Power modulation based optical fiber
loop-sensor for structural health
monitoring in composite materials
Nikhil Gupta and Kevin Chen
Mechanical and Aerospace Engineering Department
New York University, Polytechnic School of Engineering
Brooklyn, NY 11201
SysInt 2014, Bremen, Germany
1
List of Publications and Patents
• The technologies covered in this work are presented in the following
– Patents:
• Fiber-optic extensometer, US Patent #8,428,400, April 23, 2013, Nikhil Gupta,
Nguyen Q. Nguyen.
• Method for measuring the deformation of a specimen using a fiber optic
extensometer, US Patent #8,649,638, February 11, 2014, Nikhil Gupta, Nguyen Q.
Nguyen.
– Papers:
• Nishino, Z., Chen, K., and Gupta, N. Power Modulation Based Optical Sensor for
High Sensitivity Vibration Measurements. IEEE Sensors, 2014, (7): p. 2153 - 2158.
• Nguyen, N. Q. and Gupta, N. Whispering gallery mode sensor for phase
transformation and solidification studies. Philosophical Magazine Letters, 2010.
90(1): p. 61-67.
• Nguyen, N. Q. and Gupta, N., Analysis of an encapsulated whispering gallery
mode micro-optical sensor. Applied Physics B: Lasers and Optics, 2009. 96(4): p.
793-801.
• Nguyen, N. Q. and Gupta, N., Power modulation based fiber-optic loop-sensor
having a dual measurement range. Journal of Applied Physics, 2009. 106(3),
#033502.
2
Introduction
• Structural Health Monitoring (SHM)
A process of identifying one or more of
–
–
–
–
Load applied or displacement obtained on the structure
Extent of damage
Growth rate of damage
Performance of the structure as damage accumulates
• SHM can help in moving from predictive maintenance to
need-based maintenance
– Increase in safety
– Cost saving
Whispering Gallery Mode Sensors
Scanning laser

Optical fiber
r0
Photodiode
Sensor
Sensor
Fig A: Schematic of embedded sensor
• Tunable laser is used
• Evanescent field of the stripped
off section of fiber interacts with
that of the resonator (particle)
• Coupling back of the evanescent
field in the fiber gives resonance
peaks, which can be tracked
4
Whispering Gallery Mode Sensors
• Very high sensitivity
2π r n ≈  l
5
( = integer)
n  r  l
n
r
l
l
Transmission
– Detection of single chemical
molecules
– Detection of a single HIV
virus
– Measurement of subnanometer displacement
For r >> l, resonance condition:
l1
l2

n = refractive index of the microsphere
l = wavelength
r = micro-sphere radius
WGM Sensors: Effect of
Refractive Index
n1  n0  C1 1  C2  2   3 
n2  n0  C1 2  C2  1   3 
• Sensitivity comes at a
price!
n3  n0  C1 3  C2  1   2 
Where
n0 undeformed index of refraction
1, 2 and 3 are principal stresses
C1 and C2 are elasto-optic coefficients of the
material of the sphere.
– Signal to noise ratio can
be low
– Keeping the particle in
resonance can be difficult
Silica (Yves Belouard et al.
PMMA (Feridun et al. 2004)
2006)
C1 (m2/N)
-4.22 x10-12
-12 x10-12
C2 (m2/N)
-0.65 x10-12
-12 x10-12
1.467
1.4876
n0
6
Introduction
Applied force
Input laser light
Power losses at each fiber bend
to detector
• Microbend sensors
– Use multi-mode fiber
– Require high power light source
– Normally used under compression
– Large size
7
Transmitted power
optical
fiber
Displacement
Results and Discussion
• Power attenuation
Rc  20
• Critical radius (Jeunhomme, 1983)

l
2.748

0.996

3/ 2 
l
 n  
c 
l
where
difference
• For present single-mode optical
0.8
0.6
0.4
0.2
0
fiber
0
l=1.31 µm, lc=1.26 µm, n=0.0058
8
Pcurved/Pstraight
l is the operating wavelength
lc is cut-off wavelength
n: core-cladding index of refraction
1
Rc=11.8 mm
3
6
9
Loop radius (mm)
12
3
Fiber-loop sensors
• Power transmission due to curvature
PR 
Pout
Pin
– Pout is transmitted power through the loop
– Pout is power incoming to the loop
• Compressing loop creates more losses,
relative transmitted power
P 'out
P
Pout
9
– P’out is transmitted power with the applied
force
– Pout is power with no load applied
Fiber-loop sensors
• Compression of loop RB=7 mm
80
1
70
P'out/Pout
0.8
0.6
0.4
60
Force (mN)
Loading
Unloading
50
40
Loading
Unloading
30
20
0.2
10
0
0
0
2000
4000
Displacement (m)
6000
0
2000
4000
Displacement (m)
coating
6000
core
• Resonances occur between leaky mode
reflected from cladding/coating interface
and fundamental mode
10
RB
Radiation caustic
cladding
Fiber-loop sensors
• Pure bend loss-Marcuse model
Assumption: infinite cladding, large bend radius, weakly guided index fiber
P R  exp  2 l
where
e
B B

1/ 2
 2 3 RBe 
1  
2
2 B   3 e 
exp  
2 
2   RB  V 2 K12  a 
3

0 

k  2 / l
V  ak  n  n
  k n  
2

2 1/ 2
cl
2
co
lBe  2RBe


2 1/ 2
0
2
co
  0  k n
2
2 1/ 2
cl
SMF28e from Corning, NY
11
Fiber layer
Radius (m)
Index of refraction
Core
4.1
1.4517
Cladding
62.5
1.447
Coating
125
1.4786
nco and ncl are indices of refraction of the
core and cladding
0 is the propagation constant in straight
fiber, solved by the eigenvalue equation
J1   a 
H11  i a 

 i 1
J 0  a 
H 0  i a 
ReB is effective bend radius, differing from
RB by a stress correction factor, taken 1.28
for SMF28e fiber
Fiber-loop sensors
• Renner model- finite, coating and cladding thickness
P R  exp  2 BC lBe 
where
2  Z ct Z cl 
 Zct  Zcl    Zct  Zcl  cos  20 
2 BC  2 B
0.8
Pout/Pin
1/ 2
1
0.6
Zcl  k 2 ncl2 1  2b / RBe  02
 Rc


1
 e 
 RB 

4bRBe  Rc
 e  1
3Rc  RB 
3/ 2
2
2k 2 ncl2 b
Rc 
2
2m  1/ 2 for maximum

2m  3/ 2for minimum
, m is an integer
Rc is the critical radius
12
4
6
8
RB (mm)
10
12
0.8
leB =2 ReB is the effective length of the loop
•
experimental
0
2
0
Experimental data are obtained by changing the radius of
fiber-loop
0.6
Pout/Pin
 3 RBe
0  2 2
3k ncl
e
B
3/ 2
Renner model
0.2
Zct  k n 1  2b / R  
2 2
ct
Marcuse model
0.4
0.4
0.2
Marcuse model
Renner model
0
5 5.3 5.6 5.9 6.2 6.5 6.8
R (mm)
Loop sensor calibration setup
• Square wave signal is sent to the loop
• Photodetector tracks the transmitted power
• Relative transmitted power and force are monitored with respect to
increment in displacement
Translation motor
Load cell
Photodetector
Load cell
Loop sensor
Optical fiber
13
Translation stage
Laser
Translation
stage
Loop sensor calibration
• Calibration of different loop radii
80
1
R= 5 mm
70
0.8
60
Force (mN)
P'out/Pout
R= 8 mm
0.6
R= 7 mm
0.4
0.2
40
R= 8 mm
30
10
R= 6 mm
0
R= 6 mm
50
20
R= 5 mm
0
R= 7 mm
2000 4000 6000
Displacement (m)
0
8000
0
2000 4000 6000
Displacement (m)
8000
• Smaller loops have higher sensitivity but lower measurement range
• Loop-sensors allow large deformation without losing its elasticity
and repeatability
14
Loop sensor calibration
0.9
1
0.8
R= 8 mm
36
0.8
0.75
34
0.7
32
RB =6 mm
RB =5 mm
P'out/Pout
0.65
0.6
0.6
1000
R= 7 mm
0.4
0.2
R= 5 mm
1.1
R= 6 mm
0
2000 4000 6000
Displacement (m)
8000
• Resolution
– Force: 10-4 N
– Displacement: 10-5 m
30
1100
1200
Displacement (m)
P'out/Pout
1.05
P'out/Pout
0
15
Force
38
Force (mN)
P'out/Pout
0.85
P'out/Pout
Force (mN)
• In high sensitivity domain
Force
32
31
1
30
0.95
29
0.9
28
y = -0.0031x + 2.6173
0.85
490
510 530 550 570
Displacement (m)
27
Cyclic loading tests
• Pear-shaped loop and experimental setup
Optical fiber
2R0
Hollow tube
Oscilloscope
Data acquisition
Amplified
photodetector
Translation stage
Optical fiber
16
Load cell
Laser
Cyclic loading tests
• Results in 10,000 cyclic loading
P'out loading
unloading 100
P'out (a.u.)
P’out (a.u)
6
80
• Loop radius: 5 mm
60
• Displacement: 6 mm
4
40
2
20
Force
0
196263
0
Forcee (mN)
8
• Displacement rate: 0.4 mm/s
•
30 s per loading/unloading
cycle
196303
196343
Time (s)
• Total testing time: 4 days
•
The sensors survived after 10,000 cycles
•
Results show repeatability and
consistency for 104 loading/unloading cycles
17
Cyclic loading tests
• Different displacement rate
1.4
60
1.2
60
1
80
1
40
40
0.8
20
0.8
20
0.6
0
0.6
0
0
800
1600
Time (s)
2400
0
v=0.01 mm/s
P'out/Pout
Force
400
Time (s)
80
1.4
60
1.2
60
1
P'out/Pout
Force
80
40
40
0.8
20
0.8
20
0.6
0
0.6
P'out/Pout
1
P'out/Pout
1.2
200
0
40
80
Time (s)
v=0.2 mm/s
120
0
0
20
40
Time (s)
60
v=0.4 mm/s
• Loop radius: 6 mm
• Displacement: 6 mm
v=0.05 mm/s
Force (mN)
1.4
18
Force
Force (mN)
P'out/Pout
P'out/Pout
P'out/Pout
1.2
80
Force (mN)
Force
Force (mN)
P'out/Pout
SHM of laminated composites
• Loop sensors bonded to laminated composites under flexural loading
Oscilloscope
Amplified
photodetector
Single-mode
laser
Glass fabric
laminate
Full
surface
bonded
19
Bonded at
two locations
Fiber-loop
sensor
Pre-compressed loop
SHM of laminated composites
4
2
P'out/Pout
P'out/Pout
Force
0.9
0
0
4
0.94
0.92
Force
0.995
0.96
0
2000
4000
Deflection (m)
RB =4.9 mm
P'out/Pout
Force
6
5
2
0.997
1
0
20
2000
4000
Deflection (m)
RB =6.2 mm
1.02
8
P'out/Pout
Force
6
unloading
1
4
0.98
2
0.96
-0
0
6
0.999
Force (N)
P'out/Pout
3
0
8
1
4
0.996
loading
RB =5.9 mm
0.999
0.998
0
1000 2000 3000
Deflection (m)
P'out/Pout
1
2
4
0.998
0.997
P'out/Pout
2
Force
0.996
0
0
2000
4000
Deflection (m)
RB =6.5 mm
Force (N)
1.005
6
0.5 1 1.5 2
Time ( × 1000 s)
Force (N)
Force (N)
1.015
0.98
P'out/Pout
6
Force (N)
1.025
8
1
P'out/Pout
8
P'out/Pout
1.035
2.5
Quasi-static loading on
loop of radius 6 mm
Vibration Measurement
21
Optical fiber loop sensor
setup for calibration of
vibration measurement
The setup used for measuring
the free vibration characteristics
of a composite material.
Vibration Measurement
(a)
(b)
(c)
(d)
• The Vibration
measurements
are accurate
and match with
the frequency
of the shaker
• No fatigue or
hysteresis is
observed for
over 10,000
cycles
Results and Discussion
• The system is tested with and without optical fiber sensor using
only a PSD
• Then the output of the sensor is related to the PSD measurements
23
Conclusions
• A low-cost, high sensitivity loop-sensor has been
developed for stress or strain measurement
• The sensor can be used in dual measurement
ranges for displacement
• The sensor shows survivability in large number
of loading cycles
• Use of loop-sensor for vibration measurement is
possible
• Potential applications in chemical sensing
24
Acknowledgements
• National Science Foundation grant # CBET
0809240/ 0619193
• Environmental Protection Agency: Smart
Fellowship to Kevin Chen for chemical sensing
• Zachary Nishino, Dr. Nguyen Q. Nguyen
• Dr. Volkan Otugen’s group at Southern Methodist
University, Dallas
25

Fiber Optic Connectivity - Z

Normal and Reverse Flow Injection

CTE

レーティング換算表 - ZEUS Entertainment OFFICIAL WEB SITE