Ece070512.qxd


The Journal of Engineering Research  Vol. 6, No.1, (2009)  46-50

1.  Introduction

The attractive features of fiber optic sensors such as
lack of sparking and immunity to electromagnetic fields
has given it a great advantage over electronic sensors in
many applications, including civil engineering, oil, gas
and power generation industries (Abad, et al. 2003).    In
addition to its usefulness in embedded sensors applica-
tions and the inherited attractive features of optical fiber
sensors, Fiber Bragg Gratings (FBGs) are driven by their
performance accuracy and versatility of the utilized tech-
nology (Agarwal, 2001 and Bourdoucen and Al-Lawati,
2006). 

Fiber optics communication signal carriers comprise
about 95 percent of worldwide transmission capacity
(Chan, et al. 1999).  The introduction of the well known
Wavelength  Division  Multiplexing  (WDM) and later the 
Dense WDM techniques gave the fiber optic communica-
______________________________________
*Corresponding authors e-mail: lawati@squ.edu.om

tions such a leading edge in providing astronomical band-
width capabilities of  more than 10 terabits per second on
a single optical fiber. In fact, the Japanese Nippon
Telegraph and Telephones Corporation has recently
announced on its website reaching a data transmission rate
of 14 Tbps over a 160 km optical fiber.

This attractive feature of WDM makes it popular for
applications other than communications such as multi-
plexing many FBG sensors simultaneously on a single
fiber. Such applications were demonstrated in numerous
published articles, eg. (Abad, et al. 2003 and Fan, et al.
2004).   Other multiplexing techniques to interrogate FBG
sensors were also reported eg. (Gebremichael, et al. 2005
and o-eland.com/sensorproducts/sensor_strain.html).  

In this paper, a study of the effect of integrating an FBG
sensor was conducted in a four wavelength WDM com-
munication system operating at 1550 nm region with 1 nm
spacing. The operating wavelengths are shown in Fig. 1
below.  Four wavelengths are used to study the mutual

Integration of FBG Strain Sensors in WDM Networks, Effects
on Quality Factor

Ali Al-Lawati* and Hadj Bourdoucen

Department of Electrical and Computer Engineering, College of Engineering, Sultan Qaboos University, P.O. Box 33, P.C. 123, 
Al-Khodh, Muscat, Sultanate of Oman

Received 12 May 2007; accepted 3 October 2007

Abstract: A study of the effect of integrating an FBG sensor in a four wavelength WDM communications system operating
at 1550 nm is presented. The simulations considered focus on the mutual effects of both the sensing and the communica-
tions systems. The effect of power levels of the interrogating optical source on the performance of the two systems is also
investigated under excitation levels of up to 10 dBm. The network layout used in the simulations is based on an actual opti-
cal link in Oman having a variety of spans. The results obtained at data rates of 2.5 and 10 Gbps with variable strains up to
±600  µs show a good tolerance in terms of quality of transmission for the two systems. However, the greater the strain val-
ues, the more noticeable are the degradations of transmission quality parameters of the communications system.

Keywords: FBG, Fiber optic sensor, WDM, Quality factor

 äGô©°ûàe èeO FBG  á«Fƒ° dG ä’É°üJ’G äɵѰT ™eWDMIOƒ÷G πeÉ©e ≈∏Y ÒKÉàdG h

 ø°ShOQƒH êÉM h *»JGƒ∏dG ≈∏Y

¬¬°°UUÓÓÿÿGG~æY πª©Jh ¬«Lƒe ∫GƒWG á©HQCG Ω~îà°ùJ »àdGh ≈LƒŸG ∫ƒ£dG IO~©àe ä’É°üJ’G º¶f ™e     áµHÉ°ûàŸG ¬jô°üÑdG  ±É«dG äGô©°ûà°ùe èeO ÒKÉàd á°SGQO åëÑdG Gòg Ω~≤j :

≈àM ÚeɶædG AGOG ≈∏Y »Fƒ° dG Q~°üŸG IQ~b iƒà°ùe ÒKÉJ øe ≥≤ëàdG ” ∂dòc .ä’É°üJ’G º¶fh QÉ©°ûà°S’G øe πµd ¬dOÉÑàŸG äGÒKÉàdG ≈∏Y IÉcÉÙG õcôJh .Îe ƒfÉf 1550

äô¡XG ~bh .i~ŸG ¬Yƒæàe ¬Yƒª› ≈∏Y …ƒà– »àdGh ¿ÉªY ‘ ájô°üH ±É«dG áµÑ°T ≈∏Y IÉcÉÙG ‘ áe~îà°ùŸG áµÑ°ûdG §«£îJ ~æà°ùjh .  äG~MƒdG 10  ¤G π°üj IQÉKEG iƒà°ùe

∂dP ™eh .ÚeɶædG Óµd ∫É°SQ’G ¬«Yƒf å«M øe G~«L áfhôe øjGΰS hôµj Ée 600    ¤G π°üj ∫É©ØfG Ò¨àe ™e     10 h 2^5 øe äÉfÉ«ÑdG ä’~©e ‘ â≤≤– »àdG èFÉàædG

.ä’É°üJ’G Ωɶæd ∫É°SQ’G IOƒL ‘ ¢ü≤f ¤G …ODƒJ ∫É©Øf’G ᪫b IOÉjR ¿G ßMÓ«a

áá««MMÉÉààØØŸŸGG  ääGGOOôôØØŸŸGG .√Oƒ÷G πeÉ©e ,»LƒŸG ∫ƒ£dG IO~©àe ,á«Fƒ° dG ±É«d’G äGô©°ûà°ùe ,áµHÉ°ûàŸG ájô°üÑdG êGôH ±É«dG äGô©°ûà°ùe :



47

The Journal of Engineering Research  Vol. 6, No.1, (2009)  46-50

effects of sensing and communications systems. These
four wavelengths or channels are referred to hereafter as
λ1,  λ2,  λ3 and  λ4 corresponding to wavelength of 1548.5,
1549.5, 1550.5, 1551.5 nm respectively. 

The simulations considered focus on the mutual effects
on both the sensing and the communications systems. The
effect of power levels of the interrogating optical source
on the performance of the systems is also investigated
under excitation levels of up to 10 dBm. The network lay-
out used in the simulations is based on an actual optical
link in Oman having a variety of spans operating at 2.5
and 10 Gbps. The strain sensor is located 35 km away
from the nearest optical network access point that consists
of an Optical Add and Drop Multiplexer (OADM). 

2.  Theory and System Description

The FBG sensor exhibits a shift of  ∆λg in the reflect-
ed Bragg wavelength from the central wavelength  λg due
to change in temperature and strain.

∆λg can be found using the following relation (Kang,
et al. 1998):  

(1)

Where, δl,δn, δ∆ and δT are the changes in the fiber,
length, index of refraction, optical grating period and tem-
perature respectively. Assuming that the temperature
effect is negligible, the Bragg wavelength shift due to a
homogenous and isotropic strain on the grating can be
expressed as (Lim, et al. 2001;  Madhav and Asokan,
2004):

(2)

Where,  pe is the effective elastooptic coefficient of the
fiber and ε is the applied strain.  pe can be found using
the following equation:

(3)

Where  µ is the Poisson's ratio and pij are FBG fiber
strain optic tensors.

The applied strain will cause a shift in the Bragg wave-
length,  λg  or λ2 (with 99% reflectivity), towards the
neighboring WDM channels  λ1 and  λ3. This in turn, will
cause degradation in the communications system mainly
to the immediate channels surrounding  λ2.  Since the fiber
dispersion curve is not perfectly flat and is a function of
the wavelength, the effect on the neighboring channels
will not be identical and therefore both surrounding chan-
nels need to be examined.

In order to study this effect, the optical network shown
in  Fig. 2  was used.

The sensing FBG wavelength is detected 35 km away
from Samail city. An optical ADM is implemented to add
or drop  λ2 to the WDM link. The interrogating signal of
the FBG is sent from Muscat and the response measured
at Muscat as well. The FBG sensor's dynamic range
assumed in the simulations is typical and is also commer-
cially available (Peng, et al. 2002). 

Each link on the network is constructed from an SMF
and a DCF fibers to compensate the chromatic dispersion
at 1550 nm. The signal attenuation in both fibers is com-
pensated using optical amplifiers. 

3.  Results and Discussion

The system layout shown in Fig. 2, was simulated
based on the well known Schrödinger. The equation was
solved using the Split Step Fourier Method.

The principle of the split-step Fourier method (SSFM)
technique is based on the fact that the propagation of a
pulse over the full length of optical fiber is considered by
dividing the total length of the fiber into small segments
in such a way that changes in the envelopes of optical sig-
nals can be considered sufficiently small. Within these
segments, the linear and the nonlinear operators in the
Schrödinger equation can be considered to act independ-
ently of each other.   Hence, the effect of propagation
along the fiber segment is determined by first using non-
linear operator followed by the linear operator. In this
way, the accuracy and efficiency of operator splitting
techniques depend on the way discretization is done in
time and spatial domains (Sinkin, et al. 2003). 

As for the parameters of simulation and simplifications
used, these are based on the fact that the strain is homog-
enous and isotopic. The factor pe in Eq. (2),  which is a
function of  neff, pij and  π, is having a numerical value of
approximately 0.22 and hence this equation simplifies to
(1-pe) ε = 0.78 ε.  This approximation, well known in the
literature was used in all simulations.  The Q-factor is then
derived from the solution of  Schrodinger equation.

For a non-linear dispersive and lossy Single mode and
dispersion compensated fibers used in the simulations, the
evolution of the profile of a slowly varying field is
obtained by solving the Schrödinger equation given by:

 

1547 1548 1549 1550 1551 1552 1553
-40

-35

-30

-25

-20

-15

-10

-5

0

W a ve length (nm)

P
o

w
e

r 
(d

B
m

)

FBG Wavelength
Changes

Figure 1.  Spectra of the four DM sources used in
simulation

g
n n

2 n l 2 n T
l l T T

Λ Λ
∆λ Λ ∆ Λ ∆

∂ ∂ ∂ ∂⎛ ⎞ ⎛ ⎞
= + + +⎜ ⎟ ⎜ ⎟∂ ∂ ∂ ∂⎝ ⎠ ⎝ ⎠

∆λg = λg.(1-pe)ε   

pe= (n
2

eff/2)[ p12 - µ (p11 + p12)]  



48

The Journal of Engineering Research  Vol. 6, No.1, (2009)  46-50

(4)

Where, A is the amplitude of the varying electric field
envelope,  β1 the inverse group velocity,  β2 and β3 are
the first and the second order group velocity dispersion, a
is the absorption coefficient, and  γ is the nonlinear coef-
ficient of the considered fiber. These values are taken for
standard SMF and DCF fibers at a center wavelength of
1550 nm. 

Note that the Q factor is approximated using mean val-
ues and standard deviations of the signal samples as indi-
cated by relation:

(5)

where m1, m0, σ1, σ0 are the mean values and standard
deviations of the signal samples when a "1" or a "0" is
received respectively." 

In addition to these parameters, the simulations have
been performed for WDM sources having constant power
of 0 dBm, whereas the interrogating source used for the
sensor was varied from 0 dBm to 10 dBm. Two different
bit rates were used to have a more realistic scenario trans-
mitting at 2.5 Gbps and 10 Gbps. Figures 3 to 6 Illustrate
the quality factor profiles as a function of strain for each
wavelength. Figure 3, shows a flat profile of the Q-factor
of the FBG sensor for the full range of strains considered
(ie. -600 µs to +600 µs) which is an indication of a good
performance. This flattened behavior of the Q-factor seen
for different power levels of the interrogating source indi-
cates also a good tolerance to power changes. These two
observations imply that integrating  the sensing system for
the above range of strains within a WDM communication
system is viable.  

However, the effects on the communications system
need to be observed by examining Figs. 4, 5 and 6 corre-
sponding to  λ1,  λ3 and  λ4 respectively. 

The upper curve of Fig. 4, quasi insensitive to strain

 ADM 

Bahla 
Muscat 

Izki Samail Ghala 

FBG 

Sensing device  

25 km 84 km  37 km 48 km 

  35 km 

Figure 2.  Physical communication system layout connecting Muscat to Bahla used in simulation to
study the effect of FBG sensors on communication system

2 3
23

1 2 2 3
A A i A A a

A i A A
z t 2 6 2t t

β
β β γ

∂ ∂ ∂ ∂
= − × − × × + × − + × ×

∂ ∂ ∂ ∂

1 0

1 0

m m
Q

σ σ
−

=
+

 

-600 -400 -200 0 200 400 600
5

6

7

8

9

10

11

Strain (us)

Q
u

a
lit

y 
F

a
ct

or
 (

d
B

)

Figure 3. Quality Factor Q, in dB, for tran smission 
performance of FBG ( channel 2) 
operating at a data rate of 10 G bps for 
different input source power ,,  €€: 0 dBm, o: 
3 dBm, ◊◊: 6 dBm, ∇∇ : 10 dBm, 

 

-600 -400 -200 0 200 400 600

6

8

10

12

14

16

18

20

22

24

26

28

Strain (us)

Q
u

la
tit

y
 F

a
ct

o
r 

Q
 (

d
B

)

Figure 4.  Quality Factor Q, in dB, for transmission 
performance  of channel 1 (shortest 
wavelength on Fig.  1) operating at a data  
rate of 2.5 Gbps. οο: system not  coupled to 
sensing system, €€: 0 dBm, ××: 3 dBm, ◊◊: 6 
dBm, ∇ : 10 dBm, 



49

The Journal of Engineering Research  Vol. 6, No.1, (2009)  46-50

changes, corresponds to the Q-factor of  λ1 prior to sys-
tems integration. This figure shows that integrating the
sensor system causes a minor drop of the Q-factor.
However, the overall flat response is still maintained. At
strains approaching -600 µs the curves drop slightly due to
the effect of shifted  λg approaching  λ1. The same obser-
vations are valid for Fig. 5 representing  λ3.

The faster drop of Q-factor for  λ3 at the neighborhood
of +600 µs is most probably due to non-uniform disper-
sion characteristics of the fiber and other nonlinear
effects.

Figure 6 corresponds to quality of transmission of λ4.
As expected, due to not having  λg adjacent to this wave-
length, the effect on the performance of this channel is
very minor. A flat response throughout the dynamic range

of the sensor can be clearly observed.
Eye Diagrams for the four wavelengths at an applied

strain of -200  µs and a power of 6 dBm for the interrogat-
ing source wavelength are shown in  Fig  7. The diagrams
exhibit good eye opening values for all wavelengths. The
relatively wider eye opening for  λ1 despite having the
Bragg shift towards it is mainly due to the fact that it is
operating at a data rate of 2.5 Gbps, whereas the remain-
ing three wavelengths operate at 10Gbps.

4.  Conclusions

A study of the effect of integrating an FBG strain sen-
sor in a four wavelength WDM communication system
operating at 1550 nm region is presented. The simulations
considered focused on mutual effects of both systems. The
effect of power levels of the interrogating optical source
on the quality factor of the two systems has also been
investigated considering a network layout in  Oman hav-
ing variety of spans. 

The simulations performed on the systems at data rates
of 10 Gbps and 2.5 Gbps for strains up to ±600  µs, exhib-
it a good tolerance for the integration operation.  The
obtained quality factors and eye diagrams for all wave-
lengths used and at different interrogating sensor source
values, suggest that the integration has a minor effect on

 

-600 -400 -200 0 200 400 600
6

8

10

12

14

16

18

20

22

Strain (us)

Q
u

a
lit

y 
F

a
ct

o
r 

(d
B

)

Figure 5.  Quality Factor Q, in dB, for transmission 
performance of channel 3  operating at a  
data rate of 10 Gbps. οο: system not 
coupled to sensing system, €€: 0 dBm, ××: 3 
dBm, ◊◊: 6 dBm, ∇ : 10 dBm, 

 

-600 -400 -200 0 200 400 600

6

8

10

12

14

16

18

20

22

Strain (us)

Q
u

a
lit

y 
F

a
ct

o
r 

(d
B

)

Figure 6.  Quality Factor Q, in dB, for transmission 
performance of channel 4  operating at a  
data rate of 10 Gbps. οο: system not 
coupled to sensing system, €€: 0 dBm, ××: 3 
dBm, ◊◊: 6 dBm, ∇ : 10 dBm, 

 

  0.8ns
0.2ns 

 

 

 

 

 

0.2ns 0.2ns 
(c)                                           (d)

 
Figure 7.    Eye Diagrams for the four wavelengths 

for an applied strain of -200 µµs and a 
power of 6 dBm for the interrogating 
source. Eye diagrams for wavelengths 1 
to 4 are respectively sho wn in (a 
through d) .  Note that the FBG 
response is shown in Fig. b)  with a 
value of Q of about 10.3 dB 

(a)                                            (b)



50

The Journal of Engineering Research  Vol. 6, No.1, (2009)  46-50

both systems.  This suggests the possibility of utilizing
existing communications systems for FBG sensing appli-
cations.

References

Abad, S., Araújo F. M., Ferreira, L. A., Santos, J. L. and
López-Amo, M., 2003, "Interrogation of Wavelength
Multiplexed Fiber Bragg Gratings Using Spectral
Filtering and Amplitude-to-Phase Optical
Conversion",  J. Lightwave Technol., Vol. 21(1),  pp
127-131.

Agrawal, G., 2001, "Applications of Non Linear Fiber
Optics", Academic Press.

Bourdoucen, H. and Al Lawati, A., 2006, "Effect of
Dispersion Compensation on FBG Sensing Systems
Performance",  J. of Instruments and Experimental
Techniques, (Springer), Vol. 49(3),  pp. 403-407. 

Chan, P., Gong, W.J.M. and Demokan, M., 1999,
"Multiplexing of Fiber Bragg Grating Sensors Using
an FMCW Technique", IEEE Photonics Technology
Letters, Vol. 11(11), pp. 1470-1472.

Fan, Y. and Kahrizi,  M., 2004, "Applicability of an
Optimized Fiber Optic Smart Structure", Proceedings
of the International Conference on MEMS, NANOs
and Smart Structures (ICMENS'04), pp. 386-392.

Gebremichael, Y. M., Li W., Meggitt, B. T., Boyle, W. J.
O., Grattan, K. T. V.,  McKinley, B., Boswell, L. F.,
Aarnes, K. A., Aasen, S. E., Tynes, B., Fonjallaz, Y.
and Triantafillou, T., 2005, "A Field Deployable,

Multiplexed Bragg Gratings Sensor System Used in
an Extensive Highway Bridge Monitoring Tests",
IEEE Sensors Journal, Vol. 5(3), pp. 510-519.

Kang, S., Kim, S., Lee, S., Kwon, S., Choi, S. and Lee, B.,
1998, "Temperature-Independent Strain Sensor
System Using a Tilted Fiber Bragg Grating
Demodulator ", IEEE Photonics Technology Letters,
Vol. 10(10),  p. 146.

Lim, J., Yang, Q., Jones, B. and Jackson, P.,  2001, "Strain
and  Temperature Sensors using Multimode Optical
Fiber Bragg Gratings and Correlation Signal
Processing", Technology Conference, Budapest,
Hungary, pp. 1463-1466.

Madhav, K. and Asokan, S., 2004, "Spectrum Estimation
by Wavelength Shift Time-Stamping in a Fiber Bragg
Grating Sensor",  IEEE Photonics Technology
Letters, Vol. 16(5),  p. 1355.

Peng, P., Tseng, H. and  Chi, S., 2002, "Self-Healing Fibre
Grating Sensor System using Tunable Multiport Fibre
Laser Scheme for Intensity and Wavelength Division
Multiplexing", ELECTRONICS LETERS,  Vol.
38(24), pp. 1510-1512.

Sinkin, V., Holzlohner, R., Zweck, J. and Menyuk, C.R.,
2003, "Optimization of the Split-Step Fourier Method
in Modeling Optical-Fiber Communication Systems,"
J. Lightwave Technol., Vol. 21,  pp. 61-68.

Warf, B., 2006, “The Professional Geographer", Vol.
58(1), pp. 1-11.

www.o-eland.com/SensorProducts/sensor_strain.html.