Autonomous measurement system for localization of loss-induced

perturbation based on transmission-reflection analysis

Vasily V. Spirin*, Serguei V. Miridonov, Enrique Mitrani, Carlos Morales, Mikhail G. Shlyagin,
Marcial Castro, and Manuel Manriquez

Division de Fisica Aplicada, CICESE, Apdo. Postal #2732, 22860, Ensenada, B.C. Mexico

ABSTRACT

The paper describes first completely autonomous measurement system based on transmission-reflection analysis (AMS-
TRA). The autonomous system utilizes simple optical scheme with 2 mW Fabry-Perot diode laser and original data
acquisition and processing system. The location of the loss region is determined from unique relationships between
normalized transmitted and Rayleigh backscattered powers for different positions of the disturbance along the test fiber.
Accuracy, temporal and thermal stability of the autonomous system with 5.6 km-length test fiber were investigated. The
paper also presents the preliminary results of the autonomous measurement system implementation for gasoline leak
detection and localization.

Keywords: Autonomous measurement system, distributed fiber optic sensor, transmission-reflection analysis, Rayleigh
scattering.

1. INTRODUCTION

The distributed bending-based fiber optic sensors are very attractive for the measurement of pressure,
temperature, displacement, etc., where the measurand is associated with a lateral deformation of the fiber [1-3]. The
regions where light losses occur due to bending are usually localized by means of optical time-domain (OTDR) [1] or
frequency domain reflectometry. The latter uses both coherent (COFDR) [4] and incoherent (IOFDR) [5-6] techniques.
All these methods utilize time- or frequency-modulated light sources that allow to localize a number of perturbations
along the test fiber simultaneously. Meanwhile, for some applications, it is important to detect and localize only rare but
hazardous alarm condition which typically occurs as a single infrequent event, such as a pipe leak, fire or explosion.
Recently, the authors have reported a novel effective and potentially inexpensive method for localization of a
loss-inducing perturbation in a few meter [7]- and a few km-length [8-9] optical fiber, based on the measurement of
transmitted and reflected, or Rayleigh backscattered powers of an unmodulated light source. The method can be
implemented for the detection of one or several temporally successive, alarm-like perturbations.
This paper presents first completely autonomous measurement system based on transmission-reflection analysis
(AMS-TRA). In the first two parts of the paper we present design of the AMS-TRA and discuss localization errors with
this method. In the next parts of the paper we examine the thermal stability of the equipment and present experimental
results.

-------------------------------------------------------------

Corresponding author
Tel.: +52 (646) 1745050 ext. 25328; fax: +52 (646) 1750554
e-mail address: vaspir@cicese.mx

Fiber Optic Sensors and Applications V, edited by Eric Udd, Proc. of SPIE
Vol. 6770, 67700H, (2007) 0277-786X/07/$18 doi: 10.1117/12.732747

Proc. of SPIE Vol. 6770 67700H-1

 2. AUTONOMUS MEASUREMENT SYSTEM DESIGN

The schematic diagram of AMS-TRA prototype is shown in Fig. 1. CW light emitted by a FP laser operating in
few longitudinal modes at 1550 nm with a total linewidth of few nm was launched into a standard single mode test fiber
SMF-28 through 3 dB coupler. The launched optical power was about 1 mW, and the measured with OTDR attenuation
coefficient of the test fiber was equal to 0.19 dB/km. The APC connectors were used to cancel back reflections influence
on the FP laser. Three power detectors with specially designed preamplifiers and filters were used to measure the
transmitted (T), Rayleigh backscattered (R), and laser (L) powers.
II
020 dB k 3_________
dB Input
4 aU Output
10dB [10dB
t

Interface
USB

Fig.1 Schematic diagram of fiber-optic sensor

After digitization data transferred to PC via USB interface. The data acquisition, signal processing, and the
indication of the event location on the map is done by specially designed software. Fig.2 shows the overview of AMS-
TRA system with short peace of sensing cable for hydrocarbon leak detection.

Fig.2 AMS-TRA prototype.

Proc. of SPIE Vol. 6770 67700H-2

 3. LOCALIZATION ERRORS WITH AMS-TRA

The noise from the photodetectors, variations of the Rayleigh-scattered light, and signal change due to light
interference from even weak reflection in the optical system cause errors in the calculation of the distance to the fiber
segment where the additional loss occurs. Here the analysis of these errors is presented. As shown in [8-9] the power
reflection coefficient of each Rayleigh scattering fiber segment can be calculated as:

R L = S ( s / 2 )[ 1 exp( 2 L )] , (1)

where s is the attenuation coefficient due to Rayleigh scattering, is the total attenuation coefficient of the test fiber,
L is the length of the fiber segment, and recapture factor S for the single-mode step index fiber is defined as:

S = b ( n 12 n 22 ) / n 12 , (2)

where b depends on the waveguide property of the fiber and is usually in the range of 0.21 to 0.24 for single-mode step-
index fiber, n1 and n2 are the refractive indices of the fiber core and cladding respectively.

Let us assume that a loss induced by a perturbation takes place at distance l from the beginning of the test fiber.
Introducing a parameter S = S(s/2), the transmission and backscattering coefficients of plain fiber section can be
written as T = exp(-l) and R = S[1-exp(-2l)] respectively. The short fiber piece is affected by a monitored condition
so that it introduces light loss and has a transmission TS 1. Let us assume that the scattering is relatively weak and the
portion of the scattered light is very small. This allows us to simplify the analysis, neglecting multiple scattering in both
directions. The Fresnel reflections with coefficients r1 and r2 from the fiber source- and remote-ends, respectively, have
to be taken into account, because even a weak reflection can be comparable to the back scattering. However, we can
assume that r1,2 <<1 and neglect multiple reflections as well. In this case, the transmission T and back-scattering R
coefficients of this optical system can be written as

T = T1T S T 2 = T S e - L , (3)

R = r1 + S ( 1 e 2 l ) + T 1 2 T S S ( 1 e 2
2 (Ll)
) + T 1 2 T S2 T 22 r 2 , (4)

where L is the total sensor length. Further simplification of R yields

R = S (1 T S2 ) e 2 l + ( r2 S ) T S e 2 L + S + r1 .
2
(5)

Normalized coefficients are defined as Tnorm = T/Tmax and Rnorm = R/Rmax, where Tmax and Rmax can be evaluated from the
above equations when TS = 1. This leads to:

T norm = T S , (6)
S + r1 ( S r2 ) T S2 e 2 L S (1 T S ) e 2 l
2
R norm = , (7)
S + r1 ( S r2 ) e 2 L

The relationship between the normalized transmitted Tnorm and Rayleigh backscattered Rnorm powers can be found from
Eqs. (6) - (7):

( S + r1 )( R norm 1) R norm ( S r2 ) e 2 L + S e 2 l
T 2 norm = . (8)
S ( e 2 l e 2 L ) + r2 e 2 L

Proc. of SPIE Vol. 6770 67700H-3

Neglecting Fresnel reflections from both fiber ends (r1 = r2 = 0) we can express the location of the perturbation as:

1 1 Tnorm
2

l (Rnorm , Tnorm ) = ln 2 L 2 L
(9)
2 1 Rnorm + Rnorm e Tnorm e
2

where 0 Rnorm 1 y 0 Tnorm 1

As we can see the perturbation location depends on two measurands Tnorm and Rnorm which can be defined experimentally
with some errors:

Tnorm Tnorm (10)

R norm R norm (11)

where Rnorm and norm standard deviations of normalized reflected and transmitted powers, respectively.

Let us assume that errors for Tmax and Rmax were significantly less than ones for T and R, because an averaging time for
an initial measurements can significantly exceeds the averaging time at normal-mode regime.
For independently measured reflected and transmitted powers the error in location can be found as:

l= [ lRnorm ]2 + [ lTnorm ]2 (12)

where lRnorm and lTnorm are errors due to uncertainty in measurements of normalized reflection and transmission
correspondently:

l
lRnorm = Rnorm (13)
Rnorm
l
lTnorm = Tnorm (14)
Tnorm

Substituting (13-14) in (12) and using (9) we can obtain:

l=
[(1 e 2 L
) R ] + [2T (e
norm
2
norm
2 l
)
e 2 L Tnorm ] 2

2 (1 T )e 2 2 l
(15)
norm

As we can see the location error depends on the distance where the perturbation takes place. The accuracy of localization
of excess loss strongly depends on the value of the induced loss. With the TRA method, it is easier to localize strong
perturbations, but the localization of a weak perturbation requires more accurate measurements of the transmitted and
Rayleigh backscattered powers.
Fig.3 shows the calculated with (15) localization error versus transmission for two different locations of the
perturbation. The variations of reflected Rnorm and transmitted powers norm were measured at 1 Hz bandwidth.

Proc. of SPIE Vol. 6770 67700H-4

 800

700 Loss at 1 m
Loss at 9,999 m

600

500

l (m) 400

300

200

100

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Tnorm

Fig.3. Calculated localization error versus transmission for Rnorm = 0.005 and norm=0.00385.

4. THERMAL STABILITY

As we found experimentally, considerable problem with AMS-TRA prototype is related with ambient temperature
influence. Fig. shows the variation of the measurands Tnorm, Rnorm and Lnorm during 14 hours. Here Lnorm is directly
measured FP laser intensity (see Fig.1) normalized on its initial value. Smallest variations of the measured values were
recorded during the night when the temperature was stable (see Fig. 4).

1.005
1
Tnorm

0.995
0.99 2 Tnorm = 2.3997 %
0.985
0 100 200 300 400 500 600 700 800

1.01
Rnorm

1
0.99
0.98 2 Rnorm = 4.7507 %
0.97
0 100 200 300 400 500 600 700 800

1.02
Lnorm

1.01 2 Lnorm = 3.016 %

0 100 200 300 400 500 600 700 800

Time (min)

Fig.4 Measurands variations during 14 hours.

Fig.5 shows the variations of the measurands due to local temperature variations on 10C at different locations inside
AMS-TRA. As we found one of the most critical parts are connectors. Note that it is possible to eliminate this problem
by using special temperature insensitive connectors or splicing fibers.

Proc. of SPIE Vol. 6770 67700H-5

 1
b. 2 Tnorm = 3.0671 % f.
0.99 e.

Tnorm
a. c.
0.98
d.
0.97
0 5 10 15 20 25 30 35 40 45

0.99 a. c. e.
Rnorm

0.98
0.97
d. f.
b.
0.96 2 Rnorm = 4.3038 %

0 5 10 15 20 25 30 35 40 45

1.04 c.
Lnorm

1.03 e.
a.
1.02 d. f.
1.01
b. 2 Lnorm = 3.8597 %
0 5 10 15 20 25 30 35 40 45
Time (min)

Fig.5 Local temperature influence near: a) connector in laser channel, b) preamplifier, c) output connector of FP laser, d)
connector in transmission channel, e) connector it reflection channel, f) test fiber.

5. EXPERIMENTAL LOCALIZATION ERROR MEASUREMENTS

To estimate the accuracy of the localization of perturbation with AMS-TRA prototype, we induced gradually
increasing perturbations near source- and remote-end of the test fiber. Fig.6 shows variation of measured perturbation
location with change of losses. Variation about 2 m for the location of the disturbance was recorded for the strong
losses that exceeds 3dB.

20
l1 (m)

-20
0 5 10 15 20 25 30

0.8
Tnorm

0.6
0.4

0 5 10 15 20 25 30

0.8
0.6
Rnorm

0.4
0.2
0 5 10 15 20 25 30

1.02
Lnorm

1.01
1

0 5 10 15 20 25 30
Time (min)

Fig.6 Localization errors for losses induced near source-end of test fiber.

Fig.7 shows measurands variations versus time from the moment when an electrical power was switch on. The
strong losses were induced near remote-end of 1.3 km-length test fiber.

Proc. of SPIE Vol. 6770 67700H-6

 The measurands became stable only after 0.5 hour (see Fig.7) when the temperature inside the prototype box
stabilized.

1280

l1 (m)
1260

1240

0 20 40 60 80 100 120

0.6
Tnorm

0.595

0.59
0 20 40 60 80 100 120

0.99
Rnorm

0.98

0.97

0 20 40 60 80 100 120

1.005
Lnorm

0.995
0 20 40 60 80 100 120
Time (min)

Fig.7 Measurands variations with time after power on.

Fig.8 shows calculated and experimentally measured localization errors for 1.3 km-length test fiber versus
normalized transmitted power Tnorm which is proportional to the induced losses. Experimentally measured errors for the
strong perturbations do not exceed 5 m for the losses induced near remote-end of the test fiber.

30

Loss at 1 m.
20
l (m)

10

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Tnorm

30

Loss at 1299 m.
20
l (m)

10

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Tnorm with R = 0.0035 and T = 0.0005

Fig.8 Calculated and measured localization error for 1.3 km test fiber.

Fig.9 shows variation of recorded with AMS-TRA disturbance distance for the gradual change of the loss value.
The losses were induced near the remote-end of 5.6 km-length test fiber. Note that normalized Rayleigh backscattered
intensity does not depend on value of losses.

Proc. of SPIE Vol. 6770 67700H-7

 5750

l1 (m)
5700
5650

0 5 10 15 20 25 30 35 40

0.8

Tnorm
0.6

0.4

0 5 10 15 20 25 30 35 40

1.01
1.008
Rnorm

1.006
1.004
0 5 10 15 20 25 30 35 40

1.02
Lnorm

1.015

1.01
0 5 10 15 20 25 30 35 40
Time (min)

Fig.9 Localization errors for losses induced near remote-end of 5.6 km test fiber.

Fig.10 shows calculated and experimentally measured localization errors for 5.6 km-length test fiber.
Experimentally measured errors for the strong perturbations do not exceed 10 m for the losses induced near remote-end
of the test fiber. Typical variations of the normalized reflected and transmitted powers due to various instability in the
prototype of the sensor were Rnorm = 0.0015 and norm= 0.0003 for signal bandwidth 1 Hz. These data were used for the
estimation of the errors in localization.

40

30 Loss at 1 m.
l (m)

20

10

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Tnorm

80

60 Loss at 5599 m.
l (m)

40

20

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Tnorm with R = 0.0015 and T = 0.0003

Fig.10 Calculated and measured localization error for 5.6 km test fiber.
.

Proc. of SPIE Vol. 6770 67700H-8

 6. CONCLUSION

We have designed completely autonomous measurement system based on transmission-reflection analysis. We

have demonstrated that the AMS-TRA prototype allows the localization of loss-inducing perturbation along a 5.6 km-
length sensing fiber with few meters error for the losses that exceeds 3dB.
We believe the proposed technique will be very attractive for the eventual realization of a compact and inexpensive
distributed alarm fiber optic sensor.

ACKNOWLEDGEMENTS

We gratefully acknowledge financial support through grant P45919-Y from the Consejo Nacional de Ciencia y
Tecnologa, and Project 23770 SEMARNAT, Mxico.

REFERENCES

1 D.A. Nolan, P.E. Blaszyk and E. Udd. Optical fibers, Fiber Optic Sensors: An Introduction for Engineers &
Scientists, John Wiley &Sons, New York, 1991.
2 G.L. Mitchell. Intensity-based and Fabry-Perot interferometer sensors, Fiber Optic Sensors: An Introduction for
Engineers & Scientists, John Wiley &Sons, New York, 1991.
3 T. Clark and H. Smith. Microbend fiber optic sensors, in Fiber Optic Smart Structures, edited by E. Udd: John
Wiley & Sons, New York, 1995.
4 K. Tsuji, K. Shimizu, T. Horiguchi and Y. Koyamada. Coherent optical frequency domain reflectometry for long
single-mode optical fiber using a coherent lightwave source and an external phase modulator, IEEE Phot. Technol.
Lett. 7, pp. 804806, 1995.
5 S.G. Pierce, A. MacLean, and B. Culshaw. Optical frequency-domain reflectometry for microbend sensor
demodulation, Appl. Opt. 39, pp.4569-4581, 2000.
6 S. Venkatesh and D.W. Dolfi. Incoherent frequency modulated cw optical reflectometry with centimeter
resolution, Appl. Opt. 29, pp. 1323- 1326, 1990.
7 Spirin V.V., Shlyagin M.G., Miridonov S.V., and Swart P.L., Transmission/reflection analysis for distributed
optical fibre loss sensor interrogation, Electronics Letters, 38, pp. 117-118, 2002.
8 Spirin V.V., Transmission-reflection analysis for localization of temporally successive multi-point perturbations in
distributed fiber-optic loss sensor based on Rayleigh backscattering, Applied Optics, 42, pp.1175-1181, 2003.
9 Spirin V. V., Mendieta F.J., Miridonov S. V., Shlyagin M. G., Chtcherbakov Anatoli A. and Swart Pieter L.,
Localization of a loss-inducing perturbation with variable accuracy along a test fiber using transmission-reflection
analysis, IEEE Photonic Techn. Lett., 16, pp. 569-571, 2004.

Proc. of SPIE Vol. 6770 67700H-9