FL

IG

HT

A

IR

WO

RT

HI

NE

SS

S

UP

PO

RT

T

EC

HN

OL

OG

YA I R B U S T E C H N I C A L D I G E S T

26

AIRBUS INDUSTRIE

N U M B E R 2 6 S E P T E M B E R 2 0 0 0

1

This issue of FAST has been printed on paperproduced without using chlorine, to reduce wasteand help conserve natural resources.Every little helps.

Dialogue

Portable Equipment for testing radomesBernard CARAYON

Methodology for analysis of operational interruption costsJean-Pierre POUBEAUEmmanuel HERINCKX

A340-500/-600 Fuel systemRoss WALKER

Modern technology - What has changed?William WAINWRIGHT

Worldwide Airbus Customer Services

Changes to technology

Cover photo: first A340-600 wing and centre-fuselage section

2

3

9

15

23

30

32

© AIRBUS INDUSTRIE 2000. All rights reservedThe articles herein may be reprinted without permission except

where copyright source is indicated, but with acknowledgement to Airbus Industrie. Articles which may be subject to ongoing review must

have their accuracy verified prior to reprint. The statements made hereindo not constitute an offer. They are based on the assumptions shown and

are expressed in good faith. Where the supporting grounds for these statements are not shown, the Company will

be pleased to explain the basis thereof.

Editor: Denis DEMPSTER, Product MarketingGraphic Design: Sylvie LAGRE, Customer Services Marketing

Telephone: +33 (0)5 61 93 39 29E-mail: fast.digest@airbus.fr

Telex: AIRBU 530526FTelefax: +33 (0)5 61 93 27 67

Photo-engraving: Passion GraphicPrinter: Escourbiac

FAST may be read on Internet http://www.airbus.comunder Customer Services, Publications

FAST / NUMBER 26

S e p t e m b e r 2 0 0 0

A I R B U S T E C H N I C A L D I G E S T

FAST / NUMBER 26

Airbus Industrie and its Suppliershave always helped the Operators ofits aircraft to get the best out of themthrough dialogue. There is a constantwish of all parties to improve and anessential part of this improvementprocess is by holding and attendingoperators’ conferences and liaisonmeetings. This is a continuousprocess that benefits the airlines indifferent ways. For example Airbusaircraft enjoy very high levels ofDispatch Reliability and low DirectMaintenance Costs; an innovativespare parts investment policy has sig-nificantly lowered capital investment;with the Airbus fly-by-wire conceptabout 50% of the airliners being soldtoday have the same systems, samecockpit and same operating proce-dures. This allows greatly reduced

flight crew and maintenance trainingcosts which lead to Cross CrewQualification (CCQ) and Mixed FleetFlying (MFF) and provides unheardof levels operational flexibility to theairlines.

Two recent conferences covered theA330 and A340, and Training. Threeothers will take place in the comingmonths covering the A320 family,Performance and Operations, andA300/A310.

The study of human behaviour andcultural effects in the cockpit environ-ment are particularly important toAirbus and the airline industry. To thisend Airbus run a series of very popu-lar regional conferences on HumanFactors. The 11th was in Melbourneearlier this year and the 12th will be inthe USA in the autumn.

Coming upA319, A320, A321 Technical

SymposiumSevilla, Spain, 4-8 Dec 2000

The Airbus A320 product line isthe world’s fastest-selling jetlinerfamily. More than 1,300 aircraftfrom the A320 family have beendelivered to airlines and operators

worldwide. Airbus Industrie’s upcomingTechnical Symposium in Sevilla, Spain willprovide operators with an update of the tech-nical status on the A320 family in service,and give them the opportunity to report ontheir experience and share it with others.

11th Performance andOperations Conference

Puerto Vallarta, Mexico26-30 March 2001

This conference, held on a twoto three year rotational cycle,is organised in different partsof the world, and will be heldnext year in Latin America. It

provides a dialogue between AirbusIndustrie and the operators of Airbus aircrafton all operational aspects.

12th Human FactorsSymposium

Aspen, Colorado, 17-19 Oct 2000This symposium is being held in cooperationwith Human Factors Committees of the AirTransport Association and various US AirlinePilots’ Associations for management, pilots’unions and Human Factors specialists from USairlines. The themes for discussion are HumanFactors issues in design, training, operationsand safety. The symposium is structured tominimise formal presentations and maximiseinteractive discussions in ten panels.

Recent Events4th A330/A340 Technical

SymposiumCairo, Egypt, 22-26 May 2000

The purpose of this confer-ence was to present tech-nical solutions for sub-jects raised by the opera-tors, and provide a

forum for discussionbetween the operators

themselves, and with Airbusand the suppliers. In this respect it was verysuccessful with 371 attendees from 55 air-lines and a number of suppliers. A CD-ROMcontaining all the presentations, questionsand answers is available to CustomerAirlines from their Customer SupportManagers.

5th Training SymposiumToulouse, France 22-

26 May 2000The importance of goodtraining cannot be overemphasised and this

importance was reinforcedby the strength of the atten-

dance. 434 people representing94 airlines, 10 Airworthiness Authorities and72 suppliers attended. Besides getting peopletogether to discuss current and future train-ing philosophy and techniques, it was anexcellent opportunity to display advancedtraining equipment to a professional commu-nity whose advice and comments wereinvaluable. A CD-ROM containing all thepresentations and panels is available uponrequest.

11th Human FactorsSymposium

Melbourne, Australia4-6 July 2000

Airbus Industrie is an indus-try leader in human fac-tors’ aspects of aircraftdesign training and oper-ations. Regional meet-

ings have been regularlyheld since 1995. This

Symposium was organised forairlines in the Asia Pacific region with thecooperation of Ansett Australia and theAustralian Transport Safety Bureau. Its spe-cific objectives were to diffuse the Airbusculture, showing its rationale for design andautomation, philosophy for training andoperations and its balanced approachtowards prevention and reaction in safetymanagement.

To reflect the importance of culturalaspects in the cockpit, panels were organisedon the development of a “Culture ofOrganisational Safety”, and “Cultural needsfor Crew Resource Management (CRM)”.

160 persons, from airlines, civil aviationauthorities and organisations, airforces anddefence organisations, universities, accidentinvestigation boards and even a railwayauthority participated in this regional forum,having ample time for discussion andexchange with experienced experts in theirfields.

Prominent airline pilots and academicspeakers from the region were also involvedthroughout the symposium.

Dialogue

A330/A340 Technical

Symposium

Cairo 22-26 May 2000

AIRBUS INDUSTRIE

Designedby

Cu

stom

erS

ervices Marketing

Airb

us

Industrie © 2000 All rights

rese

rved

5th AIRBUSTRAININGSYMPOSIUM

5th AIRBUSTRAININGSYMPOSIUM

5th AIRBUSTRAININGSYMPOSIU

TOULOUSE 22-26 MAY 2000

AIRBUS INDUSTRIECUSTOMER SERVICES

©A

irbus Industrie 2000 All rights

rese

rved

11th AIRBUS HUMAN FACTORS SYMPOSIUM

MELBOURNEJuly 4th-6th 2000

CUSTOMER SERVICES Flight Operations Support

AIRBUS INDUSTRIE©

Airb

usIndustrie 2000 All rights reserved

2

FAST / NUMBER 26 3FAST / NUMBER 26 3

By Bernard CarayonSenior Engineer

Navigation Systems Engineering ServicesAirbus Industrie Customer Services

Reasons fortesting the

radomeThe function of theradome is to protect theantennas installed in thenose of the aircraft fromairflow, rain, hail, light-ning strike, bird strike,etc. At the same time itmust provide a radio-fre-quency transparent win-dow, suitable for themicrowave signals of theweather detection radar,instrument landing sys-tem and microwave land-ing system. The mostdemanding system interms of radome radiofrequency specification isthe weather radar.

Weather detectionmode

In the weather detection mode, the radardetects precipitation of water or iceparticles present in clouds. Weathertargets are colour coded in function ofthe precipitation intensity (drop size,density, reflectivity of target). The crewinterprets the resulting picture. It istherefore of prime importance that thepicture be as accurate as possible andrepresentative of the actual precipitation.Only this will allow the crew to make theright assessment and take theappropriate flight path to avoid adverseweather, turbulence and lightning strikes.

Predictive windsheardetection mode

In the predictive windshear detectionmode, the radar detects small raindrops“flying” in opposite directions(microburst event). It generates an auralwarning and displays the area where thewindshear is present on the navigationdisplay. The predictive windsheardetection system identifies microburstsup to five nautical miles ahead, givingup to sixty seconds of advance warningto the crew. This allows the crew toavoid entering a windshear, thus furtherincreasing aircraft safety in the air.However, the detection must be highlyreliable in order to avoid falsewarnings.

One way to reduce those falsewarnings is to limit false echoes due topoor quality of radomes. This is thereason why radomes installed onaircraft equipped with a predictivewindshear detection system, mustcomply with more stringentspecifications.

To ensure optimum operation of theweather radar system, the quality of theradome in terms of radio frequencyperformance is critical.Note: With the technology availabletoday windshear and turbulence cannotbe detected in dry air conditions.

How a degradedradome affects

the radar systemWeather detection

The radar transmits a radio-frequencypulse. Then it listens for the returnpulse reverberated by an obstacle (rain,terrain) and measures its elapsed time.The distance between aircraft andobstacle is proportional to the elapsedtime. The larger size and reflectivity ofthe target means a stronger return.

A low transparency to radiofrequency of a radome directly affectsrange detection and target sizecomputation. This could end up under-estimating a weather hazard. Themeasured signal must be the onereturned by the main lobe of theantenna. A return signal from secondarylobes (also called “side lobes”) maygenerate false weather targets fromground echoes.

FAST / NUMBER 264

A turbulence encounter is a play featuring three characters: the atmosphere, the aircraft and the pilot (whether a human pilot or an auto-pilot). The purpose of this article is to review the respective role and contribution of these three actors, through the main aspects associated with flying in severe turbulence at altitude.

Most of the considerations addressed in this article are general in nature and are equally applicable to the A300/A310/A300-600 and to the A320/A330/A340 aircraft families.

Whenever applicable, specific considerations are given for non-fly-by-wire and fly-by-wire models respectively.

by Michel TREMAUD, Manager ETOPS Performance and OperationsFlight Operations Support, Customer Services Directorate

FAST / NUMBER 18 7

Side lobes

Ground return = echo at aircraft altitude

Main lobes

The most recent type of weather-radarfeatures two functions: weather

detection function and predictive windshear detection. Both functions are

essential to a smooth flight.(see FAST magazine number 18

‘Flight In Severe Turbulence’).

FAST / NUMBER 26 5

Windshear detectionThe radar processor detects the Dopplerfrequency shift of the microwave pulsescaused by the microbursts of winddirection. Return microwave signalsmust not be disturbed, to allow themeasurement of direction and speed ofthe precipitation of the tiny droplets. Soa poor transparency to radio frequencyfrom the radome may lead toundetected windshear.

As for the weather function, returnsignals from secondary lobes maydisturb the radar system and cause theradar to trigger false windshear alarms.Therefore both functions demand anundistorted forward view in terms ofradio-frequency performance. That is, ahigh-performing radome in terms ofradio-frequency transparency and levelof side lobe capability.

Sources ofradome

degradationThere are many reasons for theperformance level of a radome todegrade below minimum requirements.They range from lightning strike, birdstrike, natural erosion, water ingress forsome types of radomes, ageing, poorrepairs such as wrong material used, orover-painting.

Characteristics tobe checked

There are two main characteristics to bechecked to ensure that the radome willnot affect the weather radar system:

◗ Its transparency to radio-frequency waves. Transparency relates to unobstructed forward view in terms of radio frequency to minimize the loss of outgoing and return signals.

◗ The side lobe level. Side lobe level difference between main beam and side beams of the antenna ensures that the return microwave signal being measured comes from the main beam and not from the side beams. So the measured return waveis the desired wave from the front of the antenna and is undisturbed. Both are defined in the RTCA document DO213 in chapters 2.2.1 and 2.2.2.

Transmission efficiency

The average and minimum transmissionefficiencies within the window areashould not be less than indicated for thefollowing classes:

Side lobe levelThe radome should not increaseantenna side lobe levels. For side lobelevels between –21dB and –26dB, theincrease should not be more than 2dBfor Category 1 requirements.

Average Minimum

Class A 90% 85%Class B 87% 82%Class C 84% 78%Class D 80% 75%Class E 70% 55%

Part Number Technology Aircraft family Class Cat

A923 20242 000 10 Glass fiber A300 C 2

A923 20242 000 11 Glass fiber A300 C 2

A923 20242 000 12 Kevlar A300/ A310/ A300-600 B 1

A923 20606 005 00 Kevlar A300/ A310/ A300-600 B 1

A923 20606 006 00 Kevlar A300/ A310/ A300-600/ A330/ A340 B 1

A923 20606 007 00 Kevlar A300/ A310/ A300-600/ A330/ A340 B 1

A340 32201 000 00 Quartz A300/ A310/ A300-600/ A330/ A340 A 1

A340 32202 000 00 Quartz A300/ A310/ A300-600/ A330/ A340 A 1

D531 10477 000 02 Kevlar A320 B 1

D531 10477 000 03 Kevlar A320 B 1

D531 10477 000 04 Kevlar A320 B 1

D531 10477 000 05 Kevlar A320 B 1

D531 10477 000 07 Quartz A319/ A320/ A321 A 1

Status of newAirbus Industrie

radomes

FAST / NUMBER 266

Methods, means& procedures

availableTo check the performance level of aradome, a number of techniques aredescribed in the ARTC 4 and RTCAMOPS D0213 documents such asanechoic chamber, test in ‘free’ area,two-horn transmission test tools.However, all of them require extensive,costly means and specialised personnel,quite unaffordable to most operators. Inaddition, a considerable burden isplaced upon the operators to send theradome to specialised repair stationssometimes halfway across the world. Atool ‘light’ enough to be usable inairline facilities and cheap enough wasdefinitely required.

History of toolsdevelopment

Airbus Industrie and Aerospatiale-Matra initiated work on thedevelopment of the portable radometest equipment in 1990 on airlinerequest. First, a review of all thetechniques used to measure the radio-frequency characteristics was carriedout to assess the most promisingmethod. It was decided that only amethod using a single horn, about sixcentimeters (2.3 inches) in diametercould meet the requirements of theoperators. Yet, the method had to befirst demonstrated as workable.

Consideration was given for a robotto move the horn over the radomesurface. Measurements would beperformed along the path of the horn.This solution was quickly abandoneddue to its prohibitive cost. It was thendecided to use a single hand-held hornincluding both transmitter and receiverand to manually move the horn point-by-point in close contact over theradome surface, performing themeasurements all along a path of theradome surface to be covered.Obviously each type and size of radomewould require a specific path to bedeveloped. The collected data wouldthen be converted to meaningful data todetermine the performance level of theradome.

The data collected from themeasurements does not provide a directreading of the radio-frequencyperformance level of the radome. Therelation (mathematical model) betweenthe collected data and the transparencylevel with side lobes level had to bedetermined. Airbus Industrie then testeda high number of radomes both with a

prototype tool and with a referencedanechoic chamber. The data collectedfrom the prototype of the ‘light’ testtool were then compared with the datacollected from the anechoic chamber. Atwo-day run on a Cray 2 computersorted out the data.

To carry out a test of the radome’sradio-frequency performance level, twosets of measurements have to becollected:

◗ Reflection coefficient in open circuitwhich, once converted, gives an indication of transparency.

◗ Reflection coefficient in short circuit which, once converted, indicates the size of the side lobe levels.

A conversion formula was compiledand the software was written for use ona laptop computer. Various types ofradomes were again tested with theAerospatiale Matra Airbus Test &Services portable test equipment andwith a referenced anechoic chamber.The two sets of results were comparedto enable the validation of the principleof the portable test equipment. In thecourse of the development differentlevels of radome degradation weresimulated on an A320 aircraft. Theresults met the official requirementslaid out in various documents such asARTC 4, RTCA D0213, ARINC 708,ARINC 708A… and therefore is a clearqualification of the design.

The portable test equipment is now inindustrial production and is availablefor purchase. (See page 8).

Radome test toolprocedure...

FAST / NUMBER 26

Transparency Measurement

1 The radome is placed on a layer ofradio-frequency-absorbing foam toavoid return echoes from the floor.

2 The grid is positioned over theradome (see photo below).

3 The operator checks the serviceabilityof the tool and calibrates it.

4 The operator does the measurementspoint-by-point following the instructionsdisplayed on the screen of the horn(see photo below).

Phase shift measurement

1 The operator places an electricallyconductive sheet or foil inside theradome. It is used to reflect themicrowaves. It must be maintained inclose contact with the radome innersurface, so a rubber sheet is placed ontop of the foil and sealed around theedges. By vacuuming the trapped airbetween the rubber sheet and radomeinner surface, the atmospheric pressurewill gently maintain the foil pressedagainst the radome surface.

2 The serviceabilityof the tool is againchecked.

3 A new set of mea-surements is per-formed.

4 The PC computesthe serviceability ofthe radome and thereport of the test maybe printed out for filingand traceability pur-poses as required bythe regulations.

The equipment is composed of:

◗ A hand-held horn which includes a radio-frequency transmitter/receiver and a converter to intermediate radio frequency. A liquid crystal display on the back of the horn provides instructions to the operator.

◗ A radio-frequency analyser to pick up the measurements.

◗ A portable computer (PC) to collect, control and process data, and manage the interface to the operator.

Referenced radome samples areprovided to check the serviceability ofthe tool and to calibrate it prior to anyradome testing. This ensures goodrepeatability of the measurements. Theequipment only requires a standard110/220V AC, 50/60 Hz power supply.No specific room is required. However,a sheet of radio-frequency-absorbingfoam (ref API 28 from Hyfral, vendorcode FAFK8 or equivalent) is requiredfor the open circuit measurement. Aconductive sheet or foil (kitchenaluminum foil) and a rubber sheetalong with a vacuum bag and vacuumpump (ref ANITA NG9201 made byGMI, vendor code F07856 orequivalent), as used for compositerepairs are also required for the shortcircuit measurement.

7

Brief descriptionof the equipment

The test procedureThe procedure is divided mainly in two parts:

transparency measurement and phase shift measurement.

...One set of measurements to compute the radio-frequency transparency.One set of measurements to compute sidelobes level.

FAST / NUMBER 26

ConclusionAerospatiale Matra Airbus Test & Services’ portable radome test equipment providesoperators with a low cost in-house capability to test radomes for serviceability and toidentify non-acceptable areas. The test procedure using this tool will be referenced in therelevant Airbus technical documentation such as the Component Maintenance Manual(CMM) and the Structure Repair Manual (SRM).

Main advantages for the operator are:◗ Short out-of-service time for the radome

◗ No specific test facilities needed

◗ Test equipment easily transportable (less than 45kg)

◗ Test equipment easy to set-up

◗ Less than one day per radome for testing by one technician

◗ On-site testing and repair (see FAST magazine number 18 ‘Infrared Thermography for In-Service Inspection), applicable to all Airbus Industrie radomes

◗ No transportation cost to specialised repair centres

◗ No cost of sub-contract test and repair

◗ Reduced quantity of spare radomes �

Portable test equipment optional features

Development is still continuing and the following optional features are foreseen:

◗ Graphic display of non-acceptable area of the radome

◗ Radome life-time tracking

◗ Point-by-point measurement to directly compare two areas of the radome

Price & delivery quotationOperators who wish to receive a price quotation for the Radome Portable Test Equipment,part number LCRAD0100HM0100, are invited to send their requests to Aerospatiale MatraAirbus, Test & Services in Toulouse France, fax: +33 (0)5 61 93 03 09.Note: The vacuum pump, the conductive foil, the rubber sheet and the radio-frequencyabsorbing foam can be procured from the suppliers mentioned on page 7.

8

For any questions or complementary information, please contact:

Mr. Bernard CARAYONSENIOR ENGINEER

NAVIGATION SYSTEMS ENGINEERING SERVICESAIRBUS INDUSTRIE

Fax: +33 (0) 5 61 93 44 25

e-mail:bernard.carayon@airbus.fr

Emmanuel HERINCKXM.Sc. Air TransportCranfield University

Jean-Pierre POUBEAUDepartment Manager

Maintenance Economics AnalysisCustomer Services

Methodologyfor Analysis

of OperationalInterruption Costs

Methodologyfor Analysis

of OperationalInterruption Costs

9FAST / NUMBER 26

ObjectiveOperational disruptions result from many different causes. The International Air TransportAssociation already lists more than 70different factors that cause delays. Thisresearch has been designed by AirbusIndustrie with the aim of providingairlines with a methodology to clearlyidentify and appraise the costs hiddenbehind the various operationalinterruptions.

As clearly presented by AEA, theAssociation of European Airlines,technical faults only account for a smallproportion of total disruptions (see Fig.1 above).

However, this proportion ofoperational interruptions, althoughsmall, is under airlines’ responsibilityand control. This is what they canimprove. This is also where themanufacturer can bring all its expertiseand support to further improve airlines’on-time performance.

QuestionnaireTo undertake the above research, twodifferent questionnaires were designedand sent to target both passengercarriers and integrators (packagecarriers) respectively. Nearly 65different Airbus operators werecontacted to provide a large enoughsample on which to base the analysis.

The questionnaires included sevensections related to the differentoperational interruption types classifiedas follows:

Ground Interruptions:◗ Flight dispatch delays◗ Ground turnbacks ◗ Aborted take-offs ◗ Aircraft substitutions◗ Flight cancellations

(at main base/out-station)

Air Interruptions:◗ Air turnbacks◗ Diversion

A checklist of cost items, adapted toeach particular type of operationalinterruption, was provided within theappropriate section. The purpose ofthese checklists was to give airlines theability to specify which costs and theirrespective percentage of the totalinterruption costs were included in theircalculations. A presentation of thesection related to delays is given in Fig. 2.

Introduction Airlines are continuously under pressure to improve their punctuality (i.e: on-timeperformance), setting ambitious and very challenging objectives. Increased awareness, new generations of travellers and changing attitudes have ledto a change in demand. Punctuality has become one of the most significant factorsfor defining a passenger’s satisfaction with an airline.

Delays, cancellations and other operational interruptions cost airlines money.Passenger ill-will, associated losses of revenue and compensations are among thesecosts. Operational interruptions also pose extra costs to airlines due to higher fuelburn, additional maintenance, crew and aircraft utilisation. Airlines are extremelyconcerned about controlling costs.

FAST / NUMBER 26

Fig. 1 Main causes of departure delays over 15 minutes

10

…

Handling, operations & technical

Airports & ATC

Weather

Q1 Q 2 Q 3 Q 4 Q 1 Q 2 Q 3 Q 4

% of flights delayed

1997

05

10

15

20

25

30 1998

Source: Association of European Airlines, Yearbook 1999

Each airline has its own specificenvironment, route and station network,aircraft utilisation, maintenance conceptand other operational specifications.Furthermore, it has its own marketingstrategy, targeting specific customers,with dedicated levels of services.Therefore, it seems obvious thatoperational interruption costs differ fromairline to airline, depending on theirrespective marketing and operationalspecifications.

Figures 3 and 4 present the delaycosts, as reported by the airlines, andclearly show how big the differencesbetween operators can be. Due to a lackof information we could not make anyspecific conclusion about thedifferences between geographic areas,their economic environments, betweendifferent fleet sizes, or even betweenscheduled and charter operations.

From such a small sample ofrespondents, it is very difficult toconclude anything about an averagedelay cost. The large amplitude ofreported delay cost values would

certainly distort any average figure. Wecame to the same conclusion aboutcancellation costs.

The chart on figures 5 and 6 presentthe reported cancellation costs for bothsingle-aisle and wide-body operations.Reported cancellation costs fromsingle-aisle operators are within thesame band, with the exception of Sce 7in figure 5. This gives a good idea ofthe cost of cancelling a single-aisleflight

This is far from being true for wide-body reported values. Therefore, tryingto apply average delay or cancellationcost values to all airlines would beinappropriate. However, what would beappropriate is that each airline shouldcarry out its own internal operationalinterruption costs analysis based on itsspecific operation type. In that respect,Airbus Industrie has identified amethodology for delay and cancellationcost analysis, and provides explanationson some of the driving factors affectingthese costs.

AnalysisRelated to passenger-carriers

FAST / NUMBER 26 11

Questionnairefeedback

Very good feedback was received frommany airlines that found thesechecklists helpful and good startingpoints for internal analyses.

Operators, wishing to undertake theirown internal analysis, may obtain thelists from the address at the end of thisarticle. It could be of interest to manycarriers to undertake such analysis, asonly 13 airlines were at this time able toparticipate in this survey.

The aim of the questionnaires was tocollect information regarding existingindustry models so as to analyseoperational interruption costs. Theinformation received has formed thebasis of the following analysis,providing a presentation of whatoperators considered as the drivingfactors affecting their operationalinterruption costs.

Data are based on their respectivefleet, including many different aircrafttypes from different aircraftmanufacturers. Unfortunately, out ofthe different operational interruptiontypes, only the delays and cancellationswere presented. Moreover, not enoughinformation came back fromintegrators, therefore limiting the scopeof research to passenger carriers only.

Cockpit Crew OvertimesCabin Crew Overtimes

Salaries of Stand-by CrewHotel & Meals Expenses

Other costs to be specified

Penalties for the delayOthers

Fuel CostExtra APU / Ground Unit Expenses

Stand-by Aircraft AmortizationAirport Fees

Others

Hotel & Meal ExpensesRerouting of PassengersLost of Delayed Luggage

Passenger ill-willRevenue lost from connecting pax

Others

Station / gate Agent OvertimesAirport Facility Expenses

Others

SUB-TOTAL / %

TOTAL / % 100 %COST OF A DELAY

SUB-TOTAL / %

SUB-TOTAL / %

SUB-TOTAL / %

SUB-TOTAL / %

Y NCrew

Ramp Services

Aircraft

Cargo

Passengers

Fig. 2 Questionnaires: list of cost items to consider for delaycosts

"

Reported Costs

Wide-body Operators

Sce 2

Sce 1

Sce 3

Sce 4

Sce 5

Sce 6Sce 7

Sce 8

0.0

20.0

40.0

60.0 US $ / seat / hour

Sce 1

Sce 2

Sce 3

Sce 4

Sce 5

Sce 6

Sce 7

US $ / seat Single-aisle Operators

0.0

50.0

100.0

Wide-body Operators

Sce 1

Sce 2

Sce 3US $ / seat

Sce 4

Sce 5

Sce 6

0.0

100.0

200.0

300.0

Sce 1

Sce 2 Sce 3

Sce 4

Sce 5

Sce 6Sce 7

Sce 9

Sce 8

US $ / seat / hour Single-aisle Operators

0.0

20.0

40.0

60.0

Fig. 3

Dela

y Co

sts

Canc

ella

tion

Cost

s

Crew-related expenses

The crew-related costs can be evaluatedby accounting the extra flight labourcosts (cockpit and cabin crews) incurredwhen a flight is delayed. Pilot and flightattendant salaries vary from airline toairline.

The substitution of crews by stand-bycrews also increases these expenses.

Additional hotel and meal expenses havealso to be taken into account. Theseincrease airlines’ delay costs.

The aircraft type and its associated staffrequirement also affects delay costs, andtherefore, should be considered whencalculating them.

Ramp-related expenses

These are the lowest contributing costsand should not exceed 3 or 4 % of totaldelay costs.

Only ramp-agent overtime and otheradditional airport facility expenses shouldbe included.

Aircraft-related expenses

Aircraft-related expenses are stronglydependent on operational specifications.Therefore the costs associated with fuelburn, navigation charges, extra aircraftutilisation and maintenance will varyaccordingly.

It is expected that the cost per seat willbe lower for wide-body than for single-aisle aircraft.

Long-haul fl ights also lead tosignificantly lower aircraft-relatedexpenses per seat.

When airlines have the opportunity tospeed-up their operations so as to catchup from delays, additional costs shouldbe added taking into account, forexample, the increased fuel burned.

When airlines have high frequencies of flights on a givenroute, their operations are exposed to delays in a completelydifferent way to those with low frequency operations. Delayscan have significant knock-on effects on their operations.Furthermore, combined with short sector lengths and shortturn-around times, the opportunity to speed-up theiroperations in order to catch up remains very limited.

To cope with the problem of knock-on effect, airl inessometimes add flexibility within their fleet by operating stand-by aircraft. This of course has a cost, but leads to significantlyimproved service quality and reduction in the cost of theoperational interruption. Therefore this should also be takeninto account for the calculation. Stand-by aircraft are eitherowned by the airline, used from a pool between airlines orchartered under specific agreements.

FAST / NUMBER 2612

"

Fig. 4"

Fig. 5"

Fig. 6"

In reviewing the different delay costanalyses done by operators, twodifferent approaches have clearly beenidentified. The chart in figures 7reflects these two different approaches.It can be seen from this chart, that someairlines reported having delay costsentirely due to passenger-related costs.However, it was confirmed that it wassimply due to their way to allocate theiroperating costs. In other words, theyconsider extra aircraft and crew-relatedexpenses as part of their annual aircraftand crew budgets, remaining out of thedelay cost equation.

All reporting operators thereforerecognise that crew, ramp, aircraft andpassenger-related costs contribute totheir respective delay costs, but simplyhave different ways to allocate thesecosts. We believe that all costs incurredfrom delays should be identified asbeing part of these delay costs. In otherwords, we recommend that theoperators, wishing to appraise theirdelay costs, should apply the following

approach. This approach considers allcontributing factors and splits delaycosts between the above four relatedexpenses.The table in figure 8 presents thisapproach with an example of delay costcalculation. The presented examples ofpercentages are not appropriate for anyparticular airline. This is the reasonwhy we recommend individualanalysis. It gives, however , theappropriate list of cost items to followfor delay cost appraisal.

The contribution of each particularcost item varies from airline to airline,depending on their environmentalspecifications. To get a betterunderstanding of these differences, eachcontribution is developed individuallyin the blue columns.

To conclude this chapter, byanalysing internally these different costitems and adding them all, airlines willbe able to estimate their respectivedelay costs.

MethodologyFor delay analysis

13

Passenger-related expenses

Passenger-related expenses are among thegreatest contributing factors to delay costs.These were estimated to have a contributionof between 35% on short-haul to more than60% on long-haul flights.

These are incurred from:

a) Hotel and meal expenses:

They are greater for long-haul flights,since in most cases frequencies arelower than for short-haul flights andpassengers often have to wait for thenext daily departure.

b) Costs associated with re-bookingand re-routing of passengers:

Dealing with a delay, the bigger the air-craft, the more passengers airlineshave to handle.

c) Luggage complaints

d) Revenue losses from:

Missed-connections. The resulting rev-enue losses can be appraised by com-piling airlines’ statistics, looking at howmany passengers are expected to misstheir connection following X minutes ofdelays.

Walk-away passengers. Airlinesoperating on high density and highlycompetitive routes, where passengerscan easily opt for competing air orground services, have much greaterlosses of this type.

Passenger ill-will. A flight is commonlyconsidered delayed when it departs orarrives 15 minutes after its scheduledtime. Passenger satisfaction is depen-dent on getting to a destination ontime. This satisfaction is affected wellbelow 15 minutes, according to previ-ous surveys on the subject.

The resulting passenger dissatisfac-tion, in other words passenger ill-will,can be considered as a loss of passen-ger loyalty. Repurchase intention willthen be reduced, generating losses offuture revenue for the airlines. Theselosses are difficult to appraise andquantify. It is assumed that these pas-senger-related costs are incurred fromthe beginning of every delay, thereforefrom the first minute of delay.

To prevent dramatic passenger dissat-isfaction, some airlines offer compen-sation to their passengers. Whendelays occur, the higher the fares, thehigher the compensation levels.

Cockpit Crew OvertimesCabin Crew Overtimes

Hotel & Meals Expenses

Gate Agent OvertimesAirport Facility Expenses

$ / Seat / hour 14%

4%

TOTAL / $

Total / %

Fuel CostNavigation Charges

Stand-by Aircraft AmortizationExtra Aircraft Utilisation

Extra Maintenance

Hotel & Meal ExpensesRerouting of Passengers

Luggage ComplaintsRevenue Losses*

*Due to miss-connections, walk-away passengers & passenger ill-will

Crew

Ramp Services

Aircraft

Passengers

48%

34% $xx

$xx

$xx

$xx

$xx

Fig. 8 Items contributing cost to delays

Fig. 7 Reported delay cost drivers

FFAST / NUMBER 26

…

…

Sce 1 Sce 2 Sce 3 Sce 4 Sce 5/s Sce 5/l Sce 6 Sce 7

% of Total delay cost

0

20

40

60

80

100

Crew

Passenger

Aircraft

Ramp

FAST / NUMBER 26

MethodologyFor cancellation analysis

The table in figure 9 below presents themethodology used by airlines toappraise their cancellation costs withexamples of percentages. It highlightsthat cancellation costs are driven bypassenger-related costs. However, someoperational savings from not operatingthe flight reduce cancellation costs.

Depending on airlines’ type of opera-tions and their respective environment,passenger-related expenses and theamount of operation cost savings areexpected to vary significantly. Thereforeit is recommended that all operatorsshould do their own calculation based ontheir respective operation type.

For example, domestic operators basedin North America face significantlylower airport and navigational chargescompared with European operators,leading to lower operating costs and,therefore lower potential savings.Moreover, the savings can also bedifferent depending on when thecancellation occurs. Catering sometimescannot be saved if it is already on boardthe aircraft.

When do cancellations occur? Thequestion is: “Are there any alternatives?”As reported by the airlines, when dealingwith a delay, if it is expected to last forsome time and alternatives are availableto dispatch the passengers, airlines oftenbalance the flight cancellation cost withthat of the expected delay and take theappropriate measure.

Cancellation costs are fixed costs. Onthe contrary, delay costs are running, as afunction of time, and are often reportedas costs per minute. It is thereforepossible to determine whether it is morecost effective to cancel the flight or not.The sooner the decision is taken (ifpossible prior to the scheduleddeparture), the lower the final costs. If,on the other hand, no alternatives areavailable, the cost of delays will go wellabove the cancellation costs, and delayscan last for more than 24 hours.

Applicationof delay and cancellation costs

Delay and cancellation costs can be used to evaluate the cost benefit of modifications toimprove the dispatch reliability of an aircraft fleet. A balanced computation between theairline’s cost of delays and cancellations, the cost of improvement actions, and thesavings over a specified amortisation period can be done.

In other words, by knowing the rate of delays and cancellations caused by a specificcomponent or a specific system, the average delay time they cause, their repercussion onoperational interruption costs and the necessary investment for reliability improvement,the break-even point and further cost savings can be calculated.

In this respect, all operators wishing to carry out such analyses will find AirbusIndustrie’s computerised Service Bulletin Cost Benefit Model of much help. It isavailable on request from the address below, free of charge for Airbus Industrie’scustomers.

Single-aisle

TOTAL: $ / SEAT

Total / %

Hotel & Meal ExpensesRerouting of Passengers

Luggage ComplaintsRevenue Losses

Hotel & Meal ExpensesRerouting of Passengers

Luggage ComplaintsRevenue Losses

Passengers-related cost

Total cost savings

130%

-30%

$xx

$xx / seat

14

Fig. 9 Items contributing cost tocancellations

"

Airlines, airframe and engine manufacturers, and equipment suppliers allhave a degree of responsibility and control of operational interruptionsdue to technical reasons. Reduction in the number of operationalinterruptions can lead to substantial cost savings. However, there isalways a cost associated with these improvements. Therefore, this costshould be properly balanced with the potential resulting savings.

Working closely with the manufacturers and suppliers, operationalreliability can be improved, to the benefit of the airlines and to thesatisfaction of their passengers. The method presented for the analysis ofdelay and cancellation costs can help airlines appraise their operationalinterruption costs, and therefore, identify future reliability improvementbenefits �

ConclusionFor any questions or complementary

information, please contact:

Mr. Jean-Pierre POUBEAUMAINTENANCE ECONOMICS

ANALYSIS DEPARTMENT

Fax: +33 (0) 5 62 11 01 56

e-mail:jean-pierre.poubeau@airbus.fr

The fuel system of the A340-500/-600 aircraft differssignificantly from the A340-200/-300. The principal

reason for the difference is the change of the wingdesign resulting in an increase in the wing sweep.The effect of this is to change the trajectory of any

debris from an uncontained engine rotor failure,preventing the use of tank boundaries similar to

those on the A340-200/-300.

For certification reasons the boundaries for theengine feed tanks and associated system

architecture had to be changed.

Additional changes to the system architecture havealso been made as a result of the requirement for

increased refuelling flow rates (400,000 litres/hour).

FAST / NUMBER 26 15

By Ross WALKERSenior Fuel Systems Engineer

Hydro-mechanical Systems, Airbus Industrie

ENG1

OUTER

LEFT INNER RIGHT INNER

OUTER

R

Q

F

K

J

EA

X

ENG2

2

CA

M

ENG3

ENG4

ZE

H

Y

N

3S

EB

CENTRE

TRIM

B D

GG

A E

TL

APU

S U R G E

S U R G E

1

V W G

P

4

SU

RG

E

Inner Two

Centre Spar

Inner One

Outer

Surge

Inner Three

Inner Four OuterSurge

Centre

Trim

Rear CentreTank (RCT)

(-500 only)

Collector Cells(One per Inner Tank)Rib 41

Rib 35

Rib 23

Rib 11

Rib 4

Rib 1

SU

RG

E

Fuel storageOn the A340-500 fuel is stored in nine fueltanks: Three Tanks in each wing, comprising:

One ‘Outer’ wing transfer tank; Two Engine feed tanks called

‘Inners’. (Within each ‘Inner’tank is a dedicated engine feed collector cell);

One Transfer tank in the centre wing box called the ‘Centre’ tank;

One Transfer tank in the horizontal stabiliser called the ‘Trim tank’;

One Transfer tank positioned at the forward end of the rear cargo hold. This tank is called the ‘Rear Centre Tank’ or ‘RCT’.

The A340-600 has only eight tanks, theRCT not being fitted.

RefuelThe refuel system isdesigned to refuel theaircraft from empty to fullin 33 minutes on the A340-500 and 30 minutes on theA340-600, when a refuelpressure of 3.45 bar isapplied at all four of therefuel couplings. (A slowerrefuel time will be obtainedif the refuel pressure islower). The four couplingsare fitted as two pairs, onepair fitted to each wing.

FAST / NUMBER 26FAST / NUMBER 2616

A340-500/-600 Fuel tanks

At each wing tip and on the right wing tip of the horizontal stabiliser is asurge tank. These tanks do not normally contain fuel but are used as partof the tank ventilation system to link the fuel tank system to the outsideair and prevent fuel spillage during refuelling, aircraft manoeuvres or dueto thermal expansion of the fuel. These tanks are the interface to outsideair through flame arrestors and flush non-icing inlets.

17

"

Left Outer Right OuterInner 1Inner 2Inner 3Inner 4CentreTrimRCTTotal

A340-500 A340-600

Volume Mass Volume Mass

litres US gals kg lb litres US gals kg lb

6,360 1,680 5,088 11,216 6,360 1,680 5,088 11,2166,360 1,680 5,088 11,216 6,360 1,680 5,088 11,216

25,690 6,786 20,552 45,305 25,690 6,786 20,552 45,30533,280 8,792 26,624 58,698 33,280 8,792 26,624 58,69833,280 8,792 26,624 58,698 33,280 8,792 26,624 58,69825,690 6,786 20,552 45,305 25,690 6,786 20,552 45,30555,980 14,788 44,784 98,729 55,980 14,788 44,784 98,7298,238 2,177 6,590 14,534 8,238 2,177 6,590 14,534

19,930 5,265 15,944 35,151 Not fitted214,808 56,746 171,846 378,854 194,878 51,481 155,902 343,703

"

The total aircraft and individual tankusable fuel capacities

DifferencesBelow, you have a summary of the differences forA340-500/-600 fuel system:

◗ An increase in the number of fuel tanks

◗ A revised ECAM fuel system page

◗ Secondary wing refuel and transfer galleries

◗ Dedicated fuel jettison/transfer & engine feed pumps

◗ Separate APU and trim transfer lines

◗ Segregation of computing and fuel probe interface functions into separate boxes

◗ (-500 only) A fuel tank positioned between the forward end of the rear cargo hold & the centre landing gear bay

◗ Change of vendor to Parker for fuel control monitoring system

FunctionsThe functions of the A340-500/-600 fuel systemremain similar to the A340-200/-300:

◗ Fuel storage and tank venting◗ Engine and APU fuel feed supply◗ Fuel quantity indication◗ Temperature indication◗ Refuel/Defuel control◗ Transfer control◗ Centre of gravity control◗ Fuel level sensing◗ Fuel system advisories and warnings◗ Maintenance BITE indication◗ Fuel jettison

In summary

REFUEL/DEFUEL VALVESAIRCRAFTR Left refuel isolationS Right refuel isolationBM Auxiliary refuelBN DefuelINNERBA Inner 1 inletF Inner 2 inletH Inner 3 inletBB Inner 4 inletOUTERM Outer inletN Outer inletCENTREG Centre inletGG Centre restrictorBC Inner 1 transferBG Inner 2 transferBH Inner 3 transferBD Inner 4 transferTRIMW Trim pipe isolationL Trim inletRCT CA RCT isolationCB RCT inlet

TRANSFER PUMPSCENTRECentre leftCentre rightTRIMLeft trim transferRight trim transferTrim transferRCTRCT forwardRCT aft

TRANSFER VALVESOUTERQ Left outer transferP Right outer transferTRIMT Trim isolationV Trim forward transferRCTCC RCT transferCD RCT auxiliary transferAFT B Left aft transferD Right aft transfer

JETTISON VALVESX LeftY Right

JETTISON/TRANSFER PUMPSInner 1 jettison/transferInner 2 jettison/transferInner 3 jettison/transferInner 4 jettison/transferLeft Centre jettison/transferRight Centre jettison/transfer

ENGINE/APU FEED PUMPSMAINEngine 1 mainEngine 2 mainEngine 3 mainEngine 4 mainSTANDBYEngine 1 standbyEngine 2 standbyEngine 3 standbyEngine 4 standbyAPUAPU forward feedAPU aft feed

ENGINE / APU FEED VALVESISOLATIONEC Transfer control 2ED Transfer control 3EA Left InnerEB Right InnerAPUJ APU LPK APU isolationLP1 Engine 1 LP2 Engine 2 LP3 Engine 3 LP 4 Engine 4 LPCROSSFEEDC Crossfeed 1A Crossfeed 2E Crossfeed 3Z Crossfeed 4

(For mass figures a density of 0.8 kg/litre = 6.676 lb/US gallons is used)

A340-200/-300

"

A340-500/-600 "

All A340A340-500/-600

A340-500A340-200/-300

TText legendext legend

Fuel system

INNER THREE INNER FOUR

OUTER

ENG1R

X

ENG2

M

SURGE

ENG1

OUTER

INNER TWO

INNER ONER

FBA

X

ENG2

2

CA

M

ENG3

ENG4

E

Y

N

BB

EDZ

3

BH

H

P

4

SCENTRE

TRIM

BMBN

GG

A

TL

APU

S U R G E

S U R G E

1

K

RCTA340-500ONLY

J

CBCC

EC

Q

BG

E

BC BD

V W CD CA G

SU

RG

E

FAST / NUMBER 2618

Refuel/defuel control panelControl of the refuel function is bythe two FCMCs (Fuel Control andMonitoring Computers) and twoFDCs (Fuel Data Concentrators). Theoperator interface is either throughthe standard refuel panel fitted in thelower surface of the fuselage just aftof the undercarriage bay, or throughthe optional refuel panel and MCDU(Multipurpose Control and DisplayUnit) in the cockpit.

Normal refuel is fully automatic, however in the event ofsystem failures a manual refuel facility is availablethrough the standard refuel/defuel control panel.

The automatic refuel distribution isdefined in two steps in a first stage byspecific masses for each tank followed bya second ‘top-up’ phase to the volumetrichigh level for all tanks.

Refuel distribution

A340-600 Refuel Distribution (All Densities 0.74 - 0.88 kg/litre)

The Outer, Trim and RCT tanks each havea single tank inlet valve. To minimise anysurge pressures in the refuel gallery theCentre and Inner tanks have two valveswhich are closed in sequence. The Centretank has a tank inlet valve in series with arestrictor valve and the Inner tanks eachhave two tank inlet valves in parallel.

To enter the aircraft, fuel must passthrough one of the two refuel isolationvalves which form part of the refuelcouplings fitted to each wing. To preventspillage of fuel, the detection of fuel withineither of the wing tip surge tanks or aJettison valve detected open willautomatically stop the refuelling of thecomplete aircraft by closing the refuelisolation valves.

In addition, detection of the fuel in thehorizontal stabiliser surge tank willautomatically stop automatic refuellingand manual refuelling of the Trim tank.

A

B

A340-500 refuel/defuelControl Panel "

60

50

40

30

20

10

100 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 2000

FOB 12 tonnes - All Inners 3 tonnes (start to fill Outers)FOB 21.4 tonnes - Outers 4.7 tonnes (continue filling Inners)FOB 85.8 tonnes - All Inners 19.1 tonnes (continue filling Inners 2&3 & start filling Centre & Trim)FOB 109.4 tonnes - Inners 2&3 24.6 tonnes (continue filling Centre & Trim)FOB 144.3 tonnes - Centre 41.4 and Trim 6.1 tonnes (start to fill RCT)FOB 159 tonnes - All tanks max. volume with min. density

(all tanks filled simultaneously to max. vol.)FOB 189 tonnes - All tanks max. volume with max. density

OUTERINNERS (1&4)INNERS (2&3)CENTRETRIMRear Centre Tank (RCT)

Ind

ivid

ual

Fu

el T

ank

Mas

s To

nn

es

Total Fuel on Board (FOB), Tonnes

60

50

40

30

20

10

100 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 2000

FOB 12 tonnes - All Inners 3 tonnes (start to fill Outers)FOB 21.4 tonnes - Outers 4.7 tonnes (continue filling Inners)FOB 85.8 tonnes - All Inners 19.1 tonnes (continue filling Inners 2&3 & start filling Centre & Trim)FOB 109.4 tonnes - Inners 2&3 24.6 tonnes (continue filling Centre & Trim)FOB 144.3 tonnes - All tanks max. volume with min. density

(all tanks filled simultaneously to max. vol.)FOB 171.5 tonnes - All tanks max. volume with max. density

OUTERINNERS (1&4)INNERS (2&3)CENTRETRIM

Ind

ivid

ual

Fu

el T

ank

Mas

s To

nn

es

Total Fuel on Board (FOB), Tonnes

A340-500 Refuel Distribution (All Densities 0.74 - 0.88 kg/litre)

A & B - The automatic refueldistribution for the complete rangeof fuel densities and fuel quantities"

Engine feedUnder normal operation each engine isfed by an independent fuel feed system.This consists of main and standbyengine feed booster pumps locatedwithin a collector cell, which in turn islocated within an engine feed tank(Inner). The main pump operatescontinuously, the standby pump onlyoperates if the main pump becomesdefective or is set to OFF.

The collector cells are maintained fulluntil the Inner wing tanks are nearempty (see fuel transfer section) to helpensure a supply of fuel to the engineunder negative ‘G’ manoeuvres. Thecollector cells are maintained full bythe use of jet pumps driven by fuel flowtaken from the main engine feedbooster pumps.

All engine feed systems can be joinedto the crossfeed gallery by theirindependent crossfeed valves. Thecrossfeed system is used underabnormal operational conditions suchas loss of all electrical power requiringgravity feeding or to connect allengines to a single engine feed boostpump when only the emergencyelectrical supply is available or to allowthe crew to correct an imbalancebetween symmetrical wing tanks.

In common with all other AirbusIndustrie programmes, all fuel pumpand valve electrical wiring is routedoutside of the fuel tanks to eliminatethe potential for introducing ignitionsources into the fuel tanks.

Note: Identical pumps are used forengine feed, jettison and transfer on allA340s.

APU FeedThe APU is fed via a dedicated linefrom a tapping off the number oneengine fuel feed line. The number oneengine booster pumps normally supplythe fuel pressure. However, if thesepumps are not selected then a dedicatedAPU pump is installed in the line tosupply the fuel pressure.

JettisonA jettison system is provided to avoidthe necessity for heavy maintenancetasks, by providing a means tominimise the potential for anoverweight landing. (The system is notrequired for certification reasonsrelating to the performance of theaircraft). The system is activated bymeans of two dedicated pushbuttons‘Arm’ and ‘Active’ located on thecockpit overhead panel. The system canbe manually stopped by de-selection ofeither of these pushbuttons, orautomatically if all jettison/transferpumps are running dry (low pressure)or when the fuel quantity drops below apredetermined target input into theMCDU by the flight crew.

The system jettisons fuel through twojettison valves positioned in the numberthree flap fairing between the twoengines on each wing. Up to tenjettison/transfer pumps are used toprovide the fuel flow, one situated ineach of the four Inner tanks and two inthe centre tank, two in the Trim tankand Two in the RCT. The refuel galleryis used to connect the jettison valves tothe pumps. The Centre, Trim and RCTpumps only function if the associatedtank contains fuel. In addition, if theOuter tanks contain fuel, a transfer ofthe fuel to the Inner tanks isautomatically initiated.

Fuel transfers & usage

On the A340-600 under normaloperation all fuel transfers, except thosefor centre-of-gravity control, are to thefour Inner tanks, prior to transfer to thecollector cells, and are controlledautomatically. Automatic transfers arecontrolled to balance the fuel quantitiesin symmetrical wing tanks to prevent animbalance which could adversely effectthe aircraft handling. On the A340-500transfers from the RCT are to theCentre tank. On the cockpit overheadpanel four pushbuttons are provided toallow manual transfer control of theOuters to Inners, Centre to Inners, Trimto Centre and RCT to Centre.

FAST / NUMBER 26 19

Centre ofgravity control

In order to minimise the aircraft’saerodynamic drag in cruise the fuelsystem is used to optimise the aircraft’sangle of attack by controlling theaircraft’s aft centre-of-gravity (CG).Depending on the aircraft’s zero fuelweight (ZFW), CG and the actual fuelon board, the fuel system will controlthe aircraft’s CG to a target of 2% meanaerodynamic chord (MAC) forward ofthe certified aft limit.

Control is achieved by means of fueltransfers to and from the Trim tank andin addition on the A340-500 transfersfrom the RCT. Aft transfers to the Trimtank are performed if the tank is not fulland the aircraft’s CG is forward of thetarget. Forward transfers from the Trimtank are performed if the aircraft’s CGdrifts aft, due to fuel burn, of the target.On the A340-500 forward transfersfrom the RCT are delayed until theaircraft’s CG drifts aft, due to fuel burn,of the target or the centre tank has lessthan 11 tonnes of fuel. (At the end ofcruise all Trim tank fuel is transferredforward).

For integrity of the system, anindependent CG monitor is performedby the flight management system.

Fuel quantitymeasurement &

level sensingUnlike all other Airbus aircraft, the fuelquantity measurement/indication andlevel sensing functions are combinedand use similar capacitive type probesto perform the two functions. The probeexcitation and return signal processingis performed within two independentFDCs. The conditioned data from theFDCs is sent to both FCMCs, whichprocess the data for fuel quantityindication (FQI), fuel level indication,refuel control, CG calculations andcontrol and general fuel transfercontrol. Automatic compensation isapplied for changes in fuel density, fuelpermitivity and aircraft effectiveattitude.

The probes are configured into twoseparate groups per tank and theinterfaces through the FDCs andFCMCs are designed so that any singlefailure (e.g. probe, harness, FDC orFCMC) will not cause a loss ofindication.

FAST / NUMBER 2620

A340-500 fuel usage& transfer sequencing

The normal tank usage sequence isin the following order:

1 Centre

2 RCT

3 Trim

4 Outers

5 Inners

6 Collector cells

forward refuel/transfer gallery andindependently controlled tank inlet valves, onefor each Inner tank. In the event of certainfailures, transfers are by means of two pumpslocated in the Centre tank which areconnected to the aft refuel/transfer gallery andindependently controlled tank inlet valves, onefor each Inner tank.

◗ Transfers from the Trim to Inners are bymeans of two pumps located in the Trim tank,two valves situated at opposite ends of theTrim tank transfer line and independentlycontrolled tank inlet valves, one for each Innertank, connected to the aft refuel/transfergallery. If the centre tank contains usable fuelthen trim transfers will be to this tank.

◗ Transfers from the Outer tanks to the Innertanks are by gravity. For transfers to Innertanks one and four the refuel inlet valves, forthe concerned tanks and refuel gallery, areused. For transfers to Inner tanks two andthree dedicated transfer valves and associatedpipework is used.

◗ Transfers from the RCT to the Centre tankare by means of two pumps located in theRCT and two valves, one valve situated at eachend of the RCT refuel/transfer line.

◗ Transfers from the Centre to the Inner tanksare normally by means of two pumps locatedin the Centre tank which are connected to the

C G _Z F W _

(FOB) TonnesFlight Time (minutes)

5000

0

0 60 120 180 240 300 360 420 480 540 600 660 720 780 840 900 960 1020172 162 153 143 133 124 114 105 95 86 76 66 57 47 38 28 18 9

10000

15000

20000

25000

30000

35000

40000

45000

50000

INNERS (1&4)INNERS (2&3)CENTRERCTTRIMOUTERS

Ind

ivid

ual

Tan

k F

uel

Qu

anti

ty (

kg)

FAST / NUMBER 26 21

Fuel temperaturemeasurement

Dual element in-tank temperature sensors are fitted to the Trim, Outerand Engine feed tanks. These sensors enable the flight crew to monitorthe evolution of the tank fuel temperatures to ensure that they are withinthe operational limitations for the specific fuel types being used (e.g.,JET A or JET A1).

Fuel quantity measurement& level sensing (cont’d)

Within the normal ground standingattitude range of the aircraft theaccuracy of the FQI system will be inthe order of 0.4% of the full capacitynear empty to 1% at full.

To further enhance the security of thefuel system designs on Airbus aircraftthe potential for an ignition source to bepresent in a fuel tank has beenmitigated by the following measures:

◗ The harnesses to the capacitiveprobes are segregated from all otheraircraft wiring ;

◗ The length of the harnesses isminimised (the FDCs being fittedwithin the centre fuselage section).

These measures limit the potential for ashort circuit to power cables.

In the case of a complete failure of thefuel control and monitoring system(FCMS), a secondary manual fuel levelindication system is installed in the sixwing tanks and the centre tank forground use. Height data from theprobes is used in conjunction withaircraft attitude information, fueldensity measurement and a set of lookup tables to calculate the fuel mass inthe tank being measured.

The level sense function is used todetect high and low fuel level states inall fuel tanks and a fuel overflow in thethree surge tanks.

This information is used to control:

◗ Refuel◗ Inter tank transfers◗ Transfer pump shut-down◗ Overflow protection◗ Indication of low fuel state.

FAST / NUMBER 2622

On the ECAM fuel page fuel systemdata is displayed including:

◗ fuel quantity◗ fuel temperature ◗ fuel transfers◗ pump status◗ fuel used◗ fuel flow.

Warnings and the total fuel on boardare displayed on the engine page.

Due to the differences in the fuel sys-tem architecture, the fuel system con-trol panel and ECAM system page dif-fer from the A340-200/-300.

Despite the differences, there is still apushbutton for each engine feed andtransfer pump, for each crossfeedvalve and each transfer function, as

well as a dedicated tog-gle switch associatedwith the Trim tank isola-tion, as can be seen onthe main fuel systemoverhead panel. Undernormal operation, afterinitialisation at the startof a mission, no crewaction is required on thepanel. Manual transfercontrol is by selection ofthe dedicated transferpushbuttons or the dese-lection of the transferpump pushbuttons.Adjacent to the mainpanel are the pushbut-tons for jettison and theInner 2 and 3 wing tankisolation valves.

For any questions or complementary information, please contact:

Mr. Ross WALKERSENIOR FUEL SYSTEMS ENGINEER

AI/EE-M HYDRO-MECHANICAL SYSTEMS

Fax:+33 (0) 5 61 93 28 98

e-mail:ross.walker@airbus.fr

A340-500/600 Cockpit Hydraulic &Fuel Control Panel

A340-500 only

A340-500/600 ECAMfuel page

"

"

Cockpit control panels& indication

ConclusionThe A340-500/-600 have been designed taking

into account lessons learnt from the in-serviceexperience of the A340-200/-300 �

FAST / NUMBER 26 23

“My view on technology haschanged a lot over the years. I seethis particularly with my attitudetowards my cars. I have alwaysliked cars, particularly fast cars.When I was younger I spent mostof my weekends working on them;tuning carburettors or tweakingelectrical points. Now, even when Ibuy a new car, quite a time passesbefore I open the bonnet, and thenit is just to marvel at the technicalbeauty of the modern engine.I never touch it. It even tells mewhen I have to add some oil. Asimilar evolution has taken placein aviation, but I believe that wehave not yet realized how it hasrevolutionized our life and how weshould change our attitudetowards modern technology. In Flight Test, we are often askedto do an investigation after acomplaint about system behaviour,and we sometimes find in the dataanalysis report, that it wasworking ‘as designed’. I believethat this is because we have notalways understood how the adventof digital systems has changed theway that automatic systemsbehave. I would like to explain mypoint of view. My intention is toprovoke some thought, andperhaps some discussion, abouthow to adapt the way that we livewith and think about automaticsystems”.

By William WAINWRIGHT Chief Test PilotAirbus Industrie

This is similar to the revolution thatoccurred with watches. The watch that Ireceived for my 21st birthday was aSwiss-made chronometer of a veryfamous mark. It is a mechanical marvel,complicated, and expensive. It kept timewonderfully, with remarkable accuracyfor many years. But then it became lessaccurate – it needed cleaning.Nowadays, even after cleaning, it issomewhat erratic. Modern watches canbe much simpler mechanically. Theinvention of the quartz watch haschanged everything. They are not onlyvery accurate when new, but there isnothing in them to wear out or get dirty. Of course, they can still break down, butit is most likely to be a sudden failure,rather than a gradual and slow

deterioration. Modern aircraft systems aresimilar. Software does not

wear out or get dirty. Italways does the samething under the same

circumstances.This is not to deny

that computers donot sometimes have‘glitches’. But that

is different, and I willtalk about it later.

And of course, the electrical compo-nents can fail, so that aircraft computerssuddenly stop working. But they do notstart working more slowly, or less effi-ciently. Generally, if they do somethingthat you do not like, it is because theyhave been programmed to do it. It is notbecause they are not working correctlyand should be changed.

As I said in my introduction, we oftenfind cases where computers have beencriticised, and after examination, it isproved that they were working ‘asdesigned’. This is often in airline ser-vice, but it also occurs during ourproduction testing, and customers some-times try to reject systems, such asautopilots or autothrottles, which areworking correctly, although not in theway that they would like them to do.I would like to take a few examples toillustrate my point.

FAST / NUMBER 2624

Computer rack in aHawker Siddeley Trident

"

from cogsto software…

Changesin technology

Not so long ago sophisticated aircraftsystems, such as autopilots andautothrottles, were made up ofcomplicated linkages involving cogs,cams, springs, and gears. They couldclog up, seize up, slip a gear, or justslow down because they werecontaminated.

They are now made up of variouselectrical components, and somesoftware.

This has all changed...

Computer rackin an A340

#

Suddenfailure!

Software DesignDesigning any flight control system is acompromise. I will take autothrust as anexample. To have good accuracy ofspeed tracking you need a high gainsystem. However, ahigh gain systemis susceptible toovercontrolling, whichin the extreme case maylead to instability,which in this example might beoscillations in power setting. And evenwith a system free of instability,frequent thrust variations in the cruisewill increase fuel consumption andannoy the passengers. Thus, you haveto find a compromise that will give youreasonable comfort together withacceptable accuracy. For example, inthe cruise we set a reasonably high gaininitially, allowing quite high thrustvariations, so that the autothrust systemcan quickly find a mean thrust settingto maintain the selected speed.

Thereafter, a lower gain is used tominimize thrust variations whilstallowing a looser speed tracking withvariations in speed of up to +/-4kt. Ifthe speed goes outside this bracket, dueto the wind suddenly changing, thesystem progressively switches back toits high gain until the stabilisedsituation is regained. Thus, in certainconditions speed tracking may be lessaccurate than some pilots think itshould be.

ExtremeConditions

Furthermore, if a large windshear isencountered, particularly when flying atspeeds near Vmo/Mmo, we need theautothrust to act quickly to prevent orminimize an overspeed. Thus, it mayhappen that the autothrust variesbetween high and low power settingsfor a short time. Even though therelatively high gain that is used is stillrather low by comparison with what apilot would have used in thesecircumstances to do his initialcorrection, we cannot change the gain,as he would do, after the possibility ofoverspeed has been avoided.

In addition, we always avoid usingvery high gains, which may causeinstability. Therefore, we usecompromises which will work verywell in most cases, but which might notbe optimum in some extreme cases.

It may even happen that the enginesreduce to idle to avoid the speedincreasing too quickly. Then, when theshear reverses, and thrust has to beincreased, the engine acceleration willbe relatively slow, as with all modernbig-fan engines at high altitude, suchthat a speed excursion occurs briefly inthe opposite sense. This behaviour hasbeen criticised when it has occurred inservice because it was considered to bedue to a faulty autothrust system. Thismay have been true in the past, withautothrottles full of gears, cams, leversetc., but this is no longer true of digitalsystems. We have designed ourautothrust system to work that way, forvery good reasons, which may notalways be obvious to the pilot.

FAST / NUMBER 26

Gain is a numberin a formula usedin a control law.High gain means ahigh multiplicationof a parameter.

With modern technologyComputers, having the same hardware,and the same software, will always workin the same way. Thus, changing theautothrust computer to another of thesame standard, will have no effect. Nomatter how many new boxes you try, theonly way to correct what has happenedis to change the software. Unlike somepopular perceptions, this is not an easymatter and it is certainly not cheap. Wehave designed our software as it is forgood reasons, and it is always based ona compromise solution. Of course,sometimes we realise that we couldhave done it in a different, perhaps abetter, way. But you always have to bevery careful when you change thesoftware to correct something that youdo not like, that you don’t make it worsesomewhere else. This is why, when wechange something, we always do a lotof extra flight tests to prove that there isno degradation.

Computers work‘asdesigned’...

25

Characteristics ofAutomatic Systems

“I believe that one of the reasons why it is not always obvious to thepilot that a system is working ‘as per design’, when it is working below

his expectations, is that we all expect a little too much of modernautomatic systems”.

Also, we sometimes meet conditions where we havenever seen them working before. They work better than

the human being in some aspects, but not as well inothers. For example, they never get tired, and thusthey can follow a speed target for hours with

excellent accuracy if the conditions do not change.But they cannot adjust their strategy when the

conditions change. In fact, they are less adaptablethan human beings.

Autoland was developed to land an aircraftwhen the pilot could not see where he wasgoing; blind landings in fog. This put theemphasis on landing the aircraft in a reason-

able touchdown zone. The com-fort of the landing was of sec-ondary importance. Nowautoland is being used in manydifferent conditions, and on manydifferent runways, which are notalways ideally suited to automat-ic landings, because they haverising ground before them, orcliffs, or they have significantslopes.

Unlike a human pilot, theautopilot cannot change theway it flies to cope with dif-ferent local conditions. It can-not change its prioritiesbetween making a smoothlanding when the weatherconditions are good and land-ing relatively firmly but rightin the centre of the touchdown

zone when they are bad. It cannot decideto change its flare height to suit a risingrunway, nor can it change its techniqueto cater for thermal activity. Of course,we could attempt to give them a lot ofcomplicated logic that would try to caterfor this. But this may cause problemselsewhere, because it is often better tohave simple logic. And you have to becareful not to have a system whichcopes perfectly with all the fringe situa-tions but does a poor job on a good run-way in real Category 3 conditions.

Thus, although the dispersion intouchdown distance is low, the disper-sion in the way the flare is done and thetouchdown rate of descent can be quite

large, particularly when landing on dif-ficult runways or in difficult weatherconditions. A variation in touchdownsbetween 2ft/second and 5ft/second isnot unusual. The autopilot should neverdo a ‘kiss’ landing. This is because wedesign it to touchdown in the correctlanding zone. Therefore, we have tomake sure that it will not float.

For this reason we have incorporateda sort of anti-long flare. If after a certaintime, the aircraft has not touched down,it is programmed to make a slight nose-down pitch change to search for theground. This means that in certain con-ditions, which cause the aircraft to havea tendency to float, it will suddenlypitch nose-down to touchdown relative-ly firmly. In fact the touchdown will stillbe within our normal dispersion, but itmay disappoint the pilot who wasexpecting something better, and it maybe rather untidy.

FAST / NUMBER 2626

Automatic systemsare less

adaptable…

They neverget tired!

The realityWe had a customer’s pilot, come toaccept a new aircraft, who asked tore-fly his aircraft after such anautoland. The repeat test was doneat the same airport in the sameweather conditions and the samething happened again. Anotherflight was requested. The pilot wasexpecting a perfect performancefrom his new aircraft, which he hadevery right to expect, but in fact, anew aircraft will not do betterautolands than an older aircraft. Theform of autoland depends on theexternal conditions and not on theage of the computer, provided it hasthe latest software.

Adaptabilityin

autoland

Blind landing in fog

"

What neverchanges

The flying characteristics in normal lawnever change. They depend on a set ofsoftware, and as I have already said,that always works in the same way. Atleast unless, or until, we make a changeto it. You may or may not like all thecharacteristics, and occasionally youmay be surprised by something that youhad not noticed before. But changingthe computers to others of the samestandard will do nothing.

During a customer acceptance flighton an A340, the customer’s pilotsuddenly complained about the rate ofroll. He was doing a part of theprogramme that checked the flyingcharacteristics in normal law. This callsfor some more vigorous manoeuvringthan is normally done in line service,and he found the roll rate to be less thanhe expected. I said that it was probablybecause he was not used to using such alarge sidestick input. Perhaps he waslooking for an instantaneous 15° persecond roll rate, as might be interpretedfrom the book. Whereas we have todesign the aircraft, with all its inertia, tohave roll characteristics which startwith a gentle acceleration initiallybefore arriving at the maximumstabilised roll rate, which will alwaysbe 15° per second.

I have to admit that we ask thecustomer’s pilot to verify the handlingcharacteristics during their acceptanceflight. We also do this during our ownproduction test flights. This is an easyway to manoeuvre the aircraft a bitmore vigorously than would normallyhappen in service with passengersaboard. In fact, we are verifying that noother anomalies occur. We are notreally checking roll rates or maximumbank angles. They are always the same,although some small variations may beseen according to the exact way inwhich the manoeuvres are done.

What canchange

Although the software will not change,an external parameter that is used bythe software may be deficient, eitherpermanently or temporarily. A goodexample is the transition into anotherphase of a law – both in manual flightand in autoland. This may be a reasonto either change the equipment or tochange the software by producing anew standard.

We have also seen cases where aircraftdid not pass into flare law duringmanual landings. The aircraft remainedin its normal ‘g’ and pitch rate law forthe landing, which is less dramatic, butstill requires us to do some correctiveaction. However, in this case, workingon the aircraft would have beenineffective; a change of design wasnecessary. In fact, it was due to the wayin which the transition was triggeredand the way the radio altitude signalwas sampled. We discovered that aparticular runway profile could cause aperturbation in the signal around theheight for changing the laws whichmeant that the transition did not takeplace. In fact, the aircraft over-flew alighting post at the exact instantwhere it passed below100ft. At that time thetransition wastriggered by theaircraft passing100ft when indescent. Thiswas hidden bythe momentaryblip in altitudecaused by thepost. The curewas to changethe triggermechanism to bewhen below 100ftand not when passing100ft.

FAST / NUMBER 26

they always workin thesame way...

Rectification actionWe had one case on a productionfirst flight of an aircraft that passedinto the de-rotation phase of an

autoland whilststill airborne.The test pilothad to interveneto prevent a hard

landing. It was due to an error in thewiring that caused the autopilot tosee the wheels as turningpermanently. As soon as the radioaltitude part of the condition wasseen, the aircraft started its de-rotation. This is a good example ofwhy autolands must be done duringproduction test flights and is a casewhere rectification action wasnecessary.

27

De-rotation is thetime betweentouch-down of themain wheels andtouch-down of thenose wheels.

We can consider the three Primaryflight control computers on theA330/340 as three generals who justhappen to be identical tripletscommanding identical army divisions(called Prim 1,2,3). Being identicaltriplets they all think and work inexactly the same way and have thesame reactions to each and everysituation. Having identical armydivisions under their command theyeach have identical resources to copewith any situation. But in any welldisciplined army, only one general canbe the supreme commander. This isthe case for our flight control system.

Only one Prim is in command at anyone time. It passes its orders to theothers who continue, however, tocalculate their own commands whichthey keep to themselves. Of course,being identical and having access tothe same information, the orderscalculated within each Prim areidentical. Thus, if one Prim falls sickand passes the command to one of itsbrothers, the brother continues to giveorders which will be identical to thosewhich would have been given by thefirst Prim if it had not fallen sick. In

fact, only one Prim is required to beactive to maintain full normal lawcapability, provided of course thatboth Secondary computers areserviceable.

Incidentally, the Secondary computers(Sec) are dissimilar to the Prims inboth hardware & software. One ofthem can control the aircraft on itsown. It is as if the three generals hadtwo twin colonels in reserve, lesscapable, with smaller forces, butabsolutely reliable.

This was a case where one of the

Fail softThis latter example is a very goodillustration of our philosophy whichleads us to design monitoring circuitsin the way that we do. Our philosophyis quite clear; it is to never leave a sickcomputer in charge of normal law.

Thus, the monitoring circuits aredesigned so that the slightest doubtleads to the relevant computerdeclaring itself faulty and droppingoff-line. Therefore, failures are soft,and we have eliminated the hard-overfailures which were a significant riskfactor with previous generationaircraft. This may lead us to have moretechnical defects, but the effect on theaircraft and the pilot will be minimaldue to the redundancy that is built intothe flying control systems.

FAST / NUMBER 26 29

How our ‘network’ worksWe can consider the three Primaryflight control computers on theA330/340 as three generals who justhappen to be identical tripletscommanding identical army divisions(called Prim 1,2,3). Being identicaltriplets they all think and work inexactly the same way and have thesame reactions to each and everysituation. Having identical armydivisions under their command theyeach have identical resources to copewith any situation. But in any welldisciplined army, only one general canbe the supreme commander. This isthe case for our flight control system.

Only one Prim is in command atany one time. It passes its orders tothe others who continue, however, tocalculate their own commands whichthey keep to themselves. Of course,being identical and having access tothe same information, the orderscalculated within each Prim areidentical. Thus, if one Prim falls sickand passes the command to one of itsbrothers, the brother continues to giveorders which will be identical to those

which would have been given by thefirst Prim if it had not fallen sick. Infact, only one Prim is required to beactive to maintain full normal lawcapability, provided of course that bothSecondary computers are serviceable.Incidentally, the Secondary computers(Sec) are dissimilar to the Prims inboth hardware and software. One ofthem can control the aircraft on itsown. It is as if the three generals hadtwo twin colonels in reserve, lesscapable, with smaller forces, butabsolutely reliable.

FAST / NUMBER 2628

“System failures aresoftware…”

Computer faults“I have to admit that computers are not infallible, and they can stop working

correctly. I imagine that we have all experienced a home computer that has blocked andhas to be re-started to clear itself. Of course a similar situation can occur to an aircraftcomputer. But it is important to realize which part of the computer can block, and whichwill always work exactly in the same way.”

We had one case where all threePrimary Computers were lost, and Iknow that various explanations arecirculating, as rumours do, whichhave nothing to do with what actuallyhappened. But, before I tell you whatdid cause this incident, I will explainhow the system works.

To illustrate my point, even though Iam not a software expert, I will use anexample from the world of homecomputers. As I understand it, it is thesoftware that manages the computerthat may block. The programmesoftware always works as designed.For example, in terms of the homecomputer, it would be an interactionbetween programmes that mightprovoke a blockage whereas each

individual programme, such as Wordor Excel, always works as it is meantto. In aviation terms, the equivalent ofWord or Excel might be the flightcontrol law software, which I havealready said will always work ‘asdesigned’.

The equivalent of Windows, or anyother management software, is theCommand/Monitoring structure thatsupervises each computer and managesthe system, which consists of severalindividual computers, just like anetwork. I will give an illustration ofhow our ‘network’ works, and anexample of a ‘glitch’, which has nowbeen cured by a re-design of themanaging software.

“…and hard-overshave been eliminated”

Summary and conclusionIn conclusion, software design is a compromise and may notalways please everyone. But the programme software, such asthat used in Normal Law or autopilot or autothrust, will alwaysbehave in the same way in an identical setof circumstances. Failures that occur tocomputers are usually in thesupervising and management part ofthe system. They will cause thesystem to fail soft. We have asimilar situation with aircraft tothat prevailing in the car industry.You no longer have to spend yourweekends tuning carburettors ortweaking electrical points.Thanks to fuel injection andelectronic ignition your car willgive maximum performance justuntil there is a failure in thehardware, usually in the electronic circuit. You don’t have anyredundancy on your car; it just stops working. In aircraft, systemsfailures are soft and hard-overs have been eliminated �

Tuning & tweaking!

For any questions or comments, please contact:

Mr. William WAINWRIGHTCHIEF TEST PILOT

Fax: +33 (0) 5 61 93 29 34

e-mail:william.wainwright@airbus.fr

The cured ‘glitch’We had one case where all three Prims were lost whereas only onePrim was sick. This was due to a weakness in the monitoring process.To ensure that we never leave a sick computer in charge of normallaw, each Prim continuously sends messages to its brothers aboutthe state of its health. In this case, when Prim 1 started to feel sick it stammered and didnot get its message out straight away. Prims 2 and 3 hearing a garbled message thought that they werethemselves sick, and they declared themselves faulty. Prim 1 then got his message out to say that he was faulty and washanding over control, but there was no one left to hear him. Thus, the aircraft passed into Direct Law under the command of itstwo Secondary computers until Prims 2 and 3 were reset.

This was a case where one ofthe Prims (Prim 1) wasdefective and neededchanging. However, changingPrims 2 and 3 for new oneswould not prevent thissituation happening again,because they would beidentical to the ones replaced.It was an unlikely occurrencethat does not happen everytime a Prim is faulty. However,it could not be ruled out that itwould never happen again. Thesolution lay in re-designing themonitoring process, which hasbeen done.

FAST / NUMBER 26 31FAST / NUMBER 2630

RCSM LOCATION COUNTRYABU DHABI United Arab Emirates AMMAN Jordan ATHENS Greece BOMBAY IndiaBANGKOK Thailand BEIJING People’s Republic of ChinaBEIRUT LebanonBERLIN GermanyBRUSSELS BelgiumBUENOS AIRES Argentina CAIRO Egypt CHARLOTTE USA (North Carolina) CHENGDU People’s Republic of ChinaCINCINNATI USA (Ohio)COLOMBO Sri Lanka COPENHAGEN DenmarkDAKAR Senegal DHAKA BangladeshDAMASCUS Syria DELHI India DERBY England DETROIT USA (Michigan)DOHA QatarDUBAI United Arab Emirates

C U S T O M E R S U P P O R T

USA / CANADAThierry van der Heyden, Vice President Customer ServicesTelephone: +1 (703) 834 3484 / Telefax:+1 (703) 834 3464

CHINAEmmanuel Peraud, Director Customer ServicesTelephone: +86 10 6456 7720 / Telefax: +86 10 6456 76942 /3 /4

REST OF THE WORLDMohamed El-Borai, Vice President Customer Support Services DivisionTelephone: +33 (0) 5 61 93 35 04 / Telefax:+33 (0) 5 61 93 41 01

RESIDENT CUSTOMER SUPPORT ADMINISTRATIONPhilippe Bordes, Director of Resident Customer Representation Administration Telephone: +33 (0) 5 61 93 31 02 / Telefax:+33 (0) 5 61 93 49 64

T E C H N I C A L , S P A R E S , T R A I N I N G

Airbus Industrie has its main spares store in Hamburg, Germany, and subsidiary stores at Frankfurt, Germany, Washington D.C., Beijing, China, and Singapore.Airbus Industrie operates 24 hours a day every day.

AOG technical and spares calls in North America should be addressed to:Telephone +1 (703) 729 9000Fax +1 (703) 729 4373

AOG technical and spares calls outside North America should be addressed to:Telephone +49 (40) 50 76 3001 / 3002 / 3003Fax +49 (40) 50 76 3011 / 3012 / 3013

Airbus Industrie's main training centre is locatedat Toulouse, France.Telephone +33 (0) 5 61 93 33 33Fax +33 (0) 5 61 93 46 65

It has major training subsidiaries located at Miami, FloridaTelephone +1 (305) 871 36 55Fax +1 (305) 871 46 49

and Beijing, ChinaTelephone +86 10 64 57 33 40Fax +86 10 64 57 09 64

RCSM LOCATION COUNTRYDUBLIN Ireland DULUTH USA (Minnesota)DUSSELDORF Germany FRANKFURT Germany GUANGZHOU People’s Republic of ChinaHANGZHOU People’s Republic of ChinaHANOI VietnamHELSINKI FinlandHONG KONG People’s Republic of ChinaINDIANAPOLIS USA (Indiana)ISTANBUL Turkey JAKARTA Indonesia JINAN People’s Repuplic of ChinaJOHANNESBURG South Africa KARACHI Pakistan KINGSTON JamaicaKUALA LUMPUR Malaysia KUWAIT KuwaitLANZHOU People’s Republic of ChinaLARNACA Cyprus LIMA PeruLISBON Portugal LONDON England LOUISVILLE USALUTON England MACAO MacaoMADRID Spain MANCHESTER England MANILA Philippines MAURITIUS Mauritius MEDELIN ColumbiaMELBOURNE Australia MEMPHIS USA (Tennessee)MEXICO CITY Mexico MIAMI USA (Florida) MINNEAPOLIS USA (Minnesota)

RCSM LOCATION COUNTRYMONASTIR TunisiaMONTREAL Canada MOSCOW Russia MUMBAI India NAIROBI Kenya NANCHANG People’s Republic of ChinaNANJING People’s Republic of ChinaNEW YORK USA (New York) NOUMEA New CaledoniaNUREMBERG GermanyPARIS (CDG) France PARIS (ORY) France PHILADELPHIA USA (Pennsylvania)PHOENIX USA (Arizona) PITTSBURG USA (Pennsylvania)QUINDAO People’s Republic of ChinaROME Italy SAN FRANCISCO USA (California) SAN JOSE Costa Rica SAN SALVADOR El SalvadorSAO PAULO BrazilSEOUL South Korea SHANGHAI People’s Republic of ChinaSHENYANG People’s Republic of ChinaSHENZHEN People’s Republic of ChinaSINGAPORE Singapore TAIPEI Taiwan TAMPA USA (Florida)TASHKENT Uzbekistan TEHRAN Iran TOKYO (HND) Japan TORONTO Canada TULSA USA (Oklahoma) TUNIS Tunisia VANCOUVER Canada VIENNA AustriaWINNIPEG Canada XIAN People’s Republic of ChinaZURICH Switzerland

FAST / NUMBER 2632

A310 static test at CEAT Toulouse in 1983

In the 1920sstatic tests ofwing structureswere fairlysimple affairs.All you neededwere a fewsand bags.

Static test of an upturnedDewoitine D27 wing.The D27 first flew on 3rd June 1928piloted by Marcel Doret.

Sixty years later life had become more complicated and expensive. Complete aircraft, specialhangars and test equipment. Fatigue tests as well as static tests had become the standard inEurope with the arrival of the jet transport age in 1949 when the de Havilland Comet first flew.

MANAGING YOUR AIRLINE? OUR TAILOR-MADE TRAINING WILL KEEP YOU AIRBORNE.Juggling schedules can be tricky. To keep disruption to a minimum, Airbus quality training is tailored to the

individual with our unique family concept, Cross Crew Qualification courses stop key members of staff being tied

up for long periods. And with continued support during peak times, we do everything to ensure your airline

remains fully operational. Airbus Customer Services. Dedicated to meet your requirements.

http://www.airbus.com

AIRBUSSETTING THE STANDARDS

FAST / NUMBER 2632

A310 static test at CEAT Toulouse in 1983

In the 1920sstatic tests ofwing structureswere fairlysimple affairs.All you neededwere a fewsand bags.

Static test of an upturnedDewoitine D27 wing.The D27 first flew on 3rd June 1928piloted by Marcel Doret.

Sixty years later life had become more complicated and expensive. Complete aircraft, specialhangars and test equipment. Fatigue tests as well as static tests had become the standard inEurope with the arrival of the jet transport age in 1949 when the de Havilland Comet first flew.