2014 International Conference on Lightning Protection (ICLP), Shanghai, China

Lightning Current Measurement with Fiber-Optic

Sensor
Invited Paper

Truong X. Nguyen, Carlos T. Mata and Angel G. Mata

Jay J. Ely and George N. Szatkowski E SC - Kennedy Space Center, FL 32899 U.S.A.
NASA Langley Research Center
Hampton, VA 23681 U.S.A. Gary P. Snyder
truong.x.nguyen@nasa.gov NASA - Kennedy Space Center, FL 32899 U. S.A.

installations. Faraday Effect causes light's polarization plane

Abstract-A fiber-optic current sensor is successfully developed
in a medium to rotate when the medium is exposed to a
with many potential applications for electric current
measurement. Originally developed for in-flight lightning
magnetic field in the direction of light propagation. Using
measurement, the sensor utilizes Faraday Effect in an optical optical fiber as the propagation medium and by forming closed
fiber. The Faraday Effect causes linear light polarization in a fiber loops, measuring the total light rotation would result in
fiber to rotate when the fiber is exposed to a magnetic field. The the total current enclosed. It is noted that the sensing element
polarization change is detected using a reflective polarimetric is an optical fiber, thus termed "intrinsic" sensor. In contrast, in
scheme. Forming fiber loops and applying Ampere's law, "extrinsic" sensor optical fiber is only used for relaying signal
measuring the total light rotation results in the determination of from a remote sensor.
the total current enclosed. The sensor is conformable to complex
This sensor overcomes many limitations of traditional
structure geometry. It is also non-conductive and immune to
electromagnetic interference, saturation or hysteresis. Installation sensors. Much like a Rogowski coil, total current is measured
is non-intrusive, and the sensor can be safely routed through by installing fiber loops around the structure of interest.
flammable areas. Two similar sensor systems are described in However, the sensor is self-integrating, thus no integrator
this paper. The first system operates at 1310nm laser wavelength needed, and DC component can be measured with suitable
and is capable of measuring approximately 300 A - 300 kA, a 60 electronics. It does not suffer from self-resonance, hysteresis
dB range. Laboratory validation results of aircraft lighting direct and saturation as with Rogowski coil or current transformers.
and in-direct effect current amplitudes are reported for this It is not susceptible to arcing/sparking from high voltage and
sensor. The second system operates at 1550nm wavelength and current. Measurement can be performed at arbitrary locations
can measure about 400 A - 400 kA. Triggered-lightning on the structure, unlike a resistive shunt that must electrically­
measurement data are presented for this system. Good results are isolate fuselage sections.
achieved in all cases.
Keywords-lightning; Faraday Effect; fiber-optic; current; There are also many unique characteristics. It is small,
lightning light-weight, flexible, yet can safely and accurately measure
very large lightning current. It is highly conformable to
arbitrarily complex structure shapes, and can fit tightly onto
I. INTRODUCTION
structure's surfaces. Installation is simple and non-intrusive,
Accurate characterization of in-flight lightning return stroke simply by wrapping the thin fiber one or more times around the
current is important in developing aircraft test requirements structure to be measured. Being non-conductive, the sensing
and protection against lightning direct and indirect effects. fiber can be safely routed directly into the aircraft fuselage,
Accurate measurements may also help to infer aircraft damages building, or through flammable regions without requiring
immediately after a strike from which appropriate actions are special converters. Versatility is excellent, as the same fiber
determined. Sensors used in previous flight efforts to can be used on both small and large structures, given sufficient
characterize lightning parameters [1,2] either required length and bandwidth. Measurement sensitivity may be
significant and expensive aircraft structure modifications (i.e. tailored by varying the number of fiber-turns around the
resistive shunt, I-dot sensor), or were unable to directly conductor.
measure total current (i.e. magnetic field sensors). Time­
derivative sensors cannot measure low frequency components The sensor is not without limitations. Fiber choice is
near DC. Rogowski coil and current transformer are not limited, as most commonly available fibers are based on silica
practical for external installations due to strict requirements on in which the Faraday Effect is weak. However, the weak effect
safety, weight, and aircraft aerodynamic performance. makes the sensor highly suitable for large current applications
such as lightning measurement. There are slight sensitivities to
This paper describes a fiber-optic current sensor based on bend/vibration and temperature variations, but there are
Faraday (Rotation) Effect that can directly measure current on compensation methods. Glass fiber can be fragile and needs
aircraft structures and is highly suitable for most aircraft suitable protection.

978-1-4799-3544-4/14/$31.00 ©2014 IEEE 1424

 Aircraft installations and their demanding safety and II. FIBER-OPTIC CURRENT SENSOR CONCEPT
aerodynamic performance requirements would benefit the most
from this new sensor's many unique characteristics previously A. Basic Sensor Operation
described. Flight measurements would be greatly simplified by
The Faraday Effect causes linearly polarized light in the
the direct measurement of total current; no longer required are
sensing fiber to rotate when exposed to an external magnetic
field sensors and analytical efforts necessary to solve an
field in the direction of light propagation [4-8]. The effect in
inverse problem. Fig. 1 illustrates fiber-optic sensor loops
an optical fiber is illustrated in Fig. 2. The total polarization
installation in contrast with traditional B-Dot sensors that can
only measure local magnetic fields.
plane rotation cP is
/
To date the sensor has not yet been flown on an aircraft due
I 1>
to high costs in any flight experiment. In addition, most
aircraft avoid flying into thunderstorms intentionally due to ... �.I{
........ .. ····
. ..
t
1·
damage risks. Regardless, future flight efforts could =
.......... ......... ........... ... . ... ··· ·· ··
E

/f-;
significantly benefit from the sensor's simplicity, accuracy and E
�
I� 1::=-----*1
the direct current measurement ability. The sensor's many .

advantages could also benefit other direct lightning

measurements such as on small buildings, windmills, or
Figure 2. Faraday Effect in optical fiber.
lightning towers.

¢ = V f B· dl = 110 V f H· dl , (1)

where /10 is the free-space penneability; V is the Verdet

constant in radians/(meter·Tesla); /1oV is the combined
permeability Verdet constant (radians/ampere); B is magnetic
flux density in Tesla (T); H is the magnetic field
(amperes/meter); and I (in meters) is the light and magnetic
field interaction path length. Forming N fiber loops and
applying Ampere's law yields result in <p being directly
proportional to the total current enclosed J (Eq. 2). Thus, J can
be detennined by measuring <p. As shown in Eq. 2, the sensor
is self-integrating, no additonal integration is needed.

NASA Illustration 2012

cP = 110 V f H . dl ,
(2)
Figure I. Optical-fiber sensors aircraft installation.
= 110VNI.

Current sensing using optical fiber has been under

B. Polarimetric Detection Scheme
development for decades. However, commercial products are
developed mostly for measuring electrical power and may not The scheme to measure polarization change induced by
be suitable for higher bandwidths in lightning. The approach current is illustrated in Fig. 3. Linearly polarized light is
presented this paper is tailored for direct lightning launched from a super-luminescence diode ( SLD) broadband
measurement and provides reasonable compromises between laser onto the sensing fiber at 4 that forms closed loops around
bandwidth, measurement range and accuracy. In addition, the a structure of interest. As light reflects from the Faraday
design and implementation approach gives emphasis to Mirror at 6, its polarization is rotated 90 degrees relative to the
measurement stability, simplicity, low-cost, the ability to incident light. This Faraday mirror implementation makes the
measure (near) DC component in continuing current, and for system less sensitive to bending, since stress induced effects on
not requiring highly specialized optical tools. polarization in one direction are canceled by the same effects
on the orthogonal polarization in the reverse direction [5]. As
The remainder of the paper describes the sensor concept
reflected light traces back to the non-polarizing beam splitter
and implementation, laboratory testing, and field testing with
(NBS) at 3, half of the power is reflected toward the half-wave
rocket-triggered lightning. Some aircraft installation
plate (HWP) at 7 and the polarizing beam splitter (PBS) at 8.
considerations and testing are also discussed. Two similar
Alternately, the PBS can be implemented in a fiber assembly.
systems discussed in this paper are based on 1310nm and
Exiting the PBS, light power in the two orthogonal
1550nm laser wavelengths. The measurement range achieved
polarizations are measured by two photo-detectors D1 and D2
is approximately 300 A - 300 kA for the 1310nm system or 400
at 9. The difference in optical powers at 9 is measured with a
A - 400 kA for the 1550nm system, a 60 dB range. Both
balanced detector, whose output voltage being proportional the
systems are less sensitive but otherwise similar to a 850nm
difference between the two detectors' outputs. The HWP at 7
system that successfully measured rocket-triggered lighting [3].
allows one to rotate and align the beam polarization incident on
the PBS.

1425
 1.0
Ideally, at zero current the incident polarization should be --
--
Round·tripReJPOfl�
F\lnction t!t 1l10nm
-
__.. � 0.8
at 450 relative to the PBS's two orthogonal principle Ibl·•••••••••• ,6 _
• Q ••
polarization axes, so that beam power is divided equally _ - If. • •• • ••• •
•
between the two optical detectors at 9. With no current the
balanced detector's output should be at zero volts. With 0.2
---
current, Faraday Effect causes the light's polarization to rotate
-400 -300 -200 -100 100 200 300 400
causing power imbalance at the two detectors, leading to a '0.2
NI (ItA)
voltage output at the balanced detector. -0.4
1<'
-0.6
- - 10I0.5'(1+sin(4VNili
-0.8 ----Ib] O.S'(1-sin(4VNili
3 Non-Polarizing -le]=lo-b]/lo+b]

�-CD-�:Qr�r---.-,,---:-���;I-
-1.0

Figure 4. Ideal sensor responses at 131 Onm.

1 Broadband ;I Polarizer :

r(y
I
Laser
7 H alf-wave The 1550nm-based system is slightly less sensitive due to
Plate � I
the lower Verdet constant at this wavelength, so it can measure
I
slightly larger current. The practical range is approximately +/-
�
I
8 Polarizing
O
� ·[ Del. b 1 � 400 kA. The design, construction and characteristics otherwise
tt
..
Beam Splitter Collimator
c:::> 9 De ec or (b) are very similar to the 1310 nm system.
---- Free-space beam
L ...... �
! -4Sd
........ Multi-Mode Fiber
. Over-current would not damage the sensor. Light
.
polarization would simply rotate beyond the intended range,
9 Detector (a)
!
+4Sd and the solution to the sino function in (4) would be
ambiguous. However, there are minor variations of the
Figure 3. Opical detection scheme.
detection scheme that permit measurement of very large
current [6]. These practices are not necessary here since for
This setup is referred to as a reflective scheme since a most direct lightning are within the 300 kA and 400 kA range
mirror is incorporated and light travels round-trip. With the of the two systems described.
Faraday Effect being non-reciprocal, the Faraday rotation is
doubled relative to single-pass scheme. The responses at the It is important that light's state-of-polarization (linear) is
two detectors should ideally be (a),(b)=O.S*[l±sin(4/1oVNI)] maintained in the fiber during light transit. This is achieved
for a reflective scheme. Mathematic operation difference-over­ with fiber selection. The two systems in this paper use two
sum, (c)=(a-b)j(a+b), yields different commercial highly spun polarization-maintaining
(PM) fibers [7], which are the result of twisting PM fibers
(c) sin (4IloVNI), or (3)
during manufacturing. Fiber twisting helps hold the state-of­
polarization that otherwise would be destroyed in a typical
NI (4) fiber. The twist rate is about 4-5 mm per turn. Additional
infonnation on fiber-optic current sensor can be found in [4-8].
where equivalent current NI (in unit Ampere-tum) is defined as
number of loops N times the current I, and /10 V = 1.01 wad/A C. Sensor Calibration and Data Correction
at 1310nm and 0.718 wad/A at 1550nm [4] for the two systems
Calibration is perfonned in laboratory by relating the
described in this paper.
sensor's response curve to data from reference sensors, and a
Fig. 4 illustrates the ideal responses at 1310nm, with the calibration function is developed. The reference sensors used
response curves (a) and (b) being voltage outputs from the two include a Rogowski coil with an electronic integrator and a
optical detectors. The difference-over-sum operation, (c)=(a­ ferrite-based Pearson™ current transformer (CT). Fig. 5
b)j(a+b), would yield a response that is more sensitive (higher compares the three sensors by plotting peak equivalent current
slope) than either response curve (a) or (b), with zero crossing using the Faraday sensor on the vertical axes against currents
at zero current, and has larger dynamic range due to common­ from the reference sensors on the horizontal axes. Large
mode noise subtraction. Current 1 or equivalent current NI is equivalent currents in the figure were achieved by using
then computed from (c) using eq. (4). multiple fiber turns and/or a wire coil - details are described in
the next section.
The typical operating range is where the curve (c) increases
monotonically, or about -350 kA to +350 kA. In this range the Ideally the Faraday sensor data would fall on the straight
response and current correspond one-to-one. In the systems diagonal line labeled "ideal" representing (1:1) correspondence
described, the non-ideal fiber medium and optical components with the two reference sensors. Instead, the data follow the
distort the curves. The practical range is slightly reduced to curve labeled "uncorrected". This non-ideal response is due to
about -300 kA to +300 kA, as shown in the next section. the reduced sensitivity in the spun fiber [7], relative to ideal
medium of the same material - along with some light
depolarization from the non-ideal fiber medium and optical

1426
components. Thus, it is important that the sensor is calibrated For simplicity, N is redefmed from Eq. (4) to be the product
to develop a correction factor or function. of the number of fiber loops and the number of wire turns. The
product NI is referred to as equivalent current as previously
A correction function can be developed by spline-fitting the
stated in Eq. (4).
"uncorrected" curve in Fig. 5 with a polynomial function
th
(typically 5 or higher, odd order). The purpose is to map the Fig. 7 illustrates results comparison between the Faraday
Faraday sensor response curve to the "ideal" curve. That sensor and the reference sensors, with different nwnber of fiber
function is then applied to subsequent measurements to achieve turns and wire loops. The equivalent currents NI are about 5
the corrected results.Fig. 5 shows the "corrected" response kA-turns and 300 kA-turns. The same calibration correction
th
curve aligns well with the "ideal" diagonal line. A 9 order function was used and good results were achieved in both
polynomial was used. Data interpolation may also be used cases. Similarly, NI up to 400 kA-turns was demonstrated with
instead of curve fitting; however, neither approach is perfect, the 1550nm wavelength system with equally good results.
and some small error may remain. In addition, only a simple Since the reference Rogowski coil and Pearson current
correction factor may be needed if measurement is restricted to transformer (CT) only measure current on one wire-turn, their
the linear region, i.e. 100 kA or less. results are numerically scaled up by the factor N for the
comparison. This practice is commonly used and accepted [4-

� 1
300 8].
I I y---
I Corrected �
200 Faraday vs. Pearson ...
,.
,it" /!-
� 100

/'1\ [
..

V I
0 Uocorrected
Faraday vs. Pearson
Faraday vs. Rogowski.

V
�b?
-200

-300

-300 -200 -100 0 100 200 300

Pearson & Rogowski Current (kA)

Figure 6. Using multi-tum coil and multiple fiber loops to achieve high
Figure 5. Example of corrected and un-corrected response curves.
current effects.

III. LABORATORY DEMONSTRATIONS

(ii)
This section describes various laboratory lightning
c . C -$0
measurement demonstrations. The measurements include a) !
�! -100
simulated large current using multiple fiber loops and/or a
current coil, b) up to 200 kA direct lightning standard test
�,
�� 1
5 kA
i -1$0
� -200
waveform performed at a commercial laboratory , c) current on 'I
� -1$0 300 kA
z '
a simulated aircraft fuselage, and d) low-level indirect lighting
current inside a simulated fuselage. Demonstrations :�"L���"-�"
, ,-.,� , ----c.,� , ----:,,�,,� _H�_ ---o----:-:-
,, ---c,=
,,-�-___c

measuring rocket-triggered-lightning are described in the next Time (llSec] Time [�iS.cl

section.
Figure 7. Laboratory testing at N/=5 kA and 300 kA. (i) 49-turn coil and
A. Laboratory Simulated Large Current one fiber loop (N=49), and (ii) 3-turn coil and 28 fiber loops (N=3*28=84).

It is difficult and unsafe to achieve in laboratory settings

current levels up to 300 kA (or 400 kA for the 1550nm system) B. Direct Lightning Large Current Measurement
for the purpose of calibrating and testing the sensor. One The 1310nm sensor system was evaluated for large current
acceptable approach to producing the associated optical effects performance using only one fiber loop around one conductor
in the fiber is by using multiple fiber loops and/or a multi-turn (N=I). Using one fiber loop would be similar to installation
conductor coil. Multiple fiber loops and/or multi-turn coil external to an aircraft fuselage, a large structure, or a lightning
amplify the Faraday rotation beyond the amount produced by a tower. The tests were perfonned at a commercial lightning test
single fiber loop around a single conductor. The amplification facility using standard aircraft lightning test waveforms and
factor is the multiplication of the number of fiber-loops and the include components D, B and C [9]. Test current amplitudes
number of wire turns used. Fig. 6 illustrates a typical setup. were 20, 40, 100 kA with double-exponential waveforms. In
Laboratory tests show using mUltiple fiber loops, a multi-turn addition, 200 kA tests with damped sinusoidal wavefonn were
coil, or combinations of both yield the same response curves. also conducted. This was necessary due to the test facility's
Using a wire coil having a high number of turns may distort the limited abilities to generate unipolar wavefonns at high current.
injected wavefonn due to the increased inductance. In contrast, The measurement piggybacked on a separate effort to evaluate
a high number of fiber loops does not affect the current lightning effects on composite panels.
waveform but would require a longer sensing fiber.

1427
 The setup shown in Fig. 8 included the fiber fonning one section, thus good isolation was achieved. A Pearson CT and a
closed loop around a flat metal plate (about 0.5m wide) Rogowski coil provide reference comparison data. Fig. 11
carrying the return currents that exit the composite panels shows good results for both the 250 A and 4 kA tests.
under test. Due cable routing restrictions, a section of the fiber
Noise is observable in the 250 A measurement, illustrating
was routed near the lightning test zone on its way to the optical
the low level sensitivity limit. The dominant source of noise is
box. Closing the fiber loop at the optical box to achieve good
the SLD laser, which is inherent a wideband optical noise
measurement isolation was not possible; thus, some error was
source. Noise reduction techniques implemented include using
anticipated. The Faraday sensor results were compared against
the balanced detector for common mode noise subtraction and
the sum of four Pearson CTs measuring currents exiting the
a l.9 MHz low-pass filter. In addition, a small moving­
four sides of the composite panel.
window data averaging is implemented in post-processing, i.e.
Fig. 9 shows results for 100 kA and 200 kA peak current. II-point window out of 10,000-points data length. A 60 dB
The comparisons are reasonably good considering the non­ range could be achieved with this setup.
optimal setup. The errors are about 3-10% depending on the
routing of the exposed fiber section through the lightning test
zone. It is clear that the Faraday sensor is capable of directly
measuring 200 kA current using just one loop, and that a better
setup could lead to better comparisons. The 20 kA and 40 kA
comparisons are similar to the 100 kA results.

Figure 10. Measurement on a l. 2m diameter aluminum cylinder

simulating an aircraft fuselage.

I I
·· ... "o.CT

(i) (ii) -Rogowski Coil

-fon,day Sensor

� > .•
Figure 8. A high-current measurement setup. l >
J 1-J

c--7----!c--=--c:
-O'O'.

Time (J.<HC) Time(J,lllK)

� 50
Figure I I. 250A and 4 kA peak current measurements on a 1.2 m (4 feet)
C ,
t diameter structure.
a-50
D. Indirect Lightning Internal Current Measurement
ol?SL, -�---c,�, --,�, , -�-�
Figs. 12-13 demonstrate excellent isolation for low level
Time (j.lSecl Time (llSecj
current measurement on an internal structure. This setup
Figure 9. Measurements of 100 kA and 200 kA on a current return plate. simulates induced current on equipment or wire bundles in the
presence of significantly larger current flowing on the outer
C. Measurement on Simulated Fuselage aluminum structure. About 2 kA peak total current was
injected on the aluminum cylinder, with about 140 A flows on
Fig. 10 illustrates a setup measuring current on a l.2m
the internal box. With some slight data averaging for noise
diameter aluminum cylinder that simulates an aircraft fuselage
reduction, generally good comparison can be observed even
or round lightning tower. Current lightning wavefonns from
though the Faraday sensor's data were close to the noise floor.
250 A to 4 kA (amplitudes limited by laboratory equipment)
were injected onto the cylinder at the bottom left location. Excellent results were also performed when the fiber loop
Return currents were extracted from the cylinder at bottom was expanded all the way to the metal frame of the aluminum
right. structure. This illustrates the ability to measure the composite
sum of all lightning induced current on all internal cables and
The 15m sensing fiber form a single loop around the
structures. This ability could precisely determine the
cylinder, with both ends co-routed to the optical box located
occurrence of strong lightning induced current, and help with
4m away in the foreground. As shown in Fig. 10, the fiber
equipment maintenance scheduling after a lightning strike.
closed the loop at the optical box without any unpaired fiber

1428
 components in the direction of the fiber are expected to be
minimized, reducing any undesirable effects. The fiber was
also protected from wild animals or being trampled on inside
combinations of rain gutters and plastic braided sleeves.

The system successfully measured rocket-triggered

lightning with excellent result comparisons. With the four
fiber loops (N=4) used, the measurement range for the sensor is
100A - 100 kA. Fig. 15 illustrates the representative results.
Also shown is low-level comparison to highlight the
measurement range and the ability to measure continuing
current. Data were slightly smoothed for noise reduction. The
results are nearly identical between the two sensors. Minor
Figure 12. Low-level measurement on an internal structure.
peak amplitude differences in some data (up to 500A) were
attributable to interference coupling from ground lightning
-Box Curren.· Faraday
M.....""" Box Curren •. Pearson CT current. The good results illustrate the system can successfully
-To.al Current· Rogowski
operate in realistic lightning environments. Additional sensor
validation data can be found in [11-12].
_ O.IS
2 kA Total Current

i
�
0.1

V 0.05

-140 A Current on Box

·o.I L-�����-��-��
- - - - - - - -
Time (j.lSecj

Figure 13. Good comparison for low-level measurement in presence of

larger current on an external structure.

IV. ROCKET-T1UGGED LIGHTNING MEASUREMENT

The 1550nm system (NI = 400A - 400 kA range) was

demonstrated measuring rocket-triggered lightning in 2012.
Figure 14. Rocket-triggered lightning current measurement setup at ICLRT.
The measurements were performed at the International Center
for Lightning Research and Testing (lCLRT) at Camp Triggered lightning ICLRT - 08/1912012 - Sweep 98 Triggered Ughtring ICLRT - 0811712012 - Sweep 46
15 15
- Faraday Sensor
I
Blanding, Florida, U.S.A. facility [10]. The measurements -Faraday Sensor
-----·Resistive Shunt ------. Resistive Shunt

�
were similar to an earlier (2011) successful effort that used
I. � I.
twisted single-mode sensing fiber [3] at 850nm wavelength. �
e e
In the setup shown in Fig. 14, triggered-lightning flashes ., �
:: .. s
would attach to the intercepting ring, and lightning current " s "
v v

\..
would travel to the ground via a down-conductor. A resistive
._ .r-. _
shunt (T&M Model R-7000-1O) in-line with the down
2 4 • -0.5 0.5 1 1.5 2.S
conductor provided reference measurements. The sensing fiber Time (mseCj Time (msec)
formed four loops around the same down-conductor. The Triggered Lightring ICLRT - 08120/2012 - Sweep 63 Triggered Lightring ICLRT - 08/2012012 - Sweep 63
s
remaining fiber segments at the two ends were co-routed - Faraday Sensor
I
� Faraday Sensor

radially away from the site. One end was connected to the ----. Resistive Shunt ----.. Resistive Shunt

supporting optical box 12m away. The other end was � 1• � 1

� �
connected to a Faraday mirror 4m away that was buried in the e e
., .,
..
ground to minimize temperature variations. Data were t: s ; 0.5
"
recorded using a commercial 14-bit digitizer sampling at 100 v V

MHz. The sensors and digitizers were powered by batteries. r... -

r--
Due to insufficient fiber length, the Faraday mirror was not 2 • • 2 • •
Time (mseCj Time (mseCj
positioned near the optical box. About 8m (from the Faraday
Figure 15. Sample triggered lightning measurements results and low-level
mirror to the optical box) section of the sensing fiber was
comparison.
"unpaired", potentially exposed to effects 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 the site, magnetic field

1429
 V. AIRCRAFT INSTALLATION AND CONSIDERATIONS Data recording is triggered on the AC component of the
output. This is more accurate than triggering on the total
It is a goal to fly one or more sensors to collect aircraft
current due to a slow and small DC voltage drift caused by
lightning data. A basic multi-sensors setup is illustrated in
various environmental factors. In the 1310nm system the
Fig. 16 with four sensing fiber loops along with optical
voltage drift can be up to an equivalent of 600 A. The DC drift
interface boxes. Sensor outputs are recorded using a high
can be subtracted out at data post processing by comparing
speed data recorder. In addition, AC components are recorded
against baseline data just before the lightning strike.
separately and used for recorder triggering (more details later).
No additional optical/electrical interface is needed as the
sensing fiber is inherently safe and can be routed directly into B. Fiber Installation Locations

the aircraft. No other field sensors are necessary to trigger the The primary fiber loop locations are around the fuselage
system. Installation is relatively simple due to optical fiber's before and after the aircraft's wing section. These two
small size and weight. However, there are a few locations would capture most of lightning current paths except
considerations to be discussed below. for the wing-to-wing path. And by having the fiber loop close
to the wing section more of swept-stroke lightning
components may be captured. Locations on aircraft wings are
also desirable; however, moving surfaces and landing gears
'" could be problematic on some aircraft. A small metal panel in
",.ox

§
'" place of a window could provide mUltiple fiber penetrations
",.ox

c=:> '"
into the fuselage.
",.ox
---
'"
",.Il(
C. Engine Ingestion and Adhesive Tape Testing
Optical
Faraday Sensor loops Data Recorder Engine ingestion is a significant concern if the fiber is
Interface
installed in front of engine intakes. This is the case for aircraft
Figure 16. Multi-sensors setup example.
with engines mounted toward the rear section. The fiber,
therefore, must be properly secured, possibly with adhesive
A. The Sensor and Data Acquisition Systems tape and non-metallic clamps.
Of three systems constructed and tested operating at To develop confidence in the installation method and the
850nm [3], 1310nm and 1550nm, the l310nm system is adhesive tape, representative fiber sections were flown on two
chosen for future aircraft installations. This wavelength different aircraft. Aircraft-grade polyurethane tapes were
provides slightly better sensitivity than the 1550nm (300 A vs. used. One installation, Fig. 18, has been flown for over a year
400 A) while offers adequate measurement ceiling at 300 kA. and includes short metal wire sections (in place of fiber) in
In addition, optical components are readily available and can three orientations. The other installation includes a
be smaller than for the 850nm system. Other wavelengths are representative fiber loop fully encircling the aircraft fuselage
also possible though not tested in this effort. at location behind the engine intakes. This second setup
The 1310 system was re-configured for aircraft installation shown in Fig. 19 has been flown for about six months.
as shown Fig. 17. Digital displays were added for health Neither setup has shown signs of degradation to date.
monitoring. The system is fitted with a 40 m long spun fiber Successful tests would minimize concerns for engine ingestion
for installation flexibility. Though multiple sensors are and would allow for the fiber installation at locations in front
envisioned in future setups, only one system is planned for the of engine intakes.
initial test.

Figure 17. Faraday-sensor flight optical box.

Figure 18. Adhesive tape testing holding wire samples in three
The sensor's output is recorded using a commercial polarizations (circled).
digitizer/recorder capable of 100 mega-samples per second at
14-bit data length. Up to 20 data channels (using 5 data
acquisition boards) can be recorded into the system's solid
state memory.

1430
 of large current up to 200 kA, current on and inside a simulated
fuselage, and triggered-lightning current. The sensor's many
advantages make it unique for many lightning measurement
applications.

VII. REFERENCES
[ 1] P. Laroche, P. Blanchet, A. Delannoy, F. Issac, "Experimental Studies of
Lightning Strikes to Aircraft," Onera Aerospace Lab Journal, Issue S,
December 20 12 (ALOS-06).
[2] F. Pitts, B. Fisher, V. Mazur, and R. Perala, "Aircraft Jolts from
Lightning Bolts," IEEE Spectrum, July 1988.
[3] T. Nguyen, 1. Ely, G. Szatkowski, C. Mata, A. Mata, G. Snyder, "Fiber­
Optic Sensor for Aircraft Lightning Current Measurement," 2012 Inl.
Conf. on Lightning Protection (lCLP).
[4] J. M. Lopex-Higuera, Editor. Handbook of Optical Fibre Sensing
Technology, 2002; Sections 27. 2 - 27.4.
[S] P. Drexler and P. Fiala,"Utilization of Faraday Mirror in Fiber Optic
Current Sensors", Radioengineering, Vol. 17, Dec. 2008.
[6] A. White, G. McHale, D. Goerz, "Advances in Optical Fiber-Based
Faraday Rotation Diagnostics," 17th IEEE Inl. Pulsed Power
Conference, Wash. DC, July 2009 (LLNL-CONF-4ISI98).
[7] R. Laming and D. Payne, "Electric Current Sensors Employing Spun
Figure 19. Adhesive tape holding test fiber loop around the fuselage. Highly Birefringent Optical Fibers," Journal of Lightwave Technology,
Dec. 1989.
[8] G.W. Day, M.N. Deeter and A.H. Rose, "Faraday effect sensors: A
D. Equipment Shielding review of recent progress," Proc. SPIE 1779, 1992, pp. 2 1-26.
The optical interface box, data recorder and cabling could [9] ARP-S412 "Aircraft Lightning Environment and Related Test
potentially be affected by strong magnetic field produced by Waveforms," Rev. B, Jan 2012.

lightning current if installed near windows. This need is [ 10] V. Rakov, "A review of Triggered-Lightning Experiments," 30th
International Conference on Lightning Protection, Cagliari, Italy,
evidenced in Section IV where the supporting equipment and
September 13-17, 20I O. Update!
cabling were affected by lightning ground current. A
[II] T. Nguyen, J. Ely, G. Szatkowski, C. Mata, A. Mata, and G. Snyder,
shielding enclosure or magnetic shielding materials installed at "Fiber-Optic Current Sensor Validation with Triggered Lightning
nearby windows could minimize the concern. Measurements," 20 13 Inl. Conf. on Lightning and Static Electricity
(lCOLSE).
VI. CONCLUSION [ 12] T. Nguyen, J. Ely, G. Szatkowski, C. Mata, A. Mata, G. Snyder, "An
Intrinsic Fiber-Optic Sensor for Structure Lightning Current
The design, accuracy, advantages and versatility of fiber­ Measurement," 20 141nl. Lightning Detection Conference (ILDC).
optic current sensor were validated though multiple
demonstrations. The demonstrations included measurements

1431