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