratings and tapers in the SHM
application. The lengthy processes of post-fabrication manufacturing such as stripping,
cleaving, splicing and packaging, may contribute deformation to the segment and more
critically, fracture. The repeated mechanical actions like lifting, dropping, bending and
15
clipping during the manufacturing processes, may incurs fatigue failures in the grating
and taper.
On that note, the fragility is introduced inevitably into the fibre grating during
the fabrication process. As the fibre is illuminated under an intense UV laser light, the
silicon-oxygen bonds were broken, resulting a slight increment in refractive index, at
the same time causes damages to the structure of the fibre (Doyle & Crispin, 2003).
Sensor System Design
With the combination of reasons of high curvature sensitivity, low cost and ease of
fabrication, MZI sensor are chosen to be justified by carries out further curvature
calibrations. Design criterion of MZI sensor in the fabrication and packaging will be
discussed in the following section.
2.3.1 Fabrication Criteria
Figure 2.10: Schematic Diagram of the Fabricated MZI Sensor
Figure 2.14 shows the schematic diagram of the fabricated MZI sensor which
consist of 2 taper regions. As reported by Wang, a minimum 0.05 mm diameter taper
is required in order to meet the threshold for cladding mode splitting (Wang, 2012).
Besides, a short interferometric length can get a wide FSR profile (as shown in figure
2.15), hence a preferable larger linearity region of spectrum can be observed (Li, et al.,
2011).
16
Figure 2.11: Linearity Region of Spectrum Profile for Different Interferometric
Length, L (Li, et al., 2011)
2.3.2 Packaging of Sensors
To protect the sensor under harsh condition of the real sensing environment, packaging
is introduced to the in-line tapered MZI sensor. Packaging criterion such as
packaging’s material, dimension and adhesive method are the possible contributing
factors to decouple the curvature sensitivity of the sensor.
Polypropylene is generally used as packaging material in optical fibre sensors,
owing to the mechanical properties of low young modulus with relatively high tensile
stress and breaking strain (Anon., 1999-2001). Optical fibre sensors are generally
sandwiched in between two polypropylene slabs by means of cyanoacrylate adhesive,
due to the good strain coupling ability of cyanoacrylate over other type of adhesives
(Clements, 2006). In addition, the non-covalent adhesive using cyanoacrylate which
only involves polymerization the surface. will neither destroys the structure of polymer,
nor causes changes of property.
In a study on the effect of FBG’s packaging dimensions established by Mokthar
at el, several dimensions of non-uniform packaging designs were compared in term of
the effective strain and temperature sensitivities. A general result show that, strain
sensitivity is high for the narrow-width and thin packaging design, whereas, significant
thermal sensitivity shows in the wide-width and thin design (Mokhtar, et al., 2012).
17
However, not as prevailed as FBGs, the effect of packaging to tapered MZI
sensor is not popular in the research. The splitting of modes in MZI sensor is differs
with its taper diameters and taper lengths. Practically, MZI sensor with exactly
identical taper diameters and taper lengths is hard to be duplicated. Thus every MZI
sensor is unique in term of the sensitivity to the curvature, which implies that the
characterization must be done using the same sensor.
18
3 METHODOLOGY
In this chapter, methodology of this project which comprises of two stages: the
preparation of sensor and characterization of sensor, will be discussed in detailed. In
the characterization stage, the optical output (power) of sensor is calibrated to the
physical curvature. The characterizations are carried out based on various wavelengths
and packaging thicknesses. Lastly, back-tracing of sensor is conducted to justify the
repeatable sensing capability of the sensor.
Figure 3.1: Outline of the Methodology
19
Fabrication of Sensor
By integrating the arc ablation and fibre pulling technique, an in-line MZI taper can
be fabricated using an in-house built arc discharge pulling rig. The diameters of the
tapers were checked under a digital microscope, with minimum acceptance diameter
smaller than 0.05 mm. Figure 3.2 shows the tapers images captured by the digital
microscope, where both of the tapers are checked smaller than 0.05 mm. As discussed
in the previous chapter (section 2.3.1), a short interferometric length is preferable to
observe a wider linearity region. However, due to the limitation of the pulling rig, the
minimum interferometric length can be achieved in this project is 5 cm as depicted in
Figure 3.3.
(a)
(b)
Figure 3.2: (a) First (b) and Second Fibre Tapers of MZI Sensor
Figure 3.3: MZI Sensor with Interferometric Length of 5 cm
Packaging of Sensor
A packaging design of MZI sensors are depicted as in Figure 3.4, which comprises of
two polypropylene slabs with dimension 12 cm ×2 cm ×0.1 cm. The fabricated sensor
is packaged by sandwiching it at the middle of two polypropylene slabs, and glued
0.025 mm 0.025 mm
20
using cyanoacrylate glue. The packaged MZI sensor is left for few hours to ensure the
cyanoacrylate is fully cured before it is ready for the tests. The packaged MZI sensor
is shown in Figure 3.5.
Figure 3.4: Schematic Diagram of Uniform Width MZI Package Design
Figure 3.5: Photo of a Packaged MZI Sensor
Curvature Calibration of Sensor
After few hours of cure time, the test is readily to be run. The packaged MZI sensor is
input to a tuneable optical source (single wavelength laser, where the wavelength is
set to be 1310 nm initially) and output to an optical power meter as shown in Figure
3.6 (b). The setup of curvature calibration is depicted as in Figure 3.6 (a). The MZI
sensor was attached to the centre of an 8 m long steel bar by mean of the cyanoacrylate
adhesive. The steel bar is then deployed into the bending test equipment as depicted in
Figure 3.7.
21
(a)
(b)
Figure 3.6: (a) Schematic Diagram of Placement of MZI Sensor on Steel Bar; (b)
Experimental Setup of MZI Sensor on Steel Bar
Figure 3.7: Experimental Setup of Bending Test
22
Figure 3.8: Schematic Diagram of Steel Bar Curvature
Figure 3.8 shows the schematic diagram of hardened steel bar with dimension
80cm ×2.5cm × 0.47cm, and its vertical displacement, d when loads applied. By
increasingly varies the load with 50g per increment, the respective power intensity is
recorded from the optical power meter. Simultaneously, the vertical displacement, 𝑑
is recorded from the Linear Variable Differential Transformer (LVDT). In the
succeeding procedure, the calibration is reversed in the order by unloading the system,
likewise, the optical power and vertical displacement was recorded to verify the
accuracy of measurement.
Figure 3.9: Trigonometric Diagram of Steel Bar Curvature
Figure 3.9 shows the trigonometric representation used to derive the curvature,
C from the vertical displacement, d, where R is the radius of curvature, L is the length
23
of harden steel bar and h is the vertical height of the centre of curvature from the
hardened steel bar. Curvature of bending can be determined through eq. 3.4, which can
be easily derived as followed. The radius of curvature can be equated as in eq. 3.3 by
substituting eq. 3.2 into and eq. 3.1, and therefore curvature, C can be derived as in eq.
3.4.
𝑅 = 𝑑 + ℎ (3.1)
ℎ2 = 𝑅2 − (𝐿
2)
2
(3.2)
𝑅 =4𝑑2 + 𝐿2
𝑑
(3.3)
3.3.1 Calibration Based on Various Wavelengths
The Curvature calibrations are conducted at various wavelengths by using the similar
setup and procedures as in section 3.3, to characterise the sensitivity of sensor in
responding to various wavelength. Setup of the whole sensing system is remained, the
wavelength of optical source is varied to 1490 nm and 1550 nm, similar set of
measurement is taken as in previous section for respective calibration.
Calibration Based on Different Packaging Thicknesses
Thickness is an important criterion that affects the sensitivity of sensor. In this section,
two configuration of sensor with thickness A (an additional 2 mm slab attached in
between the sensor and steel bar) and thickness B (an additional 2 mm slab attached
on the top of sensor) were prepared as depicted in Figure 3.10. The characterization
𝐶 =1
𝑅=
𝑑
4𝑑2 + 𝐿2
(3.4)
24
based on three thicknesses (pristine thickness, thickness A and B) is aimed to justify
which thickness gives the optimum sensitivity.
Figure 3.10: Configuration of Sensor with (a) Thickness A and (b) Thickness B
Back-tracing of the Curvature
The back-tracing process is proposed at the end of calibration to validate the real
practicality of the sensor. Where in this section, the optical power is recorded from the
optical power meter for each successive loading and unloading. The respective
curvature is then back-traced based on the characterised data from the previous
calibrations. The back-traced curvature is then compared to the real curvature
measured by the LVDT.
25
4 RESULTS AND DISCUSSION
Curvature Sensitivity Based on Various Wavelength
In the part hereof, the curvature sensitivities of the MZI sensor in packaging of pristine
thickness will be characterised based on three operating wavelengths (1310 nm, 1490
nm and 1550 nm). Figure 4.1 shows one of the raw results of optical power change
detected by the packaged MZI sensor, in responding to increase loading and unloading
at operating wavelength 1550 nm. The operating region (also referred as linearity
region) was selected from the raw data, so that the non-linearity regions are excluded
in the analyses. By comparing the R2 value of several regional plots, the region with
the highest R2 value (best fit to the linear slope) was selected.
Figure 4.1: Raw Result of Curvature Calibration of Sensor with Pristine Thickness at
1490 nm (× indicates the loading slope and △ indicates the unloading slope)
0.0390
0.0395
0.0400
0.0405
0.0410
0.0415
0.0420
0 0.002 0.004 0.006 0.008 0.01Aver
age
Op
tica
l P
ow
er (
uW
)
Curvature (m-1)
Operating region
26
Figure 4.2 depicted the linearity and non-linearity region within the responding
spectra when various weight of load is applied to the sensor. The non-linearity region
relates the optical power to the weight imposed weakly, therefore shall be excluded as
it does not carry an analysable information. In spite of it, linearity region gives a clear
and analysable relation between the optical power and the loading, which is labelled
as the operating region of the sensor. The curvature sensitivities of MZI sensor can be
characterised within the region, by calibrating the optical power to the respective
curvature.
Figure 4.2: Model of the Linearity and Non-linearity Region in the Responding
Spectra
y = -0.0373x + 22.044
R² = 0.7698
0.010
0.012
0.014
0.016
0.018
0.020
0.022
0.024
0 50 100 150 200
Aver
age
Op
tica
l P
ow
er (
uW
)
Weight (g)
y = -0.0902x + 36.729
R² = 0.985
0.010
0.015
0.020
0.025
0.030
0.035
0.040
0 50 100 150 200
Aver
age
Op
tica
l P
ow
er (
uW
)
Weight (g)
0.010
0.015
0.020
0.025
0.030
0.035
0.040
1530 1535 1540 1545 1550 1555 1560
Aver
age
Op
tica
l P
ow
er (
uW
)
Wavelength (nm)
0g 50g 100g 150g 200g
Linearity region Non-linearity region
27
Figure 4.3: Curvature Calibration of Sensor with Pristine Thickness in Responding to
Wavelength 1310 nm, 1490 nm and 1550 nm
Figure 4.4: Comparison of Flexural Moduli with Theoretical Value within the
Operating Region
Figure 4.3 shows the optical powers slopes (operating region) of the packaged
MZI sensor in responding to the three operating wavelengths. To validate the accuracy
of measurement, the flexural moduli within the operating regions are collated with the
theoretical value which ranges from 200-210 GPa (Hosford, 2010) in Figure 4.4.
Disparities in both the trend and offset are observed among responding slopes of the
three wavelengths. Consequently, the sensitivities of the packaged MZI sensor at
wavelength 1310 nm, 1490 nm and 1550 nm are different, which is 0.461 𝜇Wm-1,
0.346 𝜇Wm-1 and -1.10 𝜇Wm-1 respectively. These phenomena can be explained by
the wavelength and polarization dependent properties of MZI sensor which will be
elaborated in the following parts.
y = 0.461x + 0.0266
R² = 0.9981
y = 0.3454x + 0.0392
R² = 0.9828
y = -1.0955x + 0.0376
R² = 0.9959
0.020
0.025
0.030
0.035
0.040
0.045
0 0.002 0.004 0.006 0.008
Aver
age
Op
tica
l P
ow
er (
uW
)
Curvature (m-1)
1490 nm
1310 nm
1550 nm
200
202
204
206
208
210
212
214
0 100 200 300 400
Fle
xura
l M
od
ulu
s (G
Pa)
Weight (g)
1550 nm
1310 nm
1490 nm
28
4.1.1 Wavelength Dependent Property of MZI Sensor
Figure 4.5 shows the modelled output spectra of MZI sensor when imposed to various
loads, and the consequent optical power changes at two different operating
wavelengths. From the output spectral, attenuation in power intensity and shifting (in
practical, the FSR shifting, as has been discussed in chapter 2) are observed. The dash
lines fall on the spectra indicates the operating wavelengths 1545 nm and 1550 nm,
and their responding optical powers at respective loadings. By plotting the optical
powers against the weight imposed, two distinct optical power slopes are obtained,
where the different in changing trend is observed. Furthermore, the variation of
sensitivities is modelled as in Figure 4.6 based on wavelength 1549 nm, 1550 nm and
1551 nm within the same operating region. Thus, deduction can be made that the trend
of responding slope and sensitivity of MZI sensor is wavelength dependent.
Figure 4.5: Disparity in Power Trend of MZI Sensor in Responding to Variation of
Wavelength
y = 0.0716x + 16.123
R² = 0.9887
0.010
0.015
0.020
0.025
0.030
0.035
0 50 100 150 200
Aver
age
Op
tica
l P
ow
er (
uW
)
Weight (g)
Operating wavelength 1545 nm
y = -0.0902x + 36.729
R² = 0.985
0.010
0.015
0.020
0.025
0.030
0.035
0.040
0 50 100 150 200
Aver
age
Op
tica
l P
ow
er (
uW
)
Weight (g)
Operating wavelength 1550 nm
0.010
0.015
0.020
0.025
0.030
0.035
0.040
1530 1535 1540 1545 1550 1555 1560
Aver
age
Op
tica
l P
ow
er
(uW
)
Wavelength (nm)
0g 50g 100g 150g 200g
29
Figure 4.6: Disparity in Sensitivities of MZI Sensor in Reponding to Variation of
Wavelength
4.1.2 Polarization Dependent Property of MZI Sensor
MZI sensor is a polarization dependent sensor, where the variation in polarization state
will causes the fluctuation in the degree of signal attenuation. Figure 4.7 shows the
variation of polarization state observed using an Optical Spectrum Analyser (OSA).
By using a polarization controller (model FPC560 three paddles controller), the
pristine arbitrary state (state1) in the MZI sensor is transformed to 3 subsequent
arbitrary states (state 2, 3 and 4). The variation incurs both power fluctuation and
spectral shifting in the output spectrum. By making use of these observed effects, two
responding spectra at polarization state 1 and 4 were modelled as in Figure 4.8.
Operating wavelength of 1550 nm was selected for optical power-weight plotting, to
further compare the optical power in responding to various loading at the two
polarization states. The deviation in offset power is observed in the optical power-
y = -0.0774x + 30.304
R² = 0.986
0.010
0.015
0.020
0.025
0.030
0.035
0 100 200
Op
tica
l P
ow
er (
uW
)
Weight (g)
y = -0.0902x + 36.729
R² = 0.985
0.010
0.015
0.020
0.025
0.030
0.035
0.040
0 100 200
Op
tica
l P
ow
er (
uW
)
Weight (g)
0.010
0.015
0.020
0.025
0.030
0.035
0.040
1530 1535 1540 1545 1550 1555 1560
Aver
age
Op
tica
l P
ow
er
(uW
)
Wavelength (nm)
0g 50g 100g 150g 200g
Operating wavelength 1550 nm Operating wavelength 1549 nm
y = -0.0679x + 39.006
R² = 0.9133
0.010
0.015
0.020
0.025
0.030
0.035
0.040
0.045
0 100 200
Op
tica
l P
ow
er (
uW
)
Weight (g)
Operating wavelength 1551 nm
30
weight plotting. Thereby, we can deduce that the variation of polarization state in MZI
sensor contributes to fluctuation in offset power.
Figure 4.7: Power Changes and Spectral Shifting in Responding to Variation of
Polarization States
Figure 4.8: Offset Disparity in Responding to the Variation of Polarization State
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
1520 1525 1530 1535 1540 1545 1550 1555 1560
Aver
age
Op
tica
l P
ow
er (
nW
)
Wavelength (nm)
Arbitrary state 1 Arbitrary state 4
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0.035
0.040
1530 1535 1540 1545 1550 1555 1560
Aver
age
Op
tica
l P
ow
er (
uW
)
Wavelength (nm)
0g 50g 100g 150g 200g
Arbitrary state 1
y = -0.0811x + 42.896
R² = 0.9932
y = -0.0902x + 36.729
R² = 0.985
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0.035
0.040
0.045
0.050
0 50 100 150 200
Aver
age
Op
tica
l P
ow
er (
uW
)
Weight (g)
Arbitrary state 1
Arbitrary state 4
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0.035
0.040
0.045
1530 1535 1540 1545 1550 1555 1560
Aver
age
Op
tica
l P
ow
er (
uW
)
Wavelength (nm)
0g 50g 100g 150g 200g
Arbitrary state 4
Arbitrary state 2 Arbitrary state 3
Operating wavelength 1550 nm
31
This undesired phenomenon arises when there is any changing in curls and
spools of the transmission path of optical fibre, and was found practically causes
fluctuation in the optical power offset up to degree of ±4 dBm. To prevent this, as well
as for safety purpose, the optical fibre is fixed in a constant path using tapes for the
whole calibration process. In the following section, an additional 2mm slab will be
added to the sensor in two configurations (as discussed earlier). During the process,
detaching and re-attaching of sensor on the steel bar is required, where the curl and
spool of optical fibre (transmission path) will not more maintained the same. Therefore,
the newly thickened sensor is expected to have a deviation in the offset power from
the pristine sensor. To cope with this, and for ease of comparison, a polarization
controller is added to the sensing system, to offset the deviation.
Curvature Sensitivity Based on Different Packaging Thicknesses
Packaging of sensor is the highlight of this project, where the thickness of packaging
is an important criterion to be characterised. The thickness of packaging was expected
intuitionally affects the sensitivity of sensor, in term of strain (the degree of elongation).
To differentiate the thickness of packaging, an additional 2 mm slab was added in
between the sensor and steel bar (thickness A) and on the top of sensor (thickness B),
as depicted in Figure 4.9.
Figure 4.9: Distance from the Sensor in (a) Thickness A and (b) Thickness B
32
Figure 4.10 shows the curvature segment of sensor packaging (upper slab) which
illustrated for ease of explaining the relation between the strain, 𝜀 and the location of
MZI sensor within the packaging. The centre of packaging (location of MZI sensor) is
defined using distance, y outward away from the nature axis. The nature axis is the
axis that remains the same length (experiences zero strain) when imposed to curvature.
Geometrically, the location of MZI sensor, y for thickness A is higher than the pristine
thickness and thickness B. Equation 4.1 governs a linearly proportional relation
between the strain and distance, y. The derivation of the strain equation is shown in
Appendix A.
Figure 4.10: The Curvature Segment of the Sensor Packaging
𝜀 =∆𝑠′ − ∆𝑥
∆𝑥=
𝑦
𝑅
4.1
As can see in Figure 4.10, larger y value consequents a higher strain. Thus, sensor
with thickness A is expected to experience a higher strain than the pristine thickness
and thickness B (i.e. it bends more compared to the other two). Knowing that y value
in sensor with thickness B do not varies with the pristine thickness, it is expected to
experience an equivalent strain. Therefore, deduction can be made that the sensitivity
of sensor in thickness A is comparatively better, whereas, thickness B will not
significantly affect the sensitivity.
Nature Axis
Packaging segment (upper slab)
Centre of packaging
(location of MZI sensor)
33
4.2.1 Comparison of the Pristine Thickness, Thickness A and B.
Figure 4.11 shows the comparative results (linearity region) of the sensor with
packaging of pristine thickness, thickness A and thickness B in responding to the three
wavelengths respectively. The operating regions are selected from the respective raw
data for reason as discussed in the earlier section. The comparative results in term of
packaging thicknesses are shown in Figure 4.12, 4.14 and 4.15 for further discussion.
To validate the measurement accuracy, the flexural moduli within the operating
regions for sensor of each packaging thicknesses are collated with the theoretical value
in Figure 4.12.
(a)
(a)
(b)
(b)
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0 0.002 0.004 0.006 0.008 0.01Aver
age
Op
tica
l P
ow
er (
uW
)
Curvature (m-1)
1550 nm
1490 nm
1310 nm
190
195
200
205
210
215
220
0 100 200 300 400
Fle
xura
l M
od
ulu
s (G
Pa)
Weight (g)
1550 nm
1490 nm
1310 nm
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0 0.002 0.004 0.006 0.008 0.01Aver
age
Op
tica
l P
ow
er (
uW
)
Curvature (m-1)
1550 nm
1490 nm
1310 nm
198
200
202
204
206
208
210
212
214
0 100 200 300 400
Fle
xura
l M
od
ulu
s (G
Pa)
Weight (g)
1550 nm1490 nm
1310 nm
34
(c)
Figure 4.11: Curvature Calibration in
Responding to Wavelength 1310 nm, 1490 nm
and 1550 nm for Sensor with (a) Pristine
Thickness, (b) Thickness A and (c) Thickness B
(c)
Figure 4.12: Comparison of Flexural Moduli
with Theoretical Value within the Operating
Region for Sensor with (a) Pristine Thickness,
(b) Thickness A and (c) Thickness B
Figure 4.13: Curvature Calibration of Sensor with Various Thicknesses at
Wavelength 1310 nm
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0 0.002 0.004 0.006 0.008 0.01
Aver
age
Op
tica
l P
ow
er (
uW
)
Curvature (m-1)
1550 nm
1490 nm
1310 nm
190
195
200
205
210
215
220
0 100 200 300 400
Fle
xura
l M
od
ulu
s (G
Pa)
Weight (g)
1310 nm
1490 nm
1550 nm
y = -1.9019x + 0.1459
R² = 0.9994
y = -1.3275x + 0.1464
R² = 0.9956
y = -3.2653x + 0.1455
R² = 0.99830.115
0.120
0.125
0.130
0.135
0.140
0.145
0.150
0 0.002 0.004 0.006 0.008 0.01
Aver
age
Op
tica
l P
ow
er (
uW
)
Curvature (m-1)
Thickness B
Pristine thickness
Thickness A
35
Figure 4.14: Curvature Calibration of Sensor with Various Thicknesses at
Wavelength 1490 nm
Figure 4.15: Curvature Calibration of Sensor with Various Thicknesses at
Wavelength 1550 nm
Table 4.1: Comparison Table of Sensor of Three Different Thicknesses.
Wavelength
(nm)
Sensitivity (𝜇𝑊𝑚−1)
Pristine thickness Thickness A Thickness B
1310 -1.33 -3.27 -1.90
1490 -0.21 3.53 -0.19
1550 0.43 -0.69 0.59
Table 4.2: Optimal Detectable Radius and Curvature
Maximum detectable radius, R (km) Minimum detectable curvature, C (km-1)
1.6-4.0 0.25-0.625
y = -0.2136x + 0.1155
R² = 0.9751
y = -0.1916x + 0.1194
R² = 0.9991
y = 3.5324x + 0.1153
R² = 0.9959
0.100
0.105
0.110
0.115
0.120
0.125
0.130
0.135
0.140
0.145
0.150
0 0.002 0.004 0.006 0.008 0.01
Aver
age
Op
tica
l P
ow
er (
uW
)
Curvature (m-1)
Pristine thickness
Thickness A
Thickness B
y = 0.4288x + 0.0132
R² = 0.9907
y = 0.5953x + 0.0132
R² = 0.991
y = -0.6926x + 0.0134
R² = 0.998
0.007
0.009
0.011
0.013
0.015
0.017
0.019
0.021
0 0.002 0.004 0.006 0.008 0.01
Aver
age
Op
tica
l P
ow
er (
uW
)
Curvature (m-1)
Pristine thickness
Thickness B
Thickness A
36
To offset the power deviation caused by the changing of transmission path,
polarization controller is added into the sensing system. Nevertheless, introducing the
polarization controller into the system will result signal attenuation up to -5 dBm.
Therefore, pristine power offset is inevitable changed. Curvature calibration for sensor
with the pristine thickness is repeated, the result is taken as a new reference to compare
the effect of the two packaging thicknesses. As the calibration proceeds to thickness A
and B, the polarization controller is used to offset the deviation of optical power, so
that the offset power is same as what obtained in the calibration of the pristine
thickness.
In Figure 4.13, 4.14and 4.15, sensor for each thicknesses were calibrated to the
equivalent offset power with aids of polarization controller, for ease of comparison.
The changing trends of the optical power is found identical for the three thicknesses
when tested at wavelength 1310nm. The result tallies with the expectation in section
4.2, where the sensitivity of sensor in thickness A is expected higher than the another
two, as reported in Table 4.1, -3.27 𝜇Wm-1 for thickness A, -1.33 𝜇Wm-1 for pristine
thickness, and -1.90 𝜇 Wm-1 for thickness B. As for sensor in thickness B, the
sensitivities do not deviate a lot from pristine thickness for all the three wavelengths.
However, the trend of slope is inverted at thickness A for both wavelength 1490 nm
and 1550 nm. The contradiction in trend of slope of will be explained in the following
part.
4.2.2 Disparity in Trend of Optical Power Slope
Owing to polarization dependent property of MZI sensor, polarization controller is
used to tune the offset power for ease of comparison. However, by tuning the
polarization controller, the spectrum is not necessarily reverted to the original
spectrum even that the power is successfully tuned back to the identical offset value.
Instead of it, the spectrum might has shifted and fluctuated in the power. In this case,
the deviation in the trend of optical power slope might happens.
37
Figure 4.16: Loading Spectra in Arbitrary State 1 and 2
Figure 4.16 shows two spectrum in responding to maximum variation of
polarization states, by tuning the 2nd and 3rd paddle ( 𝜆
2 and
𝜆
4 paddle). In aids of the
figure, few points of intersection (A, B, C and D) between the two spectrum were
selected to demonstrated the how the deviations in trend is possible to happen. Among
these points, the same value of optical power is observed for the two spectrum at the
same operating wavelength, where the two spectrum are contributing to a different
trend of slope (more obvious at point B, C and D). This implies that the identical offset
power measured can come from different spectrum which might consequent to
disparity in trend. This explained the contradiction in trend of optical power slope for
the case of thickness A at both wavelength 1490 nm and 1550 nm.
Back-tracing of Curvature
From the previous section, MZI sensor was characterised according to the packaging
thicknesses, in term of the sensitivity (gradient) and the offset power. At the last part
of the project, the curvature was back-traced using the sensor with thickness A at
wavelength 1310 nm. The optical power in responding to the loading and unloading
was recorded, and back-traced using the characterised offset and gradient. As shown
below is the back-tracing equation of curvature for MZI sensor with thickness A which
operates at wavelength 1310 nm.
Arbitrary state 1
Arbitrary state 2
A
B C
D
38
𝐶𝑒𝑠𝑡𝑖𝑚𝑎𝑡𝑒𝑑 =𝑂𝑝𝑡𝑖𝑐𝑎𝑙 𝑃𝑜𝑤𝑒𝑟 − 𝑂𝑝𝑡𝑖𝑐𝑎𝑙 𝑝𝑜𝑤𝑒𝑟 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 ̅̅ ̅̅ ̅̅ ̅̅ ̅̅ ̅̅ ̅̅ ̅̅ ̅̅ ̅̅ ̅̅ ̅̅ ̅̅ ̅̅ ̅̅ ̅̅ ̅̅ ̅̅
𝐺𝑟𝑎𝑑𝑖𝑒𝑛𝑡
The mean difference between each measured optical power and the respective
characterised data is computed to offset the values, if there is any fluctuation in the
optical power from the characterised data. Figure 4.17 shows the comparison of the
back-traced curvature and the characterised curvature, with error bar of 14% and
correlation values of 0.9928 (loading slope) and 0.9887 (unloading slope).
Figure 4.17: Correlation between the Back-traced Curvature and the Characterised
Curvature (with error bar of 14%)
0.000
0.001
0.002
0.003
0.004
0.005
0.006
0.007
0.008
0.009
0.010
0 2 4 6 8 10 12
Curv
ature
(m
-1)
39
5 CONCLUSION AND RECOMMENDATIONS
Conclusion
Polypropylene packaging was introduced to the fibre-based in-line Mach-Zehnder
Interferometer sensor to protect the sensor under harsh condition of the real sensing
environment. At the first part of calibration, the packaged MZI sensor was
characterised based on the imposed curvature at various wavelengths. Disparity in
trend and offset power were observed among the three different operating wavelengths
due to the wavelength dependent property of MZI sensor. Besides, polarization
dependent property of MZI sensor also contributes to fluctuation of power. A
polarization controller was used to offset the deviation, so that the optical power of the
three thicknesses can be compared based on the same offset value. Packaging with
thickness A was found to have the best curvature sensitivity than the other two, where
the optimal sensitivity is up to is -3.27 𝜇Wm-1. The packaged MZI sensor is capable
to detect minimum curvature of 0.25 km-1 and maximum curvature radius up to 4 km,
which is considerably sensitive in monitoring the structural health.
Future Works
Proceeding to this project, the packaged sensor is suggested to embedded into the
concrete reinforcement bar to implement the real sensing condition. Besides, to
improve the accuracy of curvature sensing, temperature compensation is suggested to
be carried out using the mutual-compensating technique which has been discussed in
section 2.2.1.
40
Furthermore, the distributed sensing system is proposed to be implemented, to
cater the multiple points curvature sensing. Wherein, several multiplexing techniques
are suggested as follow for the distributed sensing module. The common found
multiplexing technique in MZI sensor is TDM. Figure 5.1 illustrates MZI multiplexing
technique using TDM in time domain. Single wavelength source is input into a
modulator and an 1×N splitter before entering the sensor array. A modulator is used
to manipulates N pulses vary with time, and split them accordingly into the arrays by
a splitter. Fibre loops in each array cause delay to the input signal in different extent,
therefore only allow pulse at certain time frame pass through.
Figure 5.1: Schematic Diagram of MZI Multiplexing Technique using TDM
Figure 5.2 illustrates the MZI multiplexing technique using Subcarrier
Multiplexing Method (SMM), which is a costlier but more complicated method. The
configuration is slightly different at the modulator part, where the modulators are
allocated in every arrays after the splitter, to produce the signal in distinct pulses at
various frequencies. The signals with distinct frequency pass though the sensor and
respond to the curvature change. (any additional points to add in?)
Figure 5.2: Schematic Diagram of MZI Multiplexing Technique using SDM
41
REFERENCES
Ali, M. et al., 2014. Tapered-EDF-Based Mach–Zehnder. IEEE Photonics Juornal
Volume 6, Number 5.
Anon., 1999-2001. Material Science and Engineering. Mechnics of materials modules.
[Online] Available at: http://ocw.mit.edu/courses/materials-science-and-
engineering/3-11-mechanics-of-materials-fall-1999/modules/props.pdf
Antunes, P., Domingues, F., Granada, M. & André, P., 2012. Mechanical Properties
of Optical Fibers. In: M. Yasin, S. W. Harun & H. Arof, eds. Selected topics on
Optical Fiber technology. s.l.:In Tech, pp. 537-550.
Billie, J., Spencer & Manuel, R., 2004. Smart Sensing Technology For Structural
Health. 13th World Conference on Earthquake Engineering, p. 1791.
Clements, A., 2006. Selection of Optical Adsesives, Arizona: University of Arizona,
College of Optical Science .
Clohessy, A. M., Healy, N., Murphy, D. F. & Hussey, C. D., 2005. Short low-loss
nanowire tapers on singlemode fibres. Electron. Lett. 41 (17), p. 954–955.
Crandall, S. H., Dahl, N. C., Lardner, T. J. & Sivakumar, M. S., 2013. An Introduction
to Mechanics of Solids. 3rd ed. Singapore: McGraw-Hill Education.
Daniele, I., Branko, G., Angelo, F. & Roberto, w., 2007. Pipeline Leakage Detection
and Localization Using Distributed Fiber Optic Sensing. Rio Pipeline Conference
& Exposition 2007.
Doyle & Crispin, 2003. Fibre Bragg Grating Sensors: An Introduction to Bragg
gratings and interrogation techniques. Senior Applications Engineer, Smart Fibres
Ltd. 2003.
Dreiden, G. V., Khusnutdinova, K. R., Samsonov, A. M. & Semenova, I. V., 2012.
Comparison of the effect of cyanoacrylate- and polyurethane-based adhesives on a
longitudinal strain solitary wave in layered PMMA waveguides. Journal of Applied
Physics.
Eric, U., William, B. & Spillman, J. R., 2011. Fiber Optic Sensors: An Introduction
for Engineers and Scientists (Second edition). New Jersey, USA: John Wiley &
Sons.
42
Harun, S. W., Lim, K., Jasim, A. & Ahmad, H., 2010. Fabrication of tapered fiber
based ring resonator. , Laser Phys. 20 (7), p. 1629–1631.
Hosford, W. F., 2010. Mechanical behavior of materials. s.l.:Cambridge University
Press.
Hu., H. et al., 2013. Generation and Field Trial Transmission. Journal of lightwave
technology, 31(99), p. 696‒701.
Huang, Y., Z. Zhuo, Y. Z. & chen, G., 2010. A Temperature Self-Compensated LPFG
Sensor for Large Strain Measurements at High Temperature. IEEE Transactions on
Instrumentation and Measurement, 59(11), pp. 2997 - 3004.
Jian, N. W. & Jaw, L. T., 2010. Feasibility of Fiber Bragg Grating and Long-Period
Fiber. Sensors 2010, 10,, pp. 10105-10127.
Juarez, C. J. & Taylor, H. F., 2005. Distributed fiber optic intrusion sensor system for
monitoring long. In Proceedings of the SPIE, Sensors, and Command, Control,
Communications, and Intelligence (C3I) Technologies for Homeland Security and
Homeland Defense IV, 5778, p. 692‒703.
Ju, S. M., Pramod, R., Watekar & Han, W. T., 2009. Highly Sensitive Temperature
Sensor Using Fiber Bragg Grating on Pb/Ge-Codoped Fiber. OSA Technical Digest
(CD) (Optical Society of America, 2009), paper JWA10.
Kersey, A. D., Berkoff, T. A. & Morey, W. W., 1993. Multiplexing fiber Bragg grating
strain-sensor system with a fiber Fabry-Perot wavelength filter. Pot. Lett., 18(16),
pp. 1370-1372.
Kinet, D. et al., 2014. Fiber Bragg Grating Sensors toward Structural Health
Monitoring in Composite Materials: Challenges and Solutions. Sensors, pp. 7394-
7419.
Kinet, D. et al., 2014. Fiber Bragg Grating Sensors toward Structural Health
Monitoring in Composite Materials: Challenges and Solutions. Sensors 2014, 14(4),
pp. 7394-7419.
Kondo, Y., Nouchi, K. & Mitsuyu, T., 1999. Fabrication of Long-period fiber grating
by focused irradiation of infrared femtosencond laser pulses. Optics Letters Vol. 24,
No. 10, pp. 646-648.
Kruezer, M., 2003. Strain Measurement with Fiber Bragg Grating Sensor, Darmstadt:
HBM measurement.
Lee, B. H. et al., 2012. Interferometric Fiber Optic Sensors. Sensors 2012, Volume 12,
pp. 2467-2486.
43
Li, B. et al., 2011. Ultra-Abrupt Tapered Fiber Mach-Zehnder Interferometer Sensors.
Sensors 2011, 11, pp. 5729-5739.
Liu, Y. & Wei, L., 2007. Low-Cost High-Sensitivity Strain and Temperature Sensing
Using Graded-Index Multimode Fibers. Physics and Computer Science Faculty
Publications.
Lu, P., Men, L. Q., Sooley, K. & Chen, Q. Y., 2009. Tapered fiber Mach–Zehnder
interferometer for simultaneous measurement. APPLIED PHYSICS LETTERS 94,
131110.
Lu, P., Men, L., Sooley, K. & Q., C., 2009. Tapered fiber Mach-Zehnder
interferometer for simultaneous measurement of refractive index and temperature.
Appl. Phys. Lett. 2009, 94, No. 131110..
Lu, T. C., Lou, K. Y., Porter, G. A. & Chen, N. K., 2013. The effects of geometric
design on the performance of abrupt-tapered fiber Mach-Zehnder interferometers.
EEE Conference Publications.
Mihailov, S. J., 2012. Fiber Bragg Grating Sensors for Harsh Environments. Sensors
2012, 12, , pp. 1898-1918.
Mizuno, Y. et al., 2015. Brillouin scattering in multi-core optical fibers for sensing
applications. Scientific Reports, 5(11388 ).
Mokhtar, M. R., Owens, K., Kwasny, J. & Taylor, S. E., 2012. Fiber-Optic Strain
Sensor System With Temperature Compensation For Arch Bridge Condition
Monitoring. IEEE SENSORS JOURNAL, VOL. 12, NO. 5, pp. 1479-1486.
Mokhtar, M. R., Sun, T. & Grattan, K. T. V., 2011. Arch-bridge Lift Process
Monitoring by Using Packaged Optical Fibre Strain Sensors with Temperature
Compensation. Journal of Physics: Conference Series, p. 307.
Mokhtar, M. R., Sun, T. & Grattan, K. T. V., 2012. Bragg Grating Packages With
Nonuniform Dimension for Strain and Temperature Sensing. IEEE SENSORS
JOURNAL, VOL. 12, NO. 1, pp. 139-145.
Morales, A. M. & Lieber, C. M., 1998. A laser ablation method for the synthesis of
crystalline semiconductor nanowires. Science 279 (January (5348)), p. 208–211.
Puneet, S. & Himani, D., 2015. Designing of a Long Period Fiber Grating (LPFG)
Using Optigrating Simulation Software. SSRG International Journal of Electronics
and Communication Engineering (SSRG-IJECE) – Volume 2 Issue 6, pp. 18-23.
44
Qiu, Y., Sheng, Y. L. & Beaulieu, C., 1999. Optimal Phase Mask for Fiber Bragg
Grating Fabrication. JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 17, NO.
11, pp. 2366-2370.
Raji, R. M. et al., 2016. Intensity-modulated Abrupt Tapered Fiber Mach-Zehnder
Interferometer For The Simultaneous Sensing of Temperature and Curvature.
Opticas and Laser Technology, Volume 86, pp. 8-13.
Sharma, A., Engles, D. & Singh, A., 2012. Study of Long Period Grating as Strain
Sensor. International Journal of Engineering Research & Technology (IJERT), 1(4).
Shin, W. et al., 2009. Temperature insensitive strain sensor based on long period. IEEE
Photonics.
Sifta, R. et al., 2015. Distributed Fiber-Optic Sensor For Detection And Localization.
Metrology And Measurement Systems, Vol. Xxii (2015), No. 1, p. 111–118.
Sun, B. et al., 2010. Asymmetrical in-fiber Mach-Zehnder Asymmetrical in-fiber
Mach-Zehnder. Optical Society of America.
Swartz, R. A., Zimmerman, A. & Lynch, J. P., 2007. Structural Health Monitoring
System With The Latest Information Technologies. 5th Infrastructure &
Environmental Management Symposium.
Thévenaz, L., 2010. Brillouin distributed time-domain sensing in optical fibers.
Frontiers of Optoelectronics in China , Volume 3, pp. 13-21.
Wang, X. Z., 2012. Characterization of Fiber Tapers for Fiber Devices and Sensors.
Ottawa-Carleton Institute for Physics, University of Ottawa.
Wen, X. D., Ning, T. G., You, H. & Kang, Z. X., 2014. Analysis and Measurement of
the Displacement Sensor Based on an Up-tapered. CHIN. PHYS. LETT., 31(3).
Xia, T. H., Zhang, A. P., Gu, B. & Zhu, J. J., 2010. Fiber-optic refractive-index sensors
based on transmissive and reflective thin-core fiber modal interferometers.
Commun. 2010, 283, pp. 2136-2139.
Xing, X., Wang, Y. & Li, B., 2008. Nanofibers drawing and nanodevices assembly in
poly(trimethylene terephthalate). Opt. Express 16 (14), p. 10815–10822.
Yang, Y. W. & Hu, Y., 2008. Sensitivity of PZT Impedance Sensors for Damage.
Sensors, pp. 327-346.
Yin, S. Z., Ruffin, P. B. & Francis, T. S. Y., 2008. Fiber Optic Sensors Second Edition.
Boca Raton, Florida: CRC Press.
Zhong, X. Y. et al., 2014. High-Sensitivity Strain Sensor Based On Inflated Long
Period Fiber Grating. Optics Letters Vol. 39, Issue 18, pp, pp. 5463-5466.
45
Zhu, T., Rao, Y. J., Wang, J. L. & Song, Y., 2007. A highly sensitive fiber-optic
refractive index sensor based on an edge-written long-period fiber grating. IEEE
Photonics Technology Letters, 19 (24), pp. 1946-1948.
Zhu, T., Song, Y., Rao, Y. & Zhu, Y., 2009. Highly sensitive optical refractometer
based on edge-written ultra-long-period fiber grating formed by periodic grooves.
IEEE Sensors Journal, 9 (6), pp. 678-681.
46
APPENDICES
APPENDIX A: Derivation of the strain equation
Figure A.1: Dimension of packaging slab in curvature, C
(a)
(b)
Figure A.2: Segment of packaging slab for derivation demonstration
47
From Figure 0.2 (b),
𝑅
∆𝑥=
𝑅 − 𝑦
∆𝑠′
From Figure 0.2 (a),
Nature axis, ∆𝑠=∆𝑥
Hence,
∆𝑠′ =𝑅 − 𝑦
𝑅∆𝑠
By knowing that the strain is equated as,
𝜀 =∆𝑠′ − ∆𝑥
∆𝑠
Therefore,
𝜀 =𝑦
𝑅
48
APPENDIX B: The Back-tracing and Characterised data.
Measured Optical
Power Measured
Curvature
(m-1)
Characterised
Optical
Power
(𝜇𝑊m-1)
Optical
Power
difference
Traced-back
Curvature
(m-1)
Error
(%) (dBm) (𝜇𝑊m-1)
-38.40 0.144543977 0.000000 0.14588143 0.001337449 -0.000073 0.000000
-38.58 0.138675583 0.001500 0.13995873 0.001283149 0.001721 -14.748575
-38.68 0.135518941 0.002875 0.13614447 0.000625527 0.002687 6.554323
-38.90 0.128824955 0.003875 0.13182567 0.003000719 0.004734 -22.158857
-39.00 0.125892541 0.005625 0.12647363 0.000581094 0.005630 -0.096403
-39.14 0.12189896 0.006875 0.12105981 0.000839147 0.006852 0.338805
-39.36 0.115877736 0.008125 0.11668096 0.000803226 0.008693 -6.991633
-39.13 0.122179966 0.006875 0.12105981 0.001120153 0.006766 1.588765
-38.92 0.128233058 0.005625 0.12705741 0.001175648 0.004915 12.628137
-38.80 0.131825674 0.004375 0.13273945 0.000913772 0.003816 12.777033
-38.68 0.135518941 0.002750 0.13677288 0.001253941 0.002687 2.306795
-38.57 0.138995263 0.001500 0.14060475 0.001609489 0.001623 -8.231139
-38.38 0.145211162 0.000625 0.14621772 0.001006556 -0.000277 144.387120
Mean Optical
power
Difference 0.001196144