An Intrinsic Fiber-Optic Sensor for Structure
Lightning Current Measurement
Carlos T. Mata and Angel G. Mata
Truong X. Nguyen, Jay J. Ely and
George N. Szatkowski
ESC - Kennedy Space Center, FL 32899 U.S.A
NASA Langley Research Center
Hampton, VA 23681 U.S.A.
truong.x.nguyen@nasa.gov
Gary P. Snyder
NASA - Kennedy Space Center, FL 32899 U.S.A.
Abstract— An intrinsic optical-fiber sensor based on Faraday
Effect is developed that is highly suitable for measuring lightning
current on aircraft, towers and complex structures. Originally
developed specifically for aircraft installations, it is light-weight,
non-conducting, structure conforming, and is immune to
electromagnetic interference, hysteresis and saturation. It can
measure total current down to DC. When used on lightning
towers, the sensor can help validate other sensors and lightning
detection network measurements.
Faraday Effect causes light polarization to rotate when the
fiber is exposed to a magnetic field in the direction of light
propagation.
Thus, the magnetic field strength can be
determined from the light polarization change. By forming
closed fiber loops and applying Ampere’s law, measuring the
total light rotation yields the total current enclosed.
A
broadband, dual-detector, reflective polarimetric scheme allows
measurement of both DC component and AC waveforms with a
60 dB dynamic range.
Two systems were built that are similar in design but with
slightly different sensitivities. The 1310nm laser system can
measure 300 A - 300 kA, and has a 15m long sensing fiber. It was
used in laboratory testing, including measuring current on an
aluminum structure simulating an aircraft fuselage or a lightning
tower. High current capabilities were demonstrated up to 200
kA at a lightning test facility. The 1550nm laser system can
measure 400 A - 400 kA and has a 25m fiber length. Used in field
measurements, excellent results were achieved in the summer of
2012 measuring rocket-triggered lightning at the International
Center for Lightning Research and Testing (ICLRT), Camp
Blanding, Florida. In both systems increased sensitivity can be
achieved with multiple fiber loops.
The fiber optic sensor provides many unique capabilities not
currently possible with traditional sensors. It represents an
important new tool for lightning current measurement where low
weight, complex shapes, large structure dimension, large current,
and low frequency capabilities are important considerations.
Keywords—lightning; Faraday Effect; current sensor; fiberoptic; aircraft; direct measurement;
I.
INTRODUCTION
Accurate characterization of lightning return stroke current
is important in protection against direct and indirect effects.
As described in literature, methods to characterize lightning
currents typically include indirect remote measurements using
various field sensors, or direct measurements using rockettriggered setups and instrumented towers. Direct measurement
also includes flight campaigns characterizing aircraft in-flight
lightning environment [1,2]. Data from direct measurements
are much more limited but are more accurate and often used in
validation of indirect measurement setups.
Direct measurement typically utilizes traditional sensors
such as resistive current shunt, Rogowski coil, current
transformer, or time-derivative sensor (i.e. I-Dot), each with its
own advantages. However, depending on applications these
sensors may have installation restrictions, such as size and
weight, or have performance limitations such as amplitude
range, frequency, bandwidth, saturation or hysteresis. These
limitations can be difficult to overcome especially for aircraft
external installations where there are considerations for
aerodynamic performance, safety, weight, and installation cost.
An on-going effort to characterize aircraft lightning
environment led to the development of a fiber-optic current
sensor tailored for direct lightning measurement. It has most of
the advantages of traditional sensors and with few
disadvantages. Desirable characteristics are many, including
being light weight, flexible, and conforming to structure
geometries. It is self-integrating and can measure total current
(not the time derivatives) including the DC component. It does
not suffer from self-resonance, saturation or hysteresis as with
Rogowski coil or current transformer. It is not susceptible to
arcing/sparking from high voltage and current. Being nonconductive, the sensor can be safely routed directly into an
aircraft fuselage or a control room.
Traditional B-Dot
Sensors
Proposed Optical
Fiber Sensor
NASA Illustration 2012
Fig. 1. Illustration of fiber-optic current sensors on aircraft.
Sensor installation is simple and non-intrusive, simply by
wrapping the thin fiber one or more times around the structure
to be measured. Versatility is excellent, as the same sensing
fiber can be used on small or large structures. The results
would be accurate in both cases given sufficient bandwidth and
length. Measurement sensitivity can be increased by using
multiple fiber-turns around the conductor.
Aircraft installation would benefit the most from this new
sensor due to strict size, weight, aerodynamic performance and
safety requirements.
Fig. 1 illustrates possible fiber
installations measuring current on structures such as fuselage,
wings, or tail sections. To date the sensor has not yet been
flown on an aircraft due to typical high costs for any flight
experiment, and that most aircraft would avoid flying into
thunderstorms intentionally. Regardless, this sensor represent
a significant leap in capabilities relative to traditional sensors
used in past flight experiments.
The sensor’s many
advantages could benefit other direct lightning measurements
such as on small buildings, windmills, or lightning towers as
illustrated in Fig. 2.
The sensor is based on Faraday (Rotation) Effect, which
causes light’s polarization plane in a medium to rotate when
the medium is exposed to a magnetic field in the direction of
light propagation. Using optical fiber as the propagation
medium and by forming close fiber loops, measuring the total
light rotation would result in
the total current enclosed. It
is noted that the sensing
element is an optical fiber,
thus
termed
“intrinsic”
sensor. In contrast, in
“extrinsic” sensor optical
fiber is only used for
relaying signal from a
remote sensor.
The sensor is not without
limitations. Fiber choice is
limited.
Most commonly
available fibers are based on
silica, and
the Faraday
Effect in silica is weak.
However, this makes the
sensor highly suitable for large currents such as in lightning.
There are slight temperature and bend/vibration sensitivities,
though there are approaches to compensate, in real-time or
through post processing, that could lead to very precise
measurements. Glass fiber is also fragile and needs suitable
protection.
This paper describes two sensor systems that are similar in
design and characteristics, only with slightly different
measurement ranges. Their lasers operate at 1310nm and
1550nm wavelengths where optical components are relatively
commonly available. In this paper the 1310nm system was
used in laboratory demonstrations while the 1550nm system
for field measurement of triggered lightning. Their design is
described in the next section. In addition, measurement results
are reported for:
Equivalent current up to 300 kA, using multiple fiber
loops or a multi-turn current coil.
Current on a 1.2 m diameter round structure emulating a
miniature lightning tower or an aircraft fuselage.
Large current having 100 kA and 200 kA peaks,
performed at a commercial lightning test facility.
Triggered-lightning currents at the International Center
for Lightning Research and Testing (ICLRT), at Camp
Blanding, Florida.
These tests and results illustrate the ability to measure large
direct lightning current on structures. The fiber-optic current
sensor is simply referred to as Faraday sensor in the remainder
of the paper.
II.
FIBER-OPTIC CURRENT SENSOR
A. Basic Sensor Operation
As stated, Faraday Effect causes light polarization in an
optical medium to rotate when the medium is exposed to a
magnetic field in the direction of light propagation. The effect
in an optical fiber is illustrated in Fig. 3. The amount of
rotation depends on the material and the strength of the
magnetic field component. The polarization plane rotation, in
radians, is [3-10]:
𝜙=𝑉
𝑩 ∙ 𝑑𝒍 = 𝜇0 𝑉
(1)
where µ0 is the free-space permeability; V is the Verdet
constant in radians/(meter∙Tesla); µ0V is the combined
permeability Verdet constant (radians/ampere); B is magnetic
flux density in Tesla (T); length l (in meters) is the light and
𝜙
B
E
V
l
Fig. 2. Fiber-optic current sensors
loops on a lightning tower.
𝑯 ∙ 𝑑𝒍 ,
Fig. 3. Faraday Effect in optical fiber.
E
magnetic field interaction path length; and H is the magnetic
field (amperes/meter). For a fiber forming N closed loops
around a conductor carrying current I (ampere), applying
Ampere’s law yields
𝜙 = 𝜇0 𝑉
𝑯 ∙ 𝑑𝒍 ,
(2)
Thus, the rotation angle is directly proportional to the
current I and the number of loops N. The sensor is selfintegrating, and no additional integration is needed. Measuring
the rotation angle directly results in current, knowing the
number loops used.
B. Polarimetric Detection Scheme
This section describes the scheme used to measure the
polarization change induced by current. The scheme is
illustrated in Fig. 4 [4]. A linearly polarized light from a superluminescence diode (SLD) laser is generated at locations
labeled 1, 2. Half of the power is transmitted through the nonpolarizing beam splitter (NBS) at 3 to the sensing fiber at 4.
The sensing fiber forms closed loops around the current
carrying conductor at 5. A Faraday mirror at 6 rotates the
reflected light polarization by 90º relative to the incident light.
This helps cancel fiber bend/stress induced effects and makes
the sensor less sensitive to bending. The reflected light traces
back through the fiber to 3, at which half of the power is
reflected through the half-wave plate (HWP) at 7 toward the
polarizing beam splitter (PBS) at 8. Exiting the PBS, light
power in the two orthogonal polarizations are measured by two
photo-detectors D1 and D2 at 9. The HWP helps rotate and
align the initial polarization incident on the PBS. Ideally, at
zero current the incident polarization should be at 45º relative
to the PBS’s two orthogonal principle polarization axes, so that
beam power is divided equally between the two optical
detectors at 9. The difference in optical powers at 9 is
measured with a balanced detector.
This setup is referred to as a reflective scheme, since a
mirror is incorporated. Using this scheme in combination with
a Faraday mirror, as light travels through the fiber twice, the
non-reciprocal Faraday rotation is doubled while external
Fig. 5. Ideal sensor responses at 1310 nm.
stress-induced effects are subtracted [5].
The responses at the two detectors should ideally be
(a),(b)=0.5*[1±sin(4μ0VNI)] for a reflective scheme.
Mathematic operation difference-over-sum, (c)=(a-b)/(a+b),
yields
(c) =
sin (4μ0VNI), or
(3)
NI =
sin-1(c)/ (4μ0V),
(4)
where equivalent current NI (in unit Ampere-turn) is defined
as number of loops N times the current I, and µ0V = 1.01
μrad/A at 1310nm and 0.718 μrad/A at 1550nm [3].
It is important that light’s linear state-of-polarization is
maintained in the fiber during light transit. This is achieved
with proper fiber design. The systems in this paper use two
different commercial spun polarization-maintaining (PM)
fibers [6], which are the result of twisting PM fibers during
manufacturing.
Fiber twisting helps hold the state-ofpolarization that otherwise would be destroyed in a typical
fiber. The twist rate is about 5 mm per turn.
Fig. 5 illustrates the ideal responses at 1310nm, with the
response curves (a) and (b) being voltage outputs from the two
optical detectors. The difference-over-sum operation (c)=(ab)/(a+b) would yield a response that is more sensitive (higher
slope) than either response curve, with zero crossing at zero
current, and has larger dynamic range due to common-mode
noise subtraction. Current I or equivalent current NI is then
computed from (c) using eq. (4).
The typical operating range is where the curve (c) increases
monotonically in Fig. 5, or about -350 kA to +350 kA. In this
range the response and current correspond one-to-one. In the
systems described in this paper, non-ideal fiber medium and
optical components distort the curves. The practical range is
slightly reduced to about -300 kA to +300 kA as to be reported
in the next section.
The 1550nm-based system is slightly less sensitive due to
the lower Verdet constant at this wavelength, so it can measure
slightly larger current. The practical range is approximately
+/- 400 kA. The design, construction and characteristics
otherwise are very similar to the previously described 1310 nm
system.
Over-current would not damage the sensor - light
polarization would simply rotate beyond the intended range.
Fig. 4. Reflective polarimetric scheme with dual detectors.
The concern is that the solution to the sin-1 function in (4)
would be ambiguous. However, there are simple solutions that
permit measurement of very large current [7].
These
techniques are not necessary here since most known direct
lightning currents are below the 300 kA and 400 kA ranges of
the two systems described.
C. Sensor Calibration and Data Correction
The 1310nm system was measured in laboratory and the
results compared against reference sensors that include a
Rogowski coil (with an electronic integrator) and a ferritebased PearsonTM current transformer (CT). Fig. 6 compares
the three sensors by plotting current from the Faraday sensors
on the vertical axes against current from the reference sensors
on the horizontal axes. Large equivalent currents were
achieved with the help of multiple fiber turns or a wire coil.
Details are described in the next section.
Ideally the Faraday sensor data would fall on the straight
diagonal line labeled “ideal”. This line represents (1:1)
correspondence between the Faraday sensor and the two
reference sensors. Instead, the data follow the red curve
labeled “uncorrected”. This non-ideal response is due to the
reduced sensitivity in the fiber (relative to ideal) along with
light depolarization due to non-ideal fiber medium and optical
components. Additional details concerning light propagation
in spun fiber can be found in [7]. It is also clear that the
uncorrected sensor’s response is non-linear unless restricted to
low current. Thus, it is important that the sensor is calibrated
over its operating range and a correction function developed.
The correct function is developed from a simple
polynomial spline-fit curve (5th to 9th, odd order) that maps the
Faraday sensor response to the “ideal” curve. Once complete
and verified the same correction function can be applied to
subsequent measurements to achieve the corrected results. Fig.
6 shows the “corrected” response curve aligns well with the
“ideal” diagonal line. An alternative to curve-fitting is
interpolation; however, neither approach is perfect as some
small error may remain.
LABORATORY TEST RESULTS
It is difficult to achieve the full range test or calibration
levels up to 300 kA (or 400 kA for the 1550nm system) in a
laboratory setting. One acceptable approach is to produce the
associated optical effects by using multiple fiber loops and/or a
multi-turn conductor coil. Multiple fiber loops and wire turns
amplify the Faraday rotation beyond that produced by a single
fiber loop around a single conductor. The amplification factor
is the product of the numbers of fiber loops and wire turns
used. Fig. 7 illustrates a typical setup. Laboratory tests show
using multiple fiber loops, a multi-turn coil, or combinations
yield the same response curves.
For simplicity, N is redefined to be the product of the
number of fiber loops and the number of wire turns. The
product NI is simply referred to as equivalent current as
previously stated in Eq. (4).
Fig. 8 illustrates excellent results comparison between the
Faraday sensor and the reference sensors, with different
number of fiber turns and wire loops. The equivalent currents
NI are about 5 kA•turns and 300 kA•turns as shown. The same
calibration correction function was used and good results were
achieved in both cases. Since the reference Rogowski coil and
Pearson current transformer (CT) only measure current on one
wire-turn, their results are numerically scaled by the factor N
for the comparison. This practice is commonly used in optical
current sensing [3-10]. Similarly, NI up to 400 kA•turns was
demonstrated with the 1550 nm wavelength system. Equally
good results were achieved though not reported here.
In calibration setups to achieve high equivalent current,
using a wire coil having a high number of turns can distort the
injected waveform. This is illustrated by comparing Fig. 8(i) to
8(ii) - the pulse width in the former widens considerably. In
contrast, using a high number of fiber loops does not affect the
current waveform but would require a longer sensing fiber.
Ideal (1:1)
300
Corrected:
Faraday vs. Pearson
200
Faraday Current (kA)
III.
100
0
Uncorrected:
Faraday vs. Pearson
Faraday vs. Rogowski
-100
-200
Fig. 7. Using multi-turn coil and multiple fiber loops to achieve high
current effects.
-300
-300
-200
-100
0
100
Pearson & Rogowski Current (kA)
200
300
Fig. 6. The 1310nm system’s response curve, corrected and uncorrected.
6
N*Current (kA•turn)
5
Pearson CT
Rogowski Coil
Faraday Sensor
(i)
4
3
5 kA
2
1
0
-1
-2
-200
50
0
200
400
600
Time (sec)
(ii)
1000
1200
Fig. 9. Measuring large current with one fiber loop.
Pearson CT
Rogowski Coil
Faraday Sensor
0
120
-50
-100
80
-150
-200
300 kA
-250
0
50
100
Time (sec)
150
40
100 kA
0
200
-20
-50
Fig. 8. Good equivalent current (N*I) result comparisons using (i) 49turn coil and one fiber loop (N=49), and (ii) 3-turn coil and 28 fiber
loops (N=3*28=84).
Fig. 9 illustrates the test setup. The Faraday sensor fiber
formed one loop around the flat-plate return conductor as
labeled. In an optimal setup both ends of the fiber loop would
be co-routed away and exit the high magnetic field test zone.
However, optimal setup was not achieved due to routing space
restrictions. One end of the fiber was routed near the test zone
into the shielded enclosure that housed the optical box and the
data acquisition system. The other end was unable to form
closed loop at the shielded enclosure and was simply formed a
coil on the floor. Thus, a fiber section was exposed to high
magnetic field whose effects would otherwise be canceled if a
50
100
Time (sec)
150
200
Four Pearson CT
Faraday Sensor
150
DIRECT LARGE CURRENT MEASUREMENT
The 1310nm sensor system was evaluated for large current
performance using only one fiber loop on one conductor
(N=1). Using one fiber loop would be similar to installation
external to an aircraft fuselage, a large structure, or a lightning
tower. The tests were performed at a commercial lightning test
facility, using standard aircraft lightning test waveforms that
include components D, B and C [11]. Test current amplitudes
were 20, 40, 100 and 200 kA. All were of double-exponential
waveform, except for the 200 kA damped sinusoidal (limited
by the test facility’s abilities to generate unipolar waveforms).
This test piggybacked on a separate effort to evaluate lightning
effects on composite panels.
0
200
100
Current (kA)
IV.
60
20
-300
-350
-50
Four Pearson CT
Faraday Sensor
100
Current (kA)
N*Current (kA•turn)
800
50
0
-50
200 kA
-100
-150
-200
-100
0
100
200 300 400
Time (sec)
500
600
700
Fig. 10. Reasonable comparison achieved measuring large current
(100 kA and 200 kA peaks) despite imperfect setup.
closed loop was achieved. Consequently, some measurement
error was anticipated.
Fig. 10 shows results for 100 kA and 200 kA peak current
against reference results. The reference results were the
mathematical sums of four Pearson CTs outputs measuring
current exiting the four sides of the composite panels.
The results are reasonably good considering the nonoptimal fiber routing. The errors are about 3-10% depending
on the routing of the unpaired fiber section. These results
demonstrate that the Faraday sensor is capable of directly
measuring 200 kA current using just one loop, similar to
expected structure installations.
V.
Lightning
Waveform
Generator
Faraday
Current Sensor
Rogowski Coil,
Pearson CT
LARGE STRUCTURE MEASUREMENT
Fig. 11 illustrates the setup measuring current on an
aluminum cylinder that simulates a round lightning tower or an
aircraft fuselage. Current lightning waveforms from 250 A to
4 kA were injected onto the cylinder at the bottom left location.
The current amplitudes were limited by the laboratory
equipment used. Return currents were extracted from the
cylinder at bottom right.
Optical Box
Fig. 11. Measurement on a large aluminum cylinder simulating an
aircraft fuselage.
The 15m fiber form a single fiber loop around the cylinder,
with both ends co-routed to the optical box located 4m away on
the table in the foreground. As can be seen in Fig. 11, the fiber
closed the loop at the optical box without any unpaired fiber
section, thus good isolation was achieved. A Pearson CT and a
Rogowski coil provide reference comparison data. Fig. 12
shows good results for both the 250 A and 4 kA tests.
0.25
(i)
0.25 kA
0.15
0.1
0.05
0
-0.05
-50
TRIGGERED LIGHTNING MEASUREMENT
Over the summer of 2012, the sensor system based on the
1550 nm wavelength was demonstrated measuring rockettriggered lightning for a more realistic lightning environment.
The test was performed at the ICLRT facility [12]. The
measurement was similar to an earlier (2011) and successful
effort utilizing a sensor system operating at 850nm laser
wavelength and a twisted single-mode sensing fiber [13-15].
In the 2012 setup shown in Fig. 13, triggered lightning flashes
would attach to the wire cage, and lightning current would
travel to the ground via a resistive shunt (T&M Model R-700010) and a down-conductor. The sensing fiber formed four loops
around the conductor. The remaining fiber segments at the two
ends were co-routed radially away from the site. One end was
connected to the optical box 12m away. The other end was
connected to a Faraday mirror that was buried in the ground to
minimize temperature variations.
Due to insufficient fiber length, the Faraday mirror was
positioned only about 4m from the launch tubes rather than
closing the loop near the optical box. Thus, about 8m (from
the Faraday mirror to the optical box) section of the sensing
fiber was “unpaired”, potentially exposed to the effects of
magnetic fields from the lightning flash and strong ground
currents that would normally be canceled with a closed loop.
However, by routing the fiber in the radial direction away from
0
50
100
4
150
Time (sec)
4.5
200
Pearson CT
Rogowski Coil
Faraday Sensor
(ii)
3.5
3
Current (kA)
VI.
Pearson CT
Rogowski Coil
Faraday Sensor
0.2
Current (kA)
Noise is clearly visible in the 250 A measurement,
illustrating the low level sensitivity limit. The dominant noise
source is the SLD laser, which is a wideband noise source.
Noise reduction techniques implemented include the balanced
detector and a 1.9 MHz low-pass filter. In addition, movingwindow data averaging is implemented in data post-processing
with a small averaging window, i.e. 11 point window out of
10,000 points data length. A 60 dB range could be achieved
with this setup.
0.3
4 kA
2.5
2
1.5
1
0.5
0
-0.5
-50
0
50
100
Time (sec)
150
200
Fig. 12. Current measurement on a 1.2 m diameter cylinder.
the lightning tower, magnetic field components in the direction
of the fiber are expected to be minimized, reducing any
undesirable effects. The fiber was protected from wild animals
or being trampled on inside combinations of rain gutters and
plastic braided sleeves (Figs. 13-14). Data were recorded using
14-bit digitizers at 100 MHz sampling rate. The sensors and
digitizers were powered by batteries.
Triggered Lightning ICLRT - 08/19/2012 - Sweep 98
16
(i)
14
Resistive Shunt
Faraday Sensor
Current (kA)
12
10
8
6
4
2
0
Resistive
Shunt
-2
-1
0
1
Down
Conductor
4 fiber
loops
2
3
4
5
Time (msec)
6
7
8
Triggered Lightning ICLRT - 08/20/2012 - Sweep 58
12
(ii)
10
Fig. 13. Four loops of fiber optic current sensor installed under the
rocket launch tubes.
Resistive Shunt
Faraday Sensor
Optical Box
Batteries
Digitizers
Current (kA)
8
6
4
2
0
-2
-1
0
1
2
3
Time (msec)
4
5
6
Triggered Lightning ICLRT - 08/17/2012 - Sweep 46
14
12
Fig. 14. Optical box and digitizers located 12m from the launch
tower.
Before any actual triggered lightning measurements, a
series of verification tests were conducted comparing outputs
from three different sensors: the Faraday sensor, the reference
resistive shunt, and a current transformer. Several one kA
positive and negative current waveforms were injected onto the
wire cage that surrounded the rocket launch tubes while return
currents were extracted from the base of the down-conductor.
The results comparisons were excellent, verifying the accuracy
of both the shunt resistor and the Faraday sensor. There was
no ground current in this test setup as with actual lightning
flashes.
Fig. 15 illustrates good result comparisons with the shunt
resistor were achieved with actual triggered lightning. Electric
current amplitude-versus-time waveforms are nearly identical
between the two sensors. The long time scales chosen in the
Resistive Shunt
Faraday Sensor
10
Current (kA)
The 1550nm system is capable of measuring NI = 400 A to
400 kA range. To improve sensitivity, four fiber loops (N=4)
were used. The current range (I) is effectively 100 A to 100
kA, which is reasonable for typical low peak lightning levels
observed at the site.
(iii)
8
6
4
2
0
-0.5
0
0.5
1
1.5
Time (msec)
2
2.5
3
Fig. 15. Results for the 1550nm system show good comparisons with
resistive shunt.
plots highlight the ability to measure long duration components
such as continuing current. It is noted that the actual lightning
currents I are reported in Fig. 15. The amplification effects
from the multiple fiber loops are removed from the data.
As a side note, the system suffered from electromagnetic
interference in early results due to strong ground currents,
affecting the peak measurements. In the later results as shown,
interference became much less simply by slightly raising off
the ground the data cables connecting the optical box to the
data acquisition system. The cable spacing above ground was
about 5 cm, supported underneath by a wood beam. Fig. 15(i)
shows about 400A error in the peak currents relative to the
resistive shunt. The remainder of the waveforms compare very
well.
kA and 400 kA for another for the two systems. Advantages
such as structure conformity, total current measurement, nonconductivity, being light weight and many others make this
sensor truly unique for direct lightning measurement.
The interference problem could be further minimized in
future setups that have strong ground current by having better
cable shielding, by elevating the cables and/or setup higher
above the ground, or by having the optical box and the
digitizers in the same shielded enclosure. It is noted that
ground current is not a problem for aircraft installations as
equipment will be protected inside the fuselage.
[2]
[5]
VII. SENSOR BANDWIDTH
[6]
Bandwidth of a sensor system is limited by the lowest
bandwidth of its components. For the fiber sensor component,
it is limited by the light transit time in the interaction length of
the fiber. This limitation is to ensure that the total transit time
is much faster than the signal change rate for proper
integration. The fiber interaction length includes the round-trip
distance around the conductor and includes the distance to and
from the Faraday sensor. The 3-dB sensor bandwidth (BW) is
[3,4]: BW ~ 0.44/t ~ 0.44c/nl, where t is transit time, c is the
speed of light in free space, n is the index of refraction in fiber
material (n=1.5), and l is the interaction length (double of fiber
length for reflective scheme).
Table I computes the maximum fiber length and structure
dimensions for different bandwidths. Higher measurement
bandwidth would require reduced structure size. Aircraft thin
structures may include wings and tail surfaces, while round
structures may include fuselage, etc. Since most of the
lightning energy that can cause structure damage is contained
in bandwidth far below one MHz, structures in excess of 14m
diameter can be measured. This is sufficient even for the
current largest passenger aircraft fuselage, at 7.8m diameter for
the Airbus A380 model.
TABLE I.
3-dB
Bandwidth
(MHz)
1
STRUCTURE DIMENSION VS. SENSOR BANDWIDTH
Max Fiber
Length (m)
44
Max Thin
Structure
Dimension (m)
22
Max Round
Structure
Diameter (m)
14
2
22
11
7
4
11
5.5
3.5
10
4.4
2.2
1.4
20
2.2
1.1
0.7
VIII. CONCLUSION
The design, accuracy, advantages and versatility of fiberoptic current sensor are described and validated though
multiple demonstrations. The measurements include large
current up to 200 kA, current on large structure, and triggered
lightning. Accurate equivalent current was achieved up to 300
IX.
[1]
[3]
[4]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
REFERENCES
F. Pitts, B. Fisher, V. Mazur, and R. Perala, “Aircraft Jolts from
Lightning Bolts,” IEEE Spectrum, July 1988.
P. Laroche, P. Blanchet, A. Delannoy, F. Issac, “Experimental Studies
of Lightning Strikes to Aircraft,” Onera Aerospace Lab Journal, Issue 5,
December 2012 (AL05-06).
J. M. Lopex-Higuera, Editor. Handbook of Optical Fibre Sensing
Technology, 2002; Sections 27.2 - 27.4.
G. Day, and A. Rose, “Faraday Effect Sensors: The State of the Art,”
Proc. SPIE, 1988, pp. 138–150.
P. Drexler and P. Fiala,”Utilization of Faraday Mirror in Fiber Optic
Current Sensors”, Radioengineering, Vol. 17, Dec. 2008.
R. Laming and D. Payne, “Electric Current Sensors Employing Spun
Highly Birefringent Optical Fibers,” Journal of Lightwave Technology,
Dec. 1989.
A. White, G. McHale, D. Goerz, “Advances in Optical Fiber-Based
Faraday Rotation Diagnostics,” 17th IEEE Int. Pulsed Power
Conference, Wash. DC, July 2009 (LLNL-CONF-415198).
A. Smith, “Polarization and Magneto-optic Properties of Single-Mode
Optical Fiber,” Applied Optic, Jan. 1978.
R. Ulrich, and A. Simon, “Polarization Optics of Twisted Single-Mode
Fibers,” Applied Optics, Vol. 18, Issue 13, pp. 2241-2251 (1979).
S. Rashleight, “Origins and Control of Polarization Effects in SingleMode Fibers,” Journal of Lightwave Technology, Vol. LT-1, No. 2,
1983.
ARP-5412 “Aircraft Lightning Environment and Related Test
Waveforms,” Rev B, Jan 2012.
V. Rakov, “A review of Triggered-Lightning Experiments,” 30th
International Conference on Lightning Protection, Cagliari, Italy,
September 13-17, 2010.
T. Nguyen, and G. Szatkowski, “Fiber Optic Sensor for Aircraft
Lightning Current Measurement”, ICOLSE, 2011.
T. Nguyen, J. Ely, G. Szatkowski, C. Mata, A. Mata, G. Snyder, “FiberOptic Sensor for Aircraft Lightning Current Measurement,” 2012 Int.
Conf. on Lightning Protection (ICLP).
T. Nguyen, J. Ely, G. Szatkowski, C. Mata, A. Mata, G. Snyder, “FiberOptic Current Sensor Validation with Triggered Lightning
Measurements,” 2013 Int. Conf. on Lightning and Static Electricity
(ICOLSE).