Content uploaded by Fabrice Auzanneau
Author content
All content in this area was uploaded by Fabrice Auzanneau on Feb 07, 2014
Content may be subject to copyright.
1
Aircraft Electrical Wiring Monitoring
System
AUTHORS: Gilles MILLET (Airbus), Serge BRUILLOT (Dassault Aviation), Didier DEJARDIN (Dassault Aviation),
Nicolas IMBERT (EUROCOPTER), Fabrice AUZANNEAU (CEA, LIST, F-91191 Gif sur Yvette, France), Luca
INCARBONE (CEA, LIST, F-91191 Gif sur Yvette, France), Marc OLIVAS (WiN MS), Loïc VINCENT (Volvo Group),
Alain CREMEZI (EUROCOPTER), Sylvain POIGNANT (Safran Engineering Services)
1. Context
The cumulated lengths of electrical cables continuously increases in aircrafts and trucks: now up to 10
kilometers in a modern truck, 40 km in a helicopter or in a fighter aircraft and 500 km in a modern civil
transport aircraft such as A380. Electrical wiring is a critical part in the nominal operation of a system. The
importance of the wired network has thus grown to the same level as the systems it is connected to [1].
Regarding the increasing complexity of the electric system (increase in the number of electric loads, in the
supply voltages), the regulation authorities (FAA, Federal Aviation Administration and EASA, European Aviation
Safety Agency ) now require to consider aircrafts’ electrical wiring as a system on its own, named EWIS
(Electrical Wiring Interconnection System).
On-board electronic equipment and systems are now equipped with integrated functions allowing a fast and
targeted diagnosis called BITE (Built-In TEst), however this is not the case today for the electrical harnesses. The
sensitivity to defects evolves because of the design complexity and integration of new technologies: various
problems can emerge at system level due to electric cables. The diagnosis of the electrical wiring is then
essential to detect and locate these defects, and to improve safety and reduce maintenance costs. This paper
presents recent results in wire diagnosis for transportation.
One can question about the necessity of such a Research and Technology work undertaken to improve cabling
defect diagnosis, since the reliability of the cables installed in modern vehicles is specified to be compliant with
its life time. This statement does not take into account neither the possible human errors during
manufacturing, installation of the electrical harness and maintenance operations, nor the stringent operational
environment in terms of shocks, vibration and humidity aggressing the wiring infrastructure, composed of
cables, connectors and interconnection units.
Several kinds of cable defects must be considered (Figure 1). Depending on them, different techniques must be
deployed. Through Ground Maintenance, only established true defects can be located, after being preliminarily
detected by classical continuity tests, for example.
Transient defects are an important issue because they only appear during usage. Most of the time they can’t
be reproduced afterwards; which often leads to electronic pieces of equipment removal with “No Fault Found”
after test. The only solution to detect and to locate them is an embedded harness maintenance system. Such a
system must be designed for scanning in real time the relevant cables, without affecting their operational
mode of operation, in order to detect a defect and then to focus on the suspicious cable to locate it precisely.
One must keep in mind that on-board cable monitoring must discriminate true defects from normal
discontinuities due to powered off equipment or electrical power distribution reconfiguration such as load
sheds.
Cable degradation diagnosis is the most challenging topic because it can reduce the “Aircraft On Ground”
(AOG) time or improve the Truck Uptime (availability for the driver), thanks to a maintenance operation before
a true defect occurrence. Methods to provide a prognostic are out of the scope of this paper.
Aircraft Electrical Wiring Monitoring System ERTS² 2014
2
Figure 1 : Classification of wire defects found in aircrafts (NASA)
2. Wire diagnosis reflectometry-based methods
Apart from transports, reliable wired networks are of utmost importance in several domains, such as energy
production, large infrastructures, buildings, telecommunication, etc. There are often quite long lengths of wires
of various types (Figure 2) and several systems and sometimes even people directly depend on their good
health. From a quality of service point of view, it is very important to be able to check the proper state of these
cables or to repair any defect as quick as possible. The stakes are huge: the overall cost of an AOG was
estimated by airlines up to USD 150000 per hour.
Figure 2 : Cumulated cable lengths in various application domains
Aircraft Electrical Wiring Monitoring System ERTS² 2014
3
In all these application domains the importance of maintenance is clear, but locating a small defect or even an
open circuit in several kilometers of wires can be very difficult, considering that these cables may not be easy
to reach. There are very few diagnosis and maintenance methods for wires, the most widely used being visual
inspection. But it can’t be exhaustive, especially in the case of very long or hidden wires. The need for more
adapted diagnosis methods has grown and several new electrical methods have been studied and developed
recently.
As of today, the most promising diagnosis methods for wired networks are based on the reflectometry
principles: sending a probe signal down the wire [1]. This signal propagates and sends back a part of its energy
when it reaches an impedance discontinuity. The analysis of the reflected signal – which is generally quite
different from the input signal – provides information on detection, localization and characterization of the
defects which generated the impedance discontinuities.
Some methods are well adapted for the monitoring of ageing cables [2] (e.g. LIRA, Line Resonance Analysis,
which requires heavy equipment), some others are more dedicated to the detection of hard defects (open and
short circuits) for maintenance purposes [3] (TDR and STDR, Sequence Time Domain Reflectometry) or
embarked diagnosis (MCTDR, Multi Carrier Time Domain Reflectometry) [4] and sometimes of soft defects [5].
Recent developments have proven their efficiency for the cancelation of location ambiguities in complex
topology networks [6]. These results have also been extended to the diagnosis of connectors [7], which are
often considered as the weak points of an interconnection system.
Standard reflectometry systems are made of 3 basic blocks: signal generation and injection, measurement
acquisition, data processing and fault localization. These electronic blocks can be studied separately, depending
on the complexity of the chosen method, and each can be prototyped and realized with quite simple electronic
design. Modern reflectometry-based methods use complex diagnosis signals with specific mathematical
properties. The reflectogram is obtained by the correlation of the measured signal and the injected signal.
Embedded diagnosis implies additional constraints such as: cost, low power, small size and harmlessness. The
latter means that the diagnosis signal must not interfere with potential useful signals on the wire, and that the
emitted EM radiation fulfills some EMC requirements. In the case of automotive and avionics, this last
constraint is of extreme importance.
To this end, MCTDR method permits a precise control of the spectrum of the injected signals, and supports a
distributed diagnosis architecture [8] that enables unambiguous and full diagnosis coverage for complex
branched topologies. The diagnosis signal of MCTDR is designed as the sum of a finite number of sinusoids at a
given set of frequencies, chosen outside of the useful signals and EMC (Electro Magnetic compatibility) spectra.
Specific conducted and emitted electromagnetic interference measurements have proven that MCTDR does
not radiate any harmful signal outside of the chosen frequency bands, thus guaranteeing zero interference with
the useful signals of the network under test.
Current achievements in various TDR-based systems have proven the possibility of diagnosing hard defects
with localization accuracy close to few centimeters both for embedded or external purposes. A single
acquisition phase is about 500ns – 1µs long. The signal to noise ratio can be improved by averaging several
results, thus increasing the acquisition duration. To be able to diagnose soft defects, such as damaged insulator
or shielding defect, new methods are currently under study [9].
3. The « Harness BITE » project
3.1 Rationale
The main industrial actors of aeronautics in France have worked in a cooperative research project called
“Harness BITE”, led by Airbus, whose main objective is to develop a real time monitoring system of aircraft
power distribution electrical wiring, aiming at detecting and locating wires defects, even intermittent (in flight),
thus decreasing the maintenance costs.
Aircraft Electrical Wiring Monitoring System ERTS² 2014
4
This project allows to:
Anticipate new FAA & EASA Airworthiness recommendations for EWIS,
Anticipate new risks induced by High-Voltage (arc-fault detection and location),
Create new on-board diagnosis systems for electrical harnesses,
Reduce maintenance cost with defect location capability.
3.2 Electronic design
An innovative reflectometry-based method, called Multi Carrier Time Domain Reflectometry (MCTDR) has been
developed. Several electronic boards were designed and realized (Figure 3) for the needs of the project, based
on the use of FPGA (Field-Programmable Gate Array) components (Altera Cyclone III). The FPGA is in charge of
signal generation and injection via a Digital to Analog Converter (DAC, @100 MSPS) and signal acquisition via an
Analog to Digital Converter and signal processing. A specific IP has been designed and implemented for the
analysis of measured signals and extraction of the useful information, i.e. defect detection, location and
characterization.
Figure 3: Reflectometry electronic boards
The location accuracy in TDR-based methods depends directly on the frequency band of the injected signal, and
the propagation speed of the cable, as shown on formula (1).
(1)
Where c0 is the light velocity in vacuum, is the relative permittivity of the insulator, f is the frequency band
of the DAC. In our case, this leads to 1 m accuracy, which is not acceptable. An innovative oversampling process
has been implemented [10], based on the hypothesis that the state of the cable does not change over a
complete measurement phase. A sequence of N data acquisition desynchronized in time by a portion
of the
DAC period is acquired and used to reconstruct a full reflectogram with N-fold enhanced location accuracy.
3.3 Description of the test benches
In order to incrementally validate the MCTDR based diagnosis methods, three test benches have been used,
with an increasing level of relevance compared to a real aircraft electrical installation:
1. OFF-LINE bench (Figure 4): it is a purely topologic bench, not powered, wired with real electrical cables
but without any load. The types and lengths of the cables (up to 30 meters) were characteristic of
cables used in various aircrafts. The purpose of this first bench is to characterize the cables with
different cable defects. Some parts of the harnesses can be replaced by defective ones.
2. ON-LINE bench (Figure 5): this bench, illustrative of the electrical installation and distribution aspects,
is equipped with powered loads. The purpose of this second bench is to validate the performances
with powered cables and the coupling with a protection device (AGFCB: Arc Ground Fault Circuit
Breaker) and to test the ability to detect intermittent defects.
Aircraft Electrical Wiring Monitoring System ERTS² 2014
5
Figure 4 : OFF-LINE bench. Red portions can be replaced by defective ones.
Figure 5 : ON-LINE bench, equipped with real loads (lights, etc), powered by a real aircraft generator
3. M2000 bench: this bench consists of the installation of a harness manufactured with an industrial
process and of topologies representative both of the cabling rules and the electrical distribution
architectures of the 3 end users involved in the project (Airbus, Eurocopter and Dassault Aviation).
Cabling performed on a Mirage 2000 fuselage (Figure 6 to Figure 9) includes the use of several kind of
fastening (plastic or not) and paths (ceiling path, path with irregular interval between fixings).
Figure 6 : HB collars
Figure 7 : Plastic collars
Aircraft Electrical Wiring Monitoring System ERTS² 2014
6
Figure 8 : Ceiling path
Figure 9 : Irregular interval between collars
Raceways used to ensure shield continuity in case of composite fuselage were also fixed to the M2000 fuselage
(Figure 10).
Figure 10 : Raceway general installation view
The purpose of this bench is to validate performances with real electrical installation and powered loads , thus
being as close as possible of a real aircraft environment but still on ground.
Benches definition is based on the aircraft / rotorcraft manufacturers’ specifications in terms of cable lengths,
electrical distribution topologies, types of cables and connectors, wire gauges and installation rules.
Loads are chosen in order to cover a wide diversity of characteristics and different types of cabling and
electrical power (28V DC or 115V AC / 400 Hz): resistive loads such as lamps, inductive loads such as pumps,
pieces of avionics equipment. For each load, return current path is achieved by the mockup structure; which is
fully representative of a non composite aircraft installation.
Several configurations have been tested in order to evaluate the robustness of the MCTDR-based defect
diagnosis algorithms:
Longest path between the power supply unit and the highest power consuming unit,
Path with the worst topology case (electrical bus with the maximum number of derivations),
Path with the maximum number of cable segments and cut-off connectors,
Path with the maximum number of heterogeneous cable segments,
Path with the load closest to the power supply unit.
Aircraft Electrical Wiring Monitoring System ERTS² 2014
7
The different types of studied defects are open-circuits, short-circuits between two wires or with the ground
and intermittent defects (Figure 11 to Figure 13):
Figure 11 : Series Arcing Fault injection
Figure 12 : Parallel Arcing Fault injection
Figure 13 : Intermittent defect injection
3.4 Measurement results
Around 3000 different measurements were performed on the 3 test benches, according to a predefined test
procedure. The objective was to be able to evaluate the diagnosis performances for almost all possible cables
and defects configurations, representative of the 3 aircrafts types (helicopter, corporate jet aircrafts and
airliners).
A few example are shown below. The cable topology shown on Figure 4 was duplicated on the ON-LINE test
bench, with a load at connector C1 and energy supply of 28V DC. The reflectogram was measured for 3
different load configurations: normal load, open circuit (load was removed) and short-circuit (cables were
connected). Figure 14 compares the 3 reflectograms: the curves change at distance close to 30 meters, which is
the length of the line under test. The presence of voltage does not prevent from detecting the defect, and the
signatures of open and short-circuits can be easily recognized. The peaks situated between distance 0 and the
defect come from the impedance mismatch at the connection of the electronic board and the cable, and also
from additional components and metallic structures between the board and the load. All these impedance
changes create noisy signals that bounce several times before coming back at the reflectometer. However, as
we are dealing with embedded diagnosis, the important feature for defect detection is the difference between
the current measurement and a reference one, which is dynamically updated to take into account small and
slow variations of the wire’s state. A fast change, such as the appearance of a defect at the load can easily be
isolated by simple comparison.
Aircraft Electrical Wiring Monitoring System ERTS² 2014
8
Figure 14: 3 different defects on a wire, ON-LINE testbench
Intermittent defects have been generated by specifically designed devices, and tested according to the
appropriate standards (AS6019 for 28VDC & AS5692 for 115VAC).
Figure 15 shows a burst of the first 4 measurements done at the detection of an intermittent open circuit
defect leading to a series arc fault on the M2000 test bench (Figure 11). The red curve is the last result just
before the signal went above the detection threshold, and then come the green, blue and purple curves. Each
measurement lasts for 160µs due to the averaging and oversampling processes. The difference between each
curve is due to the fact that this defect lasts longer than the measurement phase: the detection device can
monitor the appearance, the “steady state” and the disappearance of the defect, just as a camera would do.
Figure 15: Intermittent defect detection
Aircraft Electrical Wiring Monitoring System ERTS² 2014
9
4. Way forward
Through this in-depth test campaign, the consortium has shown the adequacy of embedded reflectometry
systems to avionics issues, and the technology developed in the project was labeled at Technology Readiness
Level 4 (TRL4), i.e. the MCTDR technology was validated in laboratory environment, although very close to
operational conditions (i.e. TRL close to level 5).
Based on the encouraging results of this first step, a second step is envisioned in order to increase readiness up
to TRL6. Its goal is to validate a system approach, with reflectometry sensors embedded as an ASIC in electronic
equipment, such as electrical power distribution center or Circuit Breaker, and globally or locally in order to
detect and locate defects, even intermittent, during the flight.
Possible enlarging of the use of this wiring defects diagnosis technology to communication links (i.e. digital
buses) will be also a new challenge of this second phase.
The results obtained during this second step will offer to trucks, aircraft or rotorcraft development programs
new opportunities to take benefit of defect localization techniques. The main milestones are:
Retrofit of an existing aircraft by only replacing in the harness the current AGFCBs by smart AGFCB
versions embedding the cable defect localization function (wire by wire management),
Test new vehicles’ electrical architecture with an embedded Harness BITE system, including:
o Wiring defects detection and localization sensors integrated in vehicle’s Electrical Power
Distribution Units or communication buses switches,
o Sensors management and measures exploitation functions, including false alarms filtering (in
case of normal avionics units power off for example),
Tests improvement thanks to the coupling of the toolkit with the vehicle’s ground database for a
better wiring defect localization and quicker maintenance.
A longer term challenge for embedded electrical harness maintenance would be defect prognostic, taking into
account wiring ageing.
5. Conclusion
To increase the reliability of wired networks or to ease their maintenance, MCTDR was identified as the most
promising method for embedded EWIS diagnosis. Based on the injection into the network of a multicarrier
signal which respects EMC and harmlessness constraints, this method provides information for the detection,
localization and characterization of electrical defects (or mechanical defects having electrical consequences) in
the wired network.
The application to avionics has been studied in this project, gathering a consortium of large industries such as
Airbus, Dassault Aviation, Eurocopter, Labinal, Latelec, Crouzet and, Zodiac with upstream research from CEA
LIST. Next step is to cover both avionics and automotive (through Volvo group’s involvement in the project).
Location accuracy was evaluated close to 20 cm for twisted wires but was degraded for one wire above metallic
ground, for up to 30 meters long cables. Branched networks are difficult to diagnose above 3 branches, but this
only concerns less than 5% of wires in aircrafts. The ability to detect intermittent defects of less than 1 ms
duration was proven.
The use of Off the Shelf components and an FPGA for embedded data processing paves the way for future ASIC
design and system integration.
Aircraft Electrical Wiring Monitoring System ERTS² 2014
10
6. Bibliography
[1] F. Auzanneau, “Wire Troubleshooting and Diagnosis: Review and Perspectives”, PIER B Journal, Vol. 49,
p. 253-279, 2013.
[2] P. F. Fantoni, "Wire System Aging Assessment and Condition Monitoring Using Line Resonance Analysis
(LIRA)", NPIC&HMIT Conference, November 12-16, 2006, Albuquerque, USA
[3] C. Furse et al., “A Critical Comparison of Reflectometry Methods for Location of Wiring Faults”, Journal
of Smart Structures and Systems, vol. 2, n°1, p. 25-46, 2006.
[4] A. Lelong, M. Olivas, “On Line Wire Diagnosis using Multicarrier Time Domain Reflectometry for Fault
Location”, IEEE Sensor Conference, August 26 – 28, 2009, Christchurch, New Zealand
[5] J. Wang et al., “Health Monitoring of Power Cable via Joint Time-Frequency Domain Reflectometry”,
IEEE Transactions on Instrumentation and Measurement, vol. 60, n° 3, 2011.
[6] W. Ben Hassen et al., “On-Line Diagnosis using Orthogonal Multi-Tone Time Domain Reflectometry in a
Lossy Cable”, IEEE International Conference on Systems, Signals and Devices, March 18-21, 2013, Hammamet,
Tunisia
[7] M. Olivas et al., “Feasibility of the detection of vibration induced faults in connectors by
Reflectometry“, 24th International Conference on Electrical Contacts, June 9-12, 2008, Saint Malo, France.
[8] A. Lelong, “Distributed Reflectometry Method for Wire Fault Location using Selective Average”, IEEE
Sensor Journal, Vol.10, Issue 2, 2010
[9] M. Franchet et al., “Modeling the effect of a defect on crosstalk signals under the weak coupling
assumption”, PIERS Proceedings, p. 119 - 123, March 22-26, 2010, Xian, China.
[10] J. Guilhemsang et al., “Method for detecting and locating defects by reflectometry in a wired electric
network and corresponding device”, patent number WO 2009087045 A1.
7. Acknowledgement
The results presented in this paper were obtained during the research project Harness BITE, a part of the
@MOST project dedicated to aircraft maintenance led by AIRBUS and partially funded by the French
Directorate General for Civil Aviation (DGAC).
