5-1

Required Technologies for Supersonic Transport Aircraft

Josef Met-tens Daimler-Benz Aerospace Airbus GmbH D-28183 Bremen Germany SUMMARY After referring to the remarkable technology level which Concorde has achieved, the most challenging new requirements for a future Supersonic Civil Transport are presented. It is proposed how to estimate influences of technology improvements on aircraft flight performance. A survey on key technologies follows with special emphasis on aerodynamic technologies. CONCORDE TECHNOLOGY LEVEL 1. Concorde (fig. 1) is the only supersonic airliner which was introduced into regular passenger service. It is still in service at British Airways and Air France without any flight accidents, and probably will stay in service at least for ten more years. than the total of all military aircraft all over the world. Concordes rangeis about6 500km, whereas the bestfighters like Su-27or F-22 -so called supercruisersachieveabout 200km in sustainedlow supersoniccruise, and military supersonic bombersor reconnaissence aircratt like B-l or SR-71reachabout3 500km without refuelling.But although supersonic flight rangeof Concordeis by far better than for any other supersonic jet built, this range was the most important limiter for a commercialsuccess of Concorde. A new viable supersonic airliner, called Supersonic CommercialTransport (SCT), must be able to serve the importanttrans-Pacific marketrequiringa rangeof 10 000 to 11000 km. This is a tremendous improvement compared to the Concorde. What arethe differences betweenConcorde and a new SCT? Besidesthe larger size, which improves a bit the range performance, technology improvements are cited to enable this big step forward. So, lets look at the technology improvementswe have achieved in aviation since the Concorde design.As thereis no other supersonic airliner, we have to compareConcordes contemporary subsonic airliners with the newestgeneration of subsonic airliners[11. Aircraft fright performance is governedby aerodynamics, structures and engines. All other disciplines-althoughoften important for the viability of an aircraft- are only weakly relatedto fright performance of transportaircraft. Therefore we will look at the improvements in thesemaindisciplines.
Concorde A340-300E

Figure

1: Concorde

Concordehas experiencedthe most supersonic flight hours and flight milesof all aircraft. Indeed,the only twelve flying Concordes have accumulatedmore supersonic flight hours
About the same development generation: Modern aircraft, 25 years later: B737-200, B737-500, B747-100, B747-400,

OWE % Concorde B737-200*) B737-500*) B747-100 B747400 A340-300E

OWE: operating

PAY % 4,8 18,4 16,l 10,5 9,61 9,82

QWJ PAY 8,90 3,06 3,23 4,67 4,79 4,88

PAYxMm 1,647 0,398 0,436 0,458 0,345 0,315

Improvment % -----9,55 --24,67 31,32 (rel. to B747-100)

: e

Range Mm 6,58 4,07 4,48 9,04 13,27 13,24

PAX No. 98 107 108 385 420 295

h4Tow Mg 185,l 52,6 60,6 332,l 394,6 271

42,52 56,14 52,Ol 48,90 45,99 47,90

empty weight,

PAY: payload, MTOW: maximum take-off weight

PAX: passengers

Part of improvement was used to increase range, part to increase payload Performance comparison by fuel per passenger-kilometer *I Data base for the different B737-versions seems to be inconsistent; because the more efficient JT-SD for B737-200) should improve aircraft efficiency, at least for long ranges engine of the B737-500 (CFM-56 instead of

Figure 2: Weight

improvements

since Concorde

Paper presented at the RTO AVT Course on Fluid Dynamics Research on Supersonic Aircraff, held in Rhode-Saint-GenZse, Belgium, 25-29 May 1998, and published in RTO EN-4.

5-2

1.1

Concorde structure years later

weight

still comparable

- 25

The first example of technologies concerns weight improvements, listedin figure 2. Boeing B737-200and B747-100were developedin parallel with Concorde [2]. The new generation aircraft of comparable sizeareB737-500,B747400 and Airbus A340-300E[3]; the latter beeinghandicaped whencompared with the B747400, because it is a bit smaller.Improvements of fuel per passenger-kilometers of more than 30% were achieved; but comparison to the old aircraft is difficult, because database haschanged (improvedseatingstandards etc.). When comparing Concordes structure weight, it is still comparable to the A340. (Comparison canonly be madeand was madeby designingboth aircraft with the samedesign tool. Assuming severalstructuretechnologystandards, results showedthe weights to reach the actual valuesof Concorde resp.A340, when usingthe same structurestandards for both designs[4]). But what are the technologies leading to the subsonic weight improvements or Concordesadvanced values? - Because Concordehasonly a small payloadfraction, it is much more sensitiv to weight increments; therefore more effort was spentfor weight savings andexpensivesolutions became useful. - Many (weight) improvementsfor transonic aircraft after B737-100, B747-100 were provided by interdisciplinary effectslike high bypass engines or optimized (nonlinear) transonicaerodynamics (fig. 3) via increased wing profile thicknessand volume or reducedwing sweep which cannotbe transfenedto a Concorde type contiguration.

time, we find strongdifferences in the aerodynamic efficiency betweenConcorde time subsonic aircraft like B747 or B737 and modernaircraft like B777, A340 or A320, expressed by improvedL/D (lift/drag). Thesesubsonic aircraft fly at high subsonic Mach numbers,when the air flow at the aircrafi locally reaches partially subsonic, partially supersonic Mach numbers, which is namedtransonic flow. Physics of transonic flow includestrongnonlinearities like shockwaves,and the governingequations changefrom elliptic to hyperbolictype. Whereas the old aircraft were designed using pure subsonic linear potential flow theory combinedwith simple sweep theory, improvements were provided exploiting nonlinear theory;the latter requiremodem high performance computers not availablefor Concorde development. But flow around supersonic airliners like the Concordeis dominatedby small disturbancesof the incoming flow, because strong disturbances would create high wave drag. Therefore, design (and especially sizing) of supersonic airliners can mostly be basedon linearized potential flow theory, aswasConcorde (slender body theory). Only for some parts nonlinear effects have to be respectedfor: strong interferenceeffectslike engineintegration,tine tuning of the configurationand strongly nonlinear boundary layer flows. But, except for laminarflow, other strongly nonlinearboundary layer flows like separation are avoidedbecause they are connected with largedrag increases. Therefore,modemnonlinear aerodynamic theory can only provide limited improvementscompared to Concorde (except for laminarflow). Although, some aerodynamic improvements may be provided: - New materials providing higher specific stiffness may allow a higheraspect ratio via interdisciplinary effects. - Local optimization for nonlinear flow phenomena will provide reduced interference drag, especially for engine integration. - Nonlinear calculations will speed up configuration optimisation. - Supersoniclaminar flow is the only new aerodynamic technology which can stronglyimproveperformance. But it is still far away from realisationfor largetransportaircraft; and laminardesignmustnot too muchpenalizewave drag design.

6I

transonic turbulent profil

conventional turbulent profil 1.3

Reduced Wing Sweep Increased Wing Thickness and Volume

Olympus

engine efficiency is still very good

Figure 3: Improvements

in Transonic

Aerodynamics

When comparing engineswe have to compare installed engine effickneies (qE), although mostly sphfic fuel consumption (SFC) is used. SFCdivides the fuel flow (i.e. an energy flow) by the thrust sophysicallyit is not a meaningfulvalue andcan only be compared at the samespeed (v = flight velocity). In contrastflE is the amountof energvflow providedby the engines thrust divided by the energVfIow provided by the fuel (H = calorific value of fuel). qE and SFC are connectedvia the equation q~ xH = v/SFC (1)

- System weight was only marginally reduced since Concorde. - Since Concordes time many improvements in structure technologydid not result in weight improvements, but were offset by advancedsafety requirements; e.g. new requirements for pressure losses, cabin evacuation or fire resistance of cabin equipment. But this is the world where a new SCTmustfly.

firce;

1.2

Limited aerodynamic Concorde-like aircraft

improvement

potential

for a

When comparing aerodynamic improvements sinceConcorde

When not using a consistentsystemof units like SI, the respective unit conversions have to be applied.For Kerosen

5-3 H is given by H = 42.817MJikg (2) body aircraft it has a size of 20 sqft (FAR 25). This poses extreme difficulties for flight altitudes above 40 000 ft. It is impossible to tight against this requirement by building a stronger aircraft, because there is no requirement on strength or probability of creation of such a hole, but only a given hole size. Concorde needs only to survive a hole in the hull of window size; and Concorde has small windows.
Environment:

Now we can compare engines. At Mach 2.0-cruise, Olympus [2] and a proposed new supersonic engine are given as well as a modem high bypass transonic engine used for widebody aircratt [ 31: Olympus, M=2.0: New engine, M=2.0: SFC = 1.19 kg/daN/h SFC = I,13 kg&N/h r,s = 0,41 (3) rls = 0,43 (4) ns = 0,37 (5)

CF6-SO-C2, M=O,85: SFC = 0,56 kg/daN/h

We see that Olympus is still very good. The improvements of subsonic engines, mainly achieved by strongly increased bypass ratio, did not yet reach Olympus efficiency. Indeed, at supersonic cruise a modem optimized engine with very low bypass ratio would provide slightly better values than the one given above. But this engine -like Olympus- will never meet the stringent noise criteria at take-off and landing which new SCT have to fulfill. Probably a bypass ratio of about 2 and extensive noise suppression (damping plus ejector) will be applied to meet noise criteria; this will decrease engine efficiency at supersonic cruise to values indicated above. 2. REQUIREMENTS FOR A NEW SUPERSONIC COMMERCIAL TRANSPORT (SCT) There are only a few Concordes which serve a small, exclusive class of passengers over limited distances at high fares, For its rare presence Concorde is allowed to meet only elementary environmental criteria which will not become more stringent for Concorde itself. Especially for noise, ICAO Annex 16 and FAR 36 explicitly require for Concorde, that it must not become even more noisy than it was just at certitication; and Concorde is very noisy. For new supersonic airliners those rules will never be applied; instead, a new SCT must fulfill new requirements which are not met by Concorde: It must comply with all valid certification rules, it must be economically viable, it must provide sufficient comfort. A new SCT must be just another aircraft [S]. In the following paragraphs only those points are mentioned which will introduce significant new challenges compared to Concorde. 2.1 Relevant certification rules Safety: Concorde has proven to be save. But since Concorde certification some new rules were introduced by the authorities -e.g. FAA, JAA- which a new SCT has to meet [6]. The most obvious challenge is cabin pressure loss: It is required, that aircraft and passengers can survive pressure loss in the cabin provoked by a sudden hole like a broken window. Therefore, sufficient pressure levels must be maintained in the aircraft when a hole opens. After DClO-accidents, when burst of underfloor cargo doors destroyed the floor of the aircraft with the hydraulic systems, size of the relevant hole was increased by pure geometrical definition (about the size of a door); for a wide

- Concordeproduces inacceptablenoise at take-off and landing,althoughhaving lessthan 110 passengers. A new SCT,which will be of at leastdoublethe size of Concorde, hasto meet current noisestandards (FAR 36, stage3) or evenfuture morestringentones.For comparison: FAR 36, stage 3 is just met by the B747-200.But SCT engines will not have a bypassratio of about 5 like the B747-200 engines [2], but only oneof about2. - Best aerodynamic performance L/D (lift/drag) is reached at elevated values of the lift coefficient (about C~=0.5 for subsonic aircratt, below C~=0.15for supersonic flight). To fly at good aerodynamic performance, to maintain acceptable pressure levels in the inlets and engines, andto till the engineswith air, dynamic pressure at supersonic cruisewill be in the rangeof about20 to 30 kPa. Therefore, flight altitude depends on cruise speedand weight, for a Mach 2 aircraft about 16 km. And the higher the cruise altitude, the moresensitiveis the atmosphere to pollution, especially the ozonelayer. At the time beeingit seems, that a Mach 2-SCT will not harm the ozone layer; but this is based on calculations which still are questionable. Supersonicaircraft will burn more fuel per passenger kilometer than subsonic transports.Although CO2 is not altitude sensitive,it is a well known greenhouse gas,and the largeamount of CO?emittedhasto bejustified. In the future it hasto be expectedthat the public becomes evenmore sensitiveto environmental impacts.Thereforea new SCT hasto demonstrate, that its impacton the earth atmosphere is tolerable. n A body cruising at supersonic speedgenerates a sonic boomwhich follows the body. This boomis an annoying andstartlingnoisein anareaof about30 to 40 km at both sidesof the SCTs track (fig. 4). To avoid harassment or

pr.rruro

hlrtoty

on (Iround

(Concorde)

Figure 4: Sonic Boom damage,civil supersonic flight will only be permitted over seaor perhaps uninhabitedland. Because noiseof the naturalenvironmentis sohigh on the sea,there is no

5-4

harassment or damage known to people, animals or ships below Concorde routes. In contrast to people on land, there exists no complaint about Concordes sonic boom over sea [7]. Operations: - Any new aircraft must be able to follow current and future ATC (air traffic control) procedures. In the future, steeper descent angles may be requested in the airport area which could be a challenge for SCT. - New SCT must meet current ground load values, i.e. wheel number, size, loading and distribution. High ground load values may damage some airports, especially on aprons with tunnels. - Loads on passenger and crew during operation must stay within acceptable boundaries. Especially long elastic fuselages provide strong vibrations during ground roll and in turbulence, and may impose high g-loads during rotation.

earn their money with this aircraft. This duction of a sufficient number of aircraft price for the manufacturer, and operationg lower than the money paid by the passengers

requires proat a profitable cost which are [lo, 1 I]:

2.2 Economic viability Range: Concorde is only able to serve some of the medium range overwater routes. This is an inattractively limited part of the market. To gain a sufficient part of the market, future SCT must serve the important overwater long range routes. This, for the trans-Pacific routes, is more than 5 500 nm range. Range is even more important than speed ! For its limited range Concorde did not find a market, although Concordes range was at the achievable limits. Also, the American SST, the Boeing B2707, failed for its insufficient range [S] and its uncontrolled growth of weight; even when environmental concerns were cited for stopping the program [9] (which luckily limited the liability of Boeing). Operations: Today, there exists a very expensive environment for air traffic operations: airports with its buildings and installations, hangars, air traffic control (ATC). New aircraft may require some additional installations, but the necessary investments must be paid by the money earned with those aircrafi. And at least for the introduction of the new aircraft, success depends on the ability to cope with the existing environment. Time spent at the airport for deand emparking of passengers, servicing of the aircraft, refueling etc. must be minimized: - Aircraft dimensions (length, span or occupied ground area) must be compatible with existing installations at the relevant airports. - Aircraft accessability (doors) must be compatible with existing airport installations; service ports should be compatible with usual procedures. - Aircraft accessability must allow for parallel de-/embark, service, fueling, etc. to enable rapid turn-around. - Aircraft supply needs must be compatible with existing (big) systems, e.g. fuel type. - To serve the many routes with overland legs, subsonic cruise performance must be about as good as the supersonic one. Concorde has only poor subsonic performance. Cost andfares: A new SCT will only become a success, if the manufacturer of the aircrafi and the airlines operating this aircraft will

- To reduce seat mile cost and to serve a large market, an SCT must transport many passengers (size effect), at least as much as subsonic long range aircratt (e.g. A340). (Smaller SCT would -at first- only serve full fare passengers; but after a short period airlines will begin to introduce reduced fares to till empty seats -because empty seats are the most expensive seats-, and this leads to the situation today with mostly low fare tickets.) - To gain a sufficient part of the market, a future SCT must serve all classes. - Speed pays (see cars, trains, air traffic) but a surcharge must stay in acceptable limits. It seems, that the travelling public accepts surcharges of about the money they would earn for the time they saved by the increased speed -even for holiday trips. - An SCT with those (low) fares must still allow a profit for manufacturer and airlines. And the manufacturer must sell many aircratt in order to provide an acceptable price. This becomes impossible with pure exclusivity.

2.3 Comfort Passengers like comfort, and especially on long trips some level of comfort is necessary. But what is the special kind of comfort required for supersonic transports ? Range: Passengers select a faster transport to save time; and they pay for the time saved. But this is the time from airport to airport or even house to house. Therefore supersonic transport makes only sense, when the flight time is an essential part of the trip time, i.e. for long ranges. And passengers do not like stop-overs; because they prolong flight time and are even more annoying than flight. Therefore an SCT must provide long range capability, to connect at least the most important areas of the world: Europe, US and eastern Asia, all separated by up to 5 000 to 6 000 run. Speed: Passengers pay for the time saved. Therefore overall trip time has to be reduced, as well by high cruise speed without stop-overs, but as importantly by accelerated checkin / check-out. Additionally, many passengers feel very uncomfortable during and after trips of more than 6 hours flight duration, especially children, or disabled and elderly people, who will feel circulatory trouble. Space: Passengers want space like in comparable subsonic transports, i.e. medium range transports with same flight duration. The narrow Concorde fuselage is only accepted for Concordes exclusivity.

3. ESTIMATION OF TECHNOLOGY INFLUENCE There are several technologies to improve a new SCT. In order to evaluate the importance of a technology, we must be

able to estimate its influence on the realisation of an SCT. At present level of knowledge, there is a hierarchy of technologies from fulfillment of constraints over cruise performance to operations. The operations are still at the last position, because we are on search for a technical solution. When a solution is found, operations will become more important, because they contribute strongly to cost performance.

weightednoise.Additionally, if noiseat onepoint is only a little moreelevated(at most2 dB), it may be compensated by lower noiseat the other points, following some complicated weighting.
- Transonic acceleration: The highest wave drag values and the worst engine efficiency nearly coincide at low supersonic speed;this

3.1

Technologies

to fulfill constraints

Here technology estimation is possible looking for the relevantphysicalpriciples.Influenceof constraints on aircraft cruiseperformance andcostis indirect, but canbe very strong andlimiting. The mostimportantconstraints for an SCTare:
- Take-08

may determine enginesize. Best supersonic cruiseperformance doesnot make any sense,if the supersonicaircraft is not able to reach supersonic speed.
- Range:

Field length: The required thrusf (i.e. engine size) is determinedby weight (acceleration)and aeroajmamic fijii (span, wing area,rotationangle)at lift-off (i.e. minimum airspeed). Take-off field lengthis definedby runway lengthof the relevant airports to be used by SCT. Today about 11 000 ft areassumed. Climb rate: The required thrust is determined by weight and aerodynamic performance UD. After take-off, a climb rate with oneengineout mustbe maintained of 0,5% with gearextended and 3% with gearretracted.
- Noise:

Noise is determinedmainly by exhaust velocity of the jet engines (aslong asthe turbomachinery is well shielded by inlet andnozzle) and mass flow (i.e. thrus< take-off thrust = massflow . exhaustvelocity). This means that an aircraft of sameweight shouldhave about the samenozzle areato producecomparable noise;e.g. SCTandB747. Stage3 compliantlow noiseexhaustvelocitiesare between 300m/s (Airbus A340, well below stage3 limits) and at most400 m/s [12]; for compliance, bypass ratiosof about2 or comparable measures arerequired. Required low speed t/~rus~ is determined by field acceleration, by weight and drag during climb, and by drag during approach. In the rulesICAO, annex 16, chapter3 andits derivations FAR 36, stage3 or JAR etc., maximumnoiseallowedat take-off and landingis defined.Figure 5 shows the three points, wherenoiseis measured. Maximum noiselevels

Fuel amounts to a very high portion of gross weight (50% or evenmore).Thereforecruiseeflciency determines range capability. Airlines indicate, that a viable supersonic transportmust be able to fly at least about 5 500nm. Many routes include segmentsover inhabited land, where only subsonic speeds are allowed. Supersonic cruise: Reduceaerodynamic drag (hereby slenderness), reduce weight andimprovesupersonic engine eflciency (herevia decreased bypass ratio) to meetrangerequirements. Drag at supersonic cruise is dominatedby wave drag, but friction drag and induced drag are important as well. Drag hasto be balanced by thrust. Subsonic cruise: Improve subsonic engine eflciency (here via increased bypass ratio) and reduceaeroa)namic drag (heremainly largerspan); both reducesupersonic efficiency. Drag at subsonic cruiseis dominated by by friction drag and induced drag, possibly vortex drag due to separation. Drag hasto be balanced by thrust.
- Controllability: Provide control author@ for all disturbanceseither external (gusts,manoeuvers) or internal (failure cases like inoperative engines, cabinpressure loss).

- Emissions:

It seems, that most importantwill be NOx-generation for its influenceon the ozonelayer. NOx-generation is determined by peak temperatures which occur only at spot points in the burner, where stochiometric conditions are met. Low NOx-burners reducepeak temperatures, but still maintainhigh mean temperatures. Some of those constraininginfluenceswork against each other. For estimationof valuescomparesubsonic aircrafi of similarweight, e.g. B747, but pay attentionto the significant physicalparameters.

approach noise measurement point

3.2
measurement point at highest

Technologies

to improve

cruise performance

noise

Figure 5: Noise measurement points allowedfor the different pointsdependon aircraft weight and number of engines.Noise levels are measured in EPNdB which is a time integral of the EPNL(dB)

To estimate influences on cruiseperformance, at first we will estimate the mostimportantphysicsof constantcruise.This is -for a real aircraft- some kind of oversimplification, becauseit ommitsthe importantsegments of take-off, climb, climb during cruise, accelerationand reservesfor go-aroundand divert as well as the minor parts for descend, landing and

5-6

taxy. But for a long range aircraft it remains the most importantsegment. Constantcruise is governedby the two equilibriumsof lift (L) andweight(Wi andof thrust (T) anddrag(D): L=W T=D Thrust is payedby fuel (F) T . SFC= dF/dt
(SFC: specific fuel consumption, see eq. 1; t: time)

aerodynamic improvementDL/D for the sameaircraft (ns, m,,,F). Now equation (13) gives

(6)
(7)

For fixed range,we cantry to estimate the influences on fuel needed (F) by the inverseBreguetequation (16) But caution: usually now both mo and F becomestrongly dependent on AL/D or Anr; because better drag or engine efftciency reduces fuel consumption F, and this reduces gross weight m. Therefore the inverse Breguet equationusually calculates an aircraft whoseperformance is not exploited.To exploit performance,the aircraft has to be resized. This cannotbe calculatedby a simpleequation.As a first, very rough estimation, the unresized improvement for fuel consumption,calculated by the inverse Breguet equation (16), hasto be multipliedby the factor: 1+ (weight part underconsideration) / m,,
(17)

(8)

whichreduces mass (m) or weight dF=-dm, W=m*g

(9)

but providesrange(r) via speed (v) dr=v.dt

(10)

When introducingthe other equations in (8), simplealgebraic operations provide the differential Breguetequation dr = - (v-L/D) / (SFCag)* dm/m (11)

e.g. for improvementof ns: (1 + F/m). But remember: Breguet equation only calculates idealised cruise flight. Thereforeit may only be usedfor estimation of tendencies in longrangeflight. Small improvementsin fuel consumptionby individual technologiesmay be compared by: percentage of the individual improvement, multipliedby weightpart influencedby this technology/ payload (18)

or with engineefficiency nE and calorific value of fuel H instead of specificfuel consumption SFC(eq. 1)

After integrationwe haveBreguets rangeformula r=L/D.rlEWg.ln[m,/(nb-F)] or r = (v-L/D) / (SFC*g)* In [m 1(W-F)]

(14) (13)

This is explainedin fig. 6 by severalexamples for anAirbus Exmnpk ~sumecl drmg reduction of15 9L by new Cchnobgln:
A34C-3008 (12 000 lm -mission): pybadfmctbn: ~a. 10% weight fraction Wluanced by drag: structure weight: ca. 28 X tieI 1) i @;li consumption ,v*ge d ipt$TFfl MTOW: muinum tab ,hw. allwryr mnNnt, ca. fuel 50 Y raved . I2 X remaining fuel (40% fuel + 10% engines) df v.viY

r is the Breguetrangewhich for a real aircraft is longerthan the aircrafts range,but it allowsfor comparison of aircraft of similarweigthsor designed for similarmissions. Aerodynamicsdetermineaerodynamic performance L/D and soinfluenceneeded fuel F, but alsoprovide requirements for structurallayout and systems. Fuel, structures,engines,equipmentand payload are total mass at beginof cruisem.. Enginescontribute to massand provide its efficiency ns which influences needed fuel F andmass for fuel storage. The selectedcruise speedv or cruise Mach M influences achievable IJD, engineefficiency na or SFC, needed fuel F, andstructureand equipment mass. The Breguetequationallowssimpleestimationof rangeand comparisonof range performance for different aircraft designs; but it is still difficult to estimateimprovements by singulartechnologies. The Breguetequationprovidesa rough estimation of rangeimprovements, when only one parameter is changed,e.g. range improvementDr provided by a pure

40

so

so

104

120

140

Figure 6: Improvement estimation A340-300Bwith an assumed 15%pure drag improvement or a 15% structure weight reduction; for all examples the maximum take-off weight (MTOW) was maintained by addingfuel or payload.

3.3

Technologies

to improve operations

Technologies for operationalimprovements usually are not directly connected to flight performance, but rather to operating costs. Therefore, simple estimation of their influence is difftcult. Most of these technologies are not specificallyrelatedto SCT.Someof themarementioned.

5-7 Materials
parts:

with improved

creep resistance for

hot engine

In supersonicenginestemperatures remain high during cruise,whereas for subsonic aircraft peak temperatures are reachedonly at take-off. Thereforehigh temperature creep resistance becomes more important for SCT. It will allow Active landing gear allows better damping of roll vibrations, higherengine(cruise)efficiency and/orreducemaintenance ground loads and supports take-oflrotation. costs. Justunderdevelopment for very largeaircraft like A3XX.
Special control systems to guaranty controllability:

DissimilarSCT-geometry requires different procedures. Elevated temperatures during cruiseintroducedissimilar loads and load cycles which alter maintenance, sometimes evenin favor for SCT(e.g. corrosion).

BecauseSCT-configurations strongly differ from subsonic or military aircraft, somedifferent control problemsmay occur;e.g. - inoperative engineswith high lever arms in supersonic flight; if this determines control surface loads(mainly rudder, but also aileron, elevator) special surfacesor procedures may provideimprovedsolutions - after a sudden pressure loss * specialdevicesor differently sizeddevicesto produce high drag may enablerapid descent and deceleration to reachsufficientatmospheric pressure levels * special devicesmayhelp to maintainsufficientpressure andtemperature in the cabin. If needed, appropriate systems mustbe developed. Neutral point and centerof pressure vary stronglywhengoing from take-off over transonicflight to supersonic cruise and back for landing. For optimum flight performanceduring supersonic cruise,only very smallcontrol flap deflections are allowed. Both requirea highly sophisticated fuel trim system.
Artt$cial nose. vision will avoid the heavy Concorde-like droop

Improved ATC-systems.

Strongly required are new ATC systems and procedures, especially for long range overwater guidanceand area navigation(direct flight, not restricted to airways linking ATC-control points), and automateddata links (FANS = Future Air Navigation System). They are not specific to SCTandarejust underdevelopment for subsonic aircraft.

4.

TECHNOLOGIES FOR A CONCORDE SOR: KIND AND RISKS

SUCCES

We have seen,that many technologies required for a new SCT are just developedor under development for subsonic aircraft. Other technology improvements cannot be transferred to an SCT. And Concordes technologylevel is still comparable. But what are the new technologies which will enablerealisationof a competitiveSCT with sufficient range andoperatingcosts ?

4.1

Aerodynamic

technologies

This coursemainly deals with aerodynamics; in chapter 5 aerodynamictechnologiesare prescribed in more detail. Therefore,hereonly the importantpointsarementioned.
Aerodynamic multi point design:

Artificial vision is just under developmentfor CAT llI landings andgroundroll of subsonic aircraft.
Manoeuver weight. and gust load alleviation may reduce wing

Just in use for several aircraft (LlOl l-500, A320, A330/A340) andwill be improvedfor future ones(A3XX).
Systems for productivity. reduced turn-around time improve
aircraft

For subsonic aircraft, in the first designstepusually a two point design is made for * high speed cruiseand * low speed take-off and landing. Often this is achievedvia a nearly pureonepoint design for high speed cruise, whereas low speedperformanceis achieved usingrather complicated flap systems. A supersonic aircrafl mustbe ableto cruiseeconomically at supersonicspeed over uninhabited areas and at high subsonic speed over inhabitedland. And supersonic cruise only becomes possible,if transonic accelerationcan be realized.So,we haveto designfor four points: * supersonic cruise, * high subsonic cruise, . low speed take-off andlanding and a transonic acceleration. Moreover, low speed performance cannotbe achieved using additionallarge flap surfaces like fowlers. Since an SCT hasa largewing surface,effect of (fowler) flaps is very limited.And lift slope(C,) decreases for smallaspect ratio. But lift via high angleof attack, as usedfor Concorde,is accompanied by large dragwhich requiresnoiseproducing thrust. Thereforea best combinationof aerodynamics (and other disciplines)design principles is required to meet the

Such systems are under development all the time, especially for expensive, very large aircraft like A3XX. Specialproblems for SCTare * a largewing root blocksaccessability to largepartsof the fuselage . a narrow fuselage retards boarding/deboarding and groundcabinservice(like cleaning) * the large fuel amountrequiresseveralpointsor longtime for fuelling.
Systems to reduce maintenance costs are under development for subsonic aircraft all the time. Specialities for SCT are:

* Supersonic inlet andnozzle are specificto SCT. * Accessability to the enginesis reducedby supersonic inlet and nozzle, and possiblyby installationjust at the wing. * In many areasof the SCT spaceis very limited (small thickness at movables, tail, nose).

5-8

diverging requirements. Most importantpart is selection of a suitedconfiguration,for which aerodynamics contribute stronglyin meetingthe four aerodynamic design points.
Nonlinear aerodynamics:

50% for a Mach 2.4 SCT. Evenwhenconsidering for the high unit price of an SCT,this seems to be very risky! But still someimprovements in other technologies -especiallyconfigurationselection-may reduce the weight improvement requirements given above.
New Materials:

Nonlinear aerodynamics enable minimisationof interferencelosses, e.g. engineandwing-fuselage integration.It is beeingdevelopedfor subsonic transports;but it becomes morechallengingfor supersonic transports by the combination of nonlineareffects with very weak obliqueshocks, stronglythree dimensional geometries and boundarylayers, and strong shocksaround the enginesand in the inlets, evenwith reflectedshocks.
Control surfaces:

Design of the control surfacesdependsstrongly on the configurationselected.Smart solutionsmay decideon the viability of a configuration or strongly easedesign and operation.
Aeroelastics:

For Mach numbers below 1.8, highesttemperatures occur in sunshineon the ground.For Mach numbers above 1.8, cruisetemperature becomes important,especially in respect to life time. Below Mach 2.0, for the airframeemphasis is still more on light weight than on temperature.Some materials envisaged arecarbonfibers (CFRP), metalmatrix composites (h4MC) for highly loaded parts,ceramics. For supersonic enginescreep resistant high temperature materials arerequired(peaktemperature during cruise).
Manufacture:

New approaches like aeroelastictailoring or aeroelastic control are beeingdevelopedfor large subsonic transports. For the slenderSCT-configuration, inclusionof aeroelastics in the early designstageis a prerequisite, at leastfor static aeroelastics (i.e. inclusion of the shapevariation due to aerodynamic loads,but still without vibrations).For many SCT-configurations,flutter itself is as important, but hitherto it can be checkedonly at a more matureddesign stage, whena moredetaileddatabasewasbuilt up. Aerodynamics,especiallynonstationary aerodynamics, and dynamic structurecalculations are much morechallenging for an SCT than for subsonic transports,at least for the flutter sensitivesymmetric(Concorde-like)configurations. Because flow and structuredimensionally hasto be treated completelythree dimensional, 2D- or quasi-2Dapproximationslike airfoil flow, sweeptheory or beamapproximationsare not possible. Aerodynamic damping decreases with speed. Therefore supersonic cruise becomes flutter sensitive.But the nonlinear transonic aerodynamics (here high subsonic)also decrease flutter damping-the so called transonicdip. So thereis a second high subsonic flutter casewhich mostly is even more flutter critical than supersonic cruise[13]. This is oneof the mostchallenging calculations in SCTdesign.
Supersonic laminarflow (SSLF):

New methods arerequiredto manufacture very light weight structures and elements. For new materials,methods must be developed to fabricateparts and to join them to assemblies. Conceptsfor inspection,repair and crashworthiness are required. All these haveto be qualifiedearly.
Structure design:

Materials and manufacturing principlesmust be integrated in the design process. For higherMach numbers, the design mustrespectfor thermaldilatation andstress. Optimisation methodswill reduce weight and integrate structuredesign with other disciplines needs.
Aeroelastics:

Provide data early in the design process,to direct the interdisciplinary design to an optimum solution which respects for aeroelastic deformationandwhich will have a saveflutter margin. Integrate stiffness(and thermal) designwith aerodynamic design (aeroelastic/ aerothennoelastic tailoring).
Certification:

Supersonic laminar flow (SSLF) may provide strong improvements during cruiseflight. But it is still far away from realisationand many questions are unresolved,even some physicalprinciplesarestill not yet understood [ 141. Today, we postulatea referenceSCT to be viable without SSLF. Because, when a new subsonic aircraft will benefit from laminar flow, also the SCT must be improved by SSLFin orderto maintainits competitiveness.

New materialsandmanufacture methods mustbe certified, before being consideredin design. Both require a long time, especially for fatigueandthermalfatigueproving. For Mach numbersabove 1.8, thermal fatigue properties must be demonstratedfor the materials, for structural concepts, and for the aircraft itself [6].

4.3

Engine

technologies

Engines must fulfill severalrequirements at different design points,like the aircrat?.

Efliciency:

4.2

Material

technologies

An SCT has only a small payload fraction; therefore it becomesvery weight sensitive. To make a Concorde-like SCT viable, the necessary reduction of structure weight (compared to availabletechnoiogy,e.g.A340) will be about 30% for a Mach 1.6SCT 40% for a Mach 2.0 SCT

SCT-engines mustbe very efficient at supersonic and high subsonic cruise,and must provide sufficient thrust at take-off, transonic(low supersonic) acceleration andcruise.
Emissions:

To protect the atmosphere, pollution hasto be minimized. Especiallylow NOx-burnersare developed.This is even more stringentthan for subsonic aircraft, because SCT fly

5-9 higher and burn at higher temperatures during cruise.

Noise:

with dissimilarinterfacesbetweenthe disciplinescompared to the classical approach. In both cases, a solution requires the toolbox of Multidisciplinary DesignOptimisation (MDO) [15, 16, 171. There are three objectivesof MDO, each of them of equal importance.
Harmonize multiple disciplines:

Engine noise during take-off and landing must be comparable to subsonic aircraft (ICAO annex 16, chapter3; FAR 36, stage3). This requireslarge nozzle exit areas, comparable to subsonic aircraft of the same weight. The multipointcapabilitiesof the engines are stronglyrelated to the configuration selected.E.g. thrust available during take-off and landing depends on enginebypassratio, noise reductionby suppression or simplythrottling down, possible integration of noise suppressors in the airframe etc. This influencesengineweight, thrust available and engineefticiency in the other designpoints. On the other hand, thrust required in the different designpoints is determined by the configuration, mainly aerodynamic performance andweight.

Organizecooperation anddatatransferbetweenall relevant disciplines. Although all over the world companiestalk about introducing MDO, total quality management (TQM), concurrentengineering (CE) etc., in real life there are strong objections of hierarchiesagainst any kind of cooperationbetweendepartments. Future will show, if marketcompetition will improvethe situation.
Cooperative design:

4.4

System technologies

Most systemsare comparablewith the ones of subsonic aircrafts. Emphasisis on low weight, small space and possiblyhigh temperatures. Specialsystems for SCTare:
Inlet and nozzle control

Supersonic engines work at subsonic speeds. The inlet must deceleratethe flow to subsonicspeed,the nozzle must adapt to free streampressure. Inlet control with several shocks and tuning with nozzle state is especially challenging. Concordes solutionis still state-of-the-art.
CG-control

Data to be respectedfor harmonisation and tools applied have to be carefully selectedand must be ordered from crudepredesign to detailedfinal design. Data transfer,from predesign to the more detaileddesignsteps and vice versa for iteration steps,must be organizedand lit into the data selected for interdisciplinary transfer. Combinethe relevant data into a designwhich must be evaluated by all participatingdisciplines. The design processmust be able to update the design exploiting the data correctionsoccuredin the cooperative evaluation.
Optimisation:

From subsonic to supersonic flight, the centerof pressure varies strongly. Control is provided by center of gravity (CG) control via fuel transfer. This systemis proved in Concorde andappliedin severalsubsonic jets aswell.
Control jlight of OEI (One Engine Inoperable) during supersonic

Use optimisation tools which are well adapted to the different design stages in order to exploit the best combinationof the available technologies. Especially for interdisciplinary development steps, numerical optimisation is recommended. is the key technology for a new supersonic transport. For subsonictransportsremarkableimprovements are also expected.
MD0

Suddeninoperability of one engine(OEI) or inlet surge/ inlet unstart during supersonic flight must be coveredby the flight control system.During this time no excessive loadsmust be provided to the airframeor the passengers. Possiblyspecial control deviceswill supportthis failure case.E.g. Concordeand SR-71 switch off the symmetric engine and restart both enginessymmetricallyto control this asymmetric failure case.
Scheduled systems

5.

AERODYNAMIC

TECHNOLOGIES

In orderto meetnoiserequirements at take-off and landing, it is envisaged to usean automated system for scheduling of flight path, flap settings and engine controls. This is envisaged for subsonic aircraft as well. But an SCT will depend moreon sucha system.

A well balanced contributionof newtechnologies in all major disciplinesis required for realisationof a new Supersonic Commercial Transport(SCT). One of them -asusualone of the most important for aircraft- is aerodynamics. Here, the required pure aerodynamic technologies are specified in moredetail, accordingto our presentknowledge.Increasing insight into the problems may change the balance of importanceof the individual technologies and may require somemorecontributions.We mustnever confineour knowledgeto the knowledge baseof an expert (or expert system) at a given time, but muststayopenfor new insights[181.

4.5

Multidisciplinary

Design Optimisation

(MDO) 5.1 Aerodynamic knowledge base

A classicalConcorde-likedesignhas to reach technology limits of all relevant disciplines. This requires perfect harmonisation of all aspects in the design. An unconventionalconfiguration can improve performance over the limits of the classicalconfiguration. But it requires a new kind of cooperation betweenthe individual disciplines,

Physicsof flow will be presented in anotherlecture. Here someunresolved or not completelyunderstood problems are presented which require some aerodynamic knowledge improvements.

5-10

5. I. I Unresolved problems in the supersonic flow regime Suction force:

geometries. A challengeremainsto improve the tools for supersonic CFD.

Physical drag contributions:

Compared to lower flight speeds, in supersonic flow the efficiency of leadingedgesuctionbecomes moreand more inefftcient. Basically this shouldbe correlatedto the Mach number components normal to the leadingedge, trailing edge,maximumthickness line etc. But following M. Mann and H. Carlson[191,it shouldbe correlatedto free stream Mach number.

Reasons for lossof suctionforce are: - Low pressuresgenerate suction forces, but reduce density. This diminishesefficiency of suction forces, especiallyin combinationwith shock losses. Both are compressibility effects which should(mainly) be related to (Le.) normalMach numbers, - Supersonictrailing edgesinhibit circulation efficiency and so reduce suction force recovery. For supersonic The following tasks summarize the challenges of supersonic leadingedges suctionforce is lost. - The flow field in the vicinity of the wing is governedby aerodynamic design: radiation processes. These processes are not exactly - Provide aerodynamicdata suited for interdisciplinary modeled by numericalcalculations: linearizedtheory does design optimisation. These data must not be at the not model effects of local Mach numbervariationsand achievableaccuracylimits, but must be reliable within therefore is unable to produce the correct radiation specifiedaccuracies for a wide rangeof configurations. directions (characteristics). Most nonlinear methods - Maximize aerodynamic performance(Lift/Drag = L/D) respectfor the local Mach numbers,but do not exactly for given geometrical constraints: model radiation; so numerical diffusion smears out . improve quality of aerodynamic tools to reflect flow radiationtransport. physics, - Supersonic wingsaremainly designed for minimum wave * balance wave drag, induced drag, friction drag (indrag. This leads to nearly conical flow situations.At cluding laminarisation concepts, riblets) for minimum higher Mach numberswith smaller Mach angles this overall drag. introduces strong pressure gradients in spanwise - Determine the limits of special flow phenomena, like direction, i.e. normal to the free stream direction, suctionforce. Boundary layer flow tends to follow local pressure gradients;so, boundarylayer air will accumulate in the low pressurevalleys on the wing and may modify the 5. I. 2 Low speed tasks: designedlow wave drag pressuredistributions. This At take-off, it is necessary to effect shouldstrongly dependon Reynoldsnumberand - improvelift might be strongerin low Reynolds wind tunnel teststhan - maintaincontrol in free flight. This effect is mainly relatedto free stream - improveL/D by limitation of drag; Mach number. whereas at landingit is requiredto Completeunderstanding of the limitations of linearized - improvelift solutions,CFD-solutionsand low Reynoldswind tunnel - increase drag, possibly by dragcontroldevices testsremains open. - guarantee handling qualities, especially when using partially separated flow Radiation in CFD soluhons: - and allow for go-arounds (sufficient thrust and/or rapid Linearized theory doesnot respectfor local Mach number dragreduction). variations.Usual CFD-methods do respectthese,but only marginally model radiationproperties.Numerical stability is achieved usually by addition of numerical viscosity. 5.1.3 Transonic tasks: Without proper modellingof radiation properties,though, At high subsonic cruiseoptimizeL/D: random contributionsare introducedinto the solution or avoid separation valid contributionsare neglected.Upwind schemes should minimizeinduceddrag. model radiation, but most upwind schemes are basically one-dimensional andcannotmodelradiationdirection, like For transonicacceleration: all the upwind schemes which only fulfill the eigenvalue - minimize wave drag which is dominated by interference sign, i.e. they approximatethe radiation direction by an effects accuracy of up to l 90. Only CFD methodswhich are - provide controlof the aircraft carefully basedon the methodof characteristics (MOC) - providecontrol of engineinlet andnozzle. provide good radiation properties, but these methods usually are not suitedfor universalCFD codes,especially for their rather inflexible handling of complicated

To improvethe aerodynamic design of an aircraft, it is very helpful1 to know the different contributors for physical drag: - Wave drag (radiated energy plus entropy already generated by shocks) - induced drag - friction drag - separation drag. For subsonic flow and linearized supersonic flow, methods exist largely basedon far field balances, e.g. [20], but for nonlinear supersonic flow the far field results are poor, because of inexact radiationmodels. Also, dueto the small leadingedge radii, surfaceintegration accuracy is more difficult to achieve; only friction drag can easily be extracted.

5-11

5.2

Flap effectiveness hinge lines

development the most important limitationsfor wind tunnel investigations are:

Reynolds number:

5.2.1 Supersonic

At control surfaces with supersonic hingelines, shocks occur at the hinge line. Shocksproduce pressurelosses and so reduceflap effectiveness. Additionally, the shock canprovoke boundarylayer separation bubbles(fig. 7); pressure in those bubblesis lower than behind the final shock.This reduces the flap force significantly.Because of the system of the three shocks behavingvery sensitiveto variationsin the incoming flow and fluctuations in the separationbubble, strong vibration loads can arise.

Reynolds numberin wind tunnel testingusually is an order of magnitudelower than in free flight, for supersonic testing often up to two orders of magnitude.For drag measurements the boundarylayer is tripped; i.e. transition stripsprovoke transition from laminarto turbulent flow at defined positions. This allows for calibrated friction measurements, but the boundarylayer is thicker at lower Reynolds numbers. Therefore the interference effects, especiallyshock-boundary layer interference,in the wind tunnel remaindifferent compared to free flight. Technologiesbetter to transpose interferenceprediction from wind tunnelto free flight needto be developed.
Transition control:

Figure 7: Shock-boundary

layer interference

5.2.2 Trailing

edge jlaps

on

highly swept wings

On wingswith highly swept(subsonic) trailing edges (OFW, arrowwings),the boundarylayer is deflectedby the spanwise pressure gradientsandtendsto become nearly parallelto the trailing edge, or even separates (fig. 8). Tendencies, known from lower sweepangles,and results for very high sweep angles,are not conclusive.Further theoreticaland experimental investigations arerequiredto understand flap efficiency at relevant sweetangles.

Transition strips must be as small as possible.Thick or wide transition strips generateto much strip drag and thicken the boundarylayer. On the other hand,if transition strips are too small,no transitionor even relaminarisation occurswhich doesnot allow useful drag measurements. Thereforethe controlof transitionin experiments is always required,i.e. to identify the laminarandturbulentboundary layer regionson the whole aircratt model. Most common transition control techniques -like the acenaphtene technique- require wind tunnel runs at constant flow conditions.This is impossible in wind tunnels of blowdown type; and many supersonic wind tunnels are blowdown tunnels. Other transition control technologiesare required here; possibly the techniques devoleped for cryogenic wind tunnels (like very sensitive infrared measurements) canbe adapted here,or special liquid cristal techniques with sufficiently smooth surfaces.
Testing around Mach 1:

At near sonic speeds the flow aroundthe model contains large supersonic flow domains.In the supersonic regions the wind tunnel model radiatesdisturbances to the wind tunnel wall which are (at least partially) reflected by the wall back onto the model. In contrast to free flight conditions,this reflection strongly changes pressures and flow properties at the model. Figure 8: Highly swept trailing edge At high subsonic speeds, the supersonic regionscan reach the wall and so generate a chokedsupersonic nozzle flow over the aircratt insteadof the opensupersonic bubbleover the free flying aircraft. This (partial) nozzle flow changes the whole flow field and doesnot further resemble to free flight conditions. Most transonic wind tunnels have slotted or perforated walls in order to minimize wall reflections. This minimisation,though, is only sufficient, if the supersonic bubbledoesnot reachthe wall or the importantreflections do not meet the model.This requirestest flow conditions avoiding the vicinity of Mach 1. New transonic wind tunnelsuseflexible walls (adaptivewalls), wherethe wall geometryis adapted during the test to follow a free stream path line. This allowsfor better adaptionof near sonictest conditions. But quality of adaptiondepends on the technical conceptof the adaptionmechanism; usually only a plane wall adaption is possible for two of the four walls

5.3

Wind tunnel measurements

5.3.1 Wind tunnels Wind tunnel experiments areessential - to get insight into still unknown flow physics like separation, turbulence, transition - to validatenumericalcalculations - to generatedata for complicatedconfigurations including interference effects - to check aerodynamicdesignsand to generatedata for improvements - to establishaerodynamic data for pre-flight validation of new aircratt. But wind tunnels have limitations as well. For SCT-

5-12

surrounding the test chamber. Although two-dimensional adaption in the most important direction is better than no adaption, three-dimensional adaption for three-dimensional models is not yet realized for relevant wind tunnels.
Engine simulation:

Field mapping measurements in the freeflowjeld:

Usually, engines in supersonic testsaremodeled by simple through-flow nozzles.But it is difficult to designspillagefree through-flow nozzles: nozzles often are choked and consumeenergy, whereasan engine adds energy. Wind tunnel simulationtherefore often is restricted to limited cases including spillage.Additionally nozzle basedrag is addeddue to internal naccelledrag. It canbe correctedby pressuremeasurements, but these will correct only the individual nozzle basedrag, and not the additionalinterference wave drag.
Laminar tests:

Several new techniquesallow for measurements in a selected plane of the flow field. Most common is Particle Image Velocity (PIV), see e.g. [24]: The flow field of interest is seeded with microscopicparticles, commonly dropletsof about 1 pm diameter.In the planeof interest, thosedropletsare photographed twice within a short time interval. The movementof the droplets is identified to provide the dropletsspeed which is equalto flow velocity except at a shock.New developments are aimedat larger measurement fields andmeasuring all three velocity camp onents.In the future, PIV measurements will be usable even for complicated interference flow measurements. Several other flow field measurement techniques are developed, but either still in its infancy, restrictedto high Mach numbers, suitableonly for very specificcases or just of pooraccuracy.
Laser-Doppler anemometry:

There exists only one wind tunnel for supersonic laminar flow testsbetweenMach 1.5 and2.5. This is a refurbished facility at the University of Stuttgart, Germany.For more information, see e.g. [14]. More supersonic quiet wind tunnels in the Mach numberrange of interest are under construction or evaluation.

5.3.2 Measurement

techniques

Measurement techniques have beendeveloped for exploiting wind tunnel experiments. Someof thesetechniques are state of the art and provided by all wind tunnels: force measurement and pressuremeasurement via small holes in the model surface.The simpler optical methodslike shadowgraphs,Schlierenor interferograms are best suited for 2Dmeasurements and availablewheresuited.More refinedtechniques are availableandwill be appliedto supersonic testing, especially opticalmethods for flow field measurements:
Pressure sensitive paint (PSP):

In the last years Laser-Doppler Anemometry (LDA) was developedas a tool for accurate,pointwise, nonintmsive flow field measurements [25]. The flow is seeded with droplets(like for PIV) which are observedwithin a small measurement volume. This volume is established by the crossing of two laserbeams, where interference produces a sequence of light and shadowlike a grid. Observedis the motionof the dropletsthroughthe interferencegrid, where the frequencyof reflected light spotsis correlatedto the dropletsspeed.LDA allows for accurate measuringof meanand fluctuating values,even resolutionof boundary layer flow, but it requires relatively long measurement time. Complementary to
correction

the

experimental measurements,

methods for experimentalerrors or insufficient

Specialpaintsare developed which, whenilluminatedby a suitedlight source,emit light depending on the amountof 02-molecules embedded by the paint surface.In air the amountof 02-molecules directly correlates to air density. This allows for direct measurementof air density distributionon the modelsurfaceand, whentemperature is known, indirectly for the measurement of pressure distribution, see.e.g. [21,22, 231.This techniqueis new andneeds further improvements before it can be appliedas a stand alone pressuremeasurement technique. Especially paint toxity, thickness or durability, painting, illumination technique and related automatic data processingneed fiuther research.
Liquid cristal coatings:

simulationare needed.Here the most importantcorrections requiredare: - accurate correctionof wall interference - correction of Reynolds effects, especially for interferencies - spillage andnozzle base dragcorrection.

5.4

Aeroelastics

5.4.1 Static aeroelastics

Surface coatingsbasedon liquid cristal technologyallow for variousmapping techniques of relevantflow parameters on the models surfacelike shearstress, temperature. This allows simultaneous measurement on large parts of the model. Problemsresult from the mostly relatively rough coating surface, the limited view anglesand often the multiple sensitivitiesof the coatingswhich require careful separationof the measured effects. These problemsstill allow only for limited useof the techniquein aerodynamic measurements.

In classical aerodynamic design,the aerodynamic wing shape is designed for onedesignpoint Mc (cruiseMach at a given weight and altitude). Knowledgebasedmarginsprovide the ability to copewith the off-design points: Some off-designpointscover the (cruise)flight regime(like MM~ = maximumoperatingMach, MCS= subsonic cruise Mach, other aircraft weightsfor begin of cruiseor end of cruise,altitude variations ...). Others concernexceptional pointswhich do not occurin normalcruise,but only e.g. for emergencies like Mo (dive Mach). Aerodynamicsassume the geometryto be rigid. Once the aerodynamic(flight) shapeis fixed and the aerodynamic loadsare known, structureloadsare determined, structureis designedand static aeroelasticdeformation is calculated. This deformationat the designpoint is taken into account when the shapeto be built (jig shape) is defined. The

5-13

procedure reestablishs the designed aerodynamic shape at the design point flight loads (I& design weight and altitude). But for any deviation of the design point, the aircrafl will have a different shape. This deviation becomes important if the wing is not very stiff and if the deviations from the design point a large. Both occur for SCTs with thin wings and multiple design conditions. To find the best compromise for an elastic wing flying at different design points, aeroelastic deformation must be considered in the aerodynamic design. For aerodynamics this can be a rather simple formulation, like a beam formulation for a slender arrow wing or an OFW, or a simple shell formulation for some kind of delta wing, including bending and torsion. The difficult problem is the simple estimation of structural values, because this requires simultaneous estimation of loads, mass distribution and structural thicknesses.

investigationsare, the more sophisticatedthe numerical methods are, the more detailed the geometry must be described.But for all levels the same requirementsfor geometry generation hold:
For geometry genemtion by a human design engineer:

Geometrymustbe described by a limited set of parameters. But those parameters mustbe meaningfulandwell ordered in orderto allow a humanto reachgeometric designgoals. Alternative ways are allowed, e.g. multiple parametersets or parameter setselections.
For automatic optimizer: geomety generation by a numerical

Geometry must be described by as few independent parameters as possible.Those parameters may have any level of abstraction.Not allowed are alternativesto the optimizer for selectionbetweendifferent, but equivalent parameter sets.
Any geometv generator tending to wiggles: must provide smooth shapes not

54.2

Flutter

Concorde has inacceptabletake-off noise levels. For a Concorde-type SCT lower noiselevels can only be achieved usinglarger enginediameters and larger wing span.But to maintainor even increase cruiseperformance, wing thickness mustto be reduced. Suchwingsbecome very flutter sensitive. Aerodynamicdampingis an indicator for flutter onset.It is the smaller, the higher the flight speedis. But at high subsonicspeeds,nonlinear transonic aerodynamics reduce aerodynamicdamping,the so called transonic dip (fig. 9) [131, A new SCThasthereforeto be investigated for flutter at transonicspeeds andat supersonic cruisespeeds.
aerodynamic damping

If wigglescannotbe avoided, smoothing procedures must be provided. For human applications, the smoothing procedurescan be applied off-line as the last step of geometrygeneration. For numericaloptimizerssmoothing, if not avoidable, must be included in the geometry generation.
Any geometry generator CAD-systems: must provide intet$aces to andfrom

When aerodynamics hasdeveloped a shape, this shape will be transferredto other companywork groupslike project, structure, aeroelastics, model designand fabrication. All aircraft related data transfer uses CAD-systems. The aerodynamic shape thereforehasto be transferredinto the CAD-system without intolerableaccuracylosses. On the other side,aerodynamics hasto useinput from other departments for geometryconstraintslike fairing size etc. Or the real modelgeometryhas to be checkedprior to a wind tunnel test. Or geometries generatedby a partner mustbe investigated. Or wind tunnel results-like pressure measurementshave to be appliedto a given geometryfor aerodynamic improvements. In all those cases it must be possible to transfer the CAD-geometxy into the aerodynamicgeometrygeneratoras an input geometry,e.g. to startan improvement calculation. Especially for application of numerical optimisation strategies, more progressin systematicshapedefinition is needed.Sometimes it is proposed to use the CAD-systems directly for geometry generation, but CAD-systemsare oriented towards structural design: these do not contain geometry definition tools suitable for aerodynamic optimisation.Powerful aerodynamic 3D-geometrygenerators areunderdevelopment aspreprocessor for CAD systems [28]

Figure 9: Transonic dip Because of flutter becoming very critical for the thin wingsof symmetric (Concorde-like) SCT-configurations,at least a roughapproximation of flutter tendencies mustbe includedin the first stepsof configurationoptimisation.Hitherto nobody knowshow to do it. Perhaps, artificial flutter damping canbe appliedin the future, if its certificationbecomes possible.

5.6
5.5 Geometry generation

Fast computer

codes for aircraft

design

Aerodynamic designis developmentof a suited shape.For SCT development, extensive application of numerical optimizersis required [26, 271. When using optimizers,the first very important stepis to describe the spaceof possible shapes by as few parameters as possible,but still without inacceptable restrictions.In the first stepof interdisciplinary optimisation, only global parameters are needed to described the basic aircratt geometry. The more refined the

In the first interdisciplinarydesignloop, the whole aircraft is investigated.To allow the optimizer an investigationof the wholeflight mission of a sufficientnumberof configurations, the individual calculations mustbe very fast anduseonly few variables.In more detailedinvestigations, not all disciplines are involved at the sametime, and perhapsnot all mission points. The aerodynamic code can thereforeuse more time

5-14

and variables to become more accurate. As a result, available turn around time, variables involved and accuracy achieved rise from step to step until the ultimate step of the flying aircraft. As long as design modifications by theoretical predictions are relevant, turn around times are needed which allow for many repeated design loops. This is, depending on the design step: one hour, one night, one weekend. Very fast codes are closed formulas for the interdisciplinary investigations. They only need some main geometry parameters as input for global estimation of the aircrafts performance to allow for configuration selection. Fast codes are all codes which allow for turn around times of one hour for pure aerodynamic calculations (with many individual code calls) or one night/weekend for interdisciplinary tasks. Fast codes relay on linearized theory with empirical corrections. They need more geometry parameters to allow for a first aerodynamic design optimisation including volume distribution and a first approximation of twist and camber, and they check the aerodynamic predictions of the interdisciplinary model. Both codes calculate (at different accuracy levels) the global aerodynamic coefficients for performance calculations and first flight mechanics estimations. They identify the physical drag contributors and provide a load estimation. As any code used for numerical optimisation, the codes must be robust. This means: The code should be able to calculate all problems which the optimizer may pose, even some strange parameter combinations, If the code breaks down, this must not stop the design process, but the code should deliver an inacceptably bad result which is the worse the heavier the code crash was. For instance, if negative pressures occur, the result can be a bad value proportional to the detected negative pressure value. This leads an optimizer to solutions, where the code does not crash. If the code is reliable, only those are interesting solutions. Such cases must be controlled by the design engineer ! Today, most research effort is devoted to highly sophisticated CFD-codes. These codes are needed and must be improved furthermore, but for a better and practical interdisciplinary aircraft optimisation, quality and applicability of the simple fast codesmustbe improved. Much moreresearcheffort is needed in this direction.

To allow efficient exploitation of CFD-codes, the codes must fulfill the following requirements: - They must represent the relevant physical properties.For SCT design, theseare: reliable radiation of disturbances (not fulfilled by most CID-codes), predictionof shocks and shock reflections, prediction of separations (this still requiresmuch more research on turbulence). - They must be able to use the exact geometrydefinitions includingsuitednumerical grids. - They mustprovide insight in flow physicsby visual&ion postprocessing of results. - They mustbe ableto predictaerodynamic loads. - They must provide reliable performance predictions(drag predictionis still difficult for mostCFD-codes). - They must be able to identify the different physical contributiorsof drag (still a researchtask, especiallyfor supersonic flow with strongradiationproperties). - They must provide reliable aerodynamicderivatives for flight mechanics calculations. - They mustsupportthe analysis of experiments. If thosecodes are only usedto checksome resultsof previous predictions, the old fashioned procedure of man hour consuming grid adaptionand numericaltine tuning may be applied. But as soonas the code is used for configuration optimisation, new requirements mustbe fulfilled: - A geometrygeneratorwith very few variable parameters must modelthe variationsof interestwhich the optimizer hasto investigate. - The grid generator mustautomatically provide a suitedgrid of high quality. - The codemustfast and automaticallyconvergeto a useful result. If the code breaksdown, a (bad) result must be providedwhich directsthe optimizerto usefulvariations. - The results produced by the optimizers parameter variationsmustreflect the variation of physicalresults.

5.8

Inverse design capabilities

5.7

Accurate

computer

codes

Accurate computer codes here are CFD-codesbased on solutionsof the Euler equations,sometimes with a coupled boundarylayer solution and solutionsof the Navier-Stokes equations. They areusedfor: - configuration optimisation, to check and improve the previous design steps based on simpler codes, and to includeresultsof wind tunneltests, - interference drag reduction, which is impossibleusing simplercodes, - inlet and nozzle designwith strong shock/boundary layer interactions.

Since the introduction of direct numerical optimisation, importance of inverse design methods has decreased. Sometimes inverse designis seenas a relict of old design techniques. But inversedesignremainsimportant.There are still cases, where the numericaleffort of direct optimisation is still inacceptable. Thoughthis will changein future, there remain other cases:Inverse design allows to construct solutions for comparisonwith incomplete or defective solutions. E.g. using only partial inputs or other than geometryinputs,a geometrycan be designed to be compared with the geometryused for a CFD-calculation.Research is neededhere, especially, if not only the classicalinverse pressure designmethods are to be used,but alsoother input alternatives[29, 151.
5.9 Special control devices

An SCT has a flight envelope strongly enlarged in comparison to subsonic transports.All new configurations, eitherConcorde-like aircraft with thin wingsor an OFW, may provide some configuration deficiencies unknown for subsonic transports.If the existing control devices cannot

5-15

handle specific situations, or if the handling of those situations heavily penalizes those devices, then special control devices may improve handling of these situations.
Examples:

SCT optimizedwithout ejectorflaps mustbe compared with a completelyindependent optimumdesignwhich is adapted to the exploitationof the ejectorflaps.

If oneenginestallsor an engineburst occursat supersonic speed (OEI = One Engine Inoperative), strong lateral moments androlling moments can establish. If handlingof them penalizesrudder (and/or aileron) sizing, a special spoilerdeflection on the other wing may compensate the occuringyaw, roll and pitching moments.Fine tuning is possible usingthe conventionalcontrols. The obliquewing has superioraerodynamic performance, especiallyat low speeds. This may inhibit an acceptable landingprocedure with steepdescent.Specialdevicescan produce the requesteddrag without inacceptableintroduction of pitching moments. Such devicesstrongly dependon the selectedconfiguration. Theyare only recommended, if they considerably reducesize, weight or complexity of the already existing system.It is possible, that suchdeviceseasedesignof the control system layout, but complicating the system may occuraswell.

5.11

Supersonic

laminar

flow (SSLF)

5.10

Ejector

flaps

A new SCT, especiallya Concorde-type SCT, hasdifficulties to fulfill take-off noise requirements.Any possibility to improve take-off performance and reduce noise must thereforebe investigated. The enginecompanies haveproposed severalenginetypesfor SCT.Thereare engines which provide somuchhigh pressure air at take-off, that they canonly apply full (thermal)powerif a large amountof bleed air is usedelsewhere. It is worth therefore to investigate ejector flaps (fig. lo), mainly to increase thrust. Problems to be investigated are:

Supersonic transportsare very drag sensitive.Technologyto reduce dragby applicationof laminarflow, therefore,will be important; it is a prerequisiteto achieve very long range capability, or to reduceaerodynamic heatingat higher cruise Mach numbers[30]. In earlier studiesit was assumed that SCTswould only become possible by applicationof laminar flow [31]. But today, we request an SCTto be viable without application of laminarflow in order to maintainits competitivenesswhen laminar flow becomes available for subsonic and supersonic transports. By reducingfuel burned, laminar flow dragreductionreduces sizeandweight of the aircrafi, or increases rangecapability-whereas otherwise sizeandweight would grow towards infinity. Transition mechanisms from laminarto turbulent state of the boundarylayer flow (ALT, CFI, TSI) function as for transonictransports,but at more severe conditions: higher sweep angles, cooled surfaces; highermodeinstabilities(HMI) must at leastbe taken into account, although they may not become important below Mach 3. Hitherto there is a worldwide lack of groundtest facilities to investigate TSI at the expected cruise Mach numbers between1.6 and2.4; in Stuttgart, Germanyonesuch facility -a Ludwiegtube- is in the validation phase.A quiet Ludwiegtunnel couldbe a favourablechoicefor Europe.But it will require a new approachin designingaircraft which includesimprovedtheoreticalpredictions,usageof classical wind tunnelsfor turbulentflow andflight testsfor validation. A moredetailedoverview is given in [141.

6. CONCLUSION A new SCT will only becomereality, if many technologies are improved or newly developed. Some of them are aerodynamic technologies, asmentioned above.But many of them require contributions by other disciplines or need interdisciplinary connection with others. Both pure aerodynamics and interdisciplinaryproblems provide enough opportunities for manynew intelligentcontributions.

REFERENCES

Figure 10: Ejector flaps - What is the efficiency of the complex tubing and ejector flap system ? - Does the additional installation weight of the complex tubing and flap systemoffset the improvement of take-off performance ? - How complexandreliablewill the system be ? - What is the noiseof suchan ejectorsystem? The F117 is an example providing low noise, but design goal was mainly radarandinfrared signature. - Will exploitation of the ejector systemfor lit? generation improve the design, when trim penalties and safety requirements are respected ? To estimatethe pros and consof sucha system,probablyan

J. Mertens:Sonof Concorde, a Technology Challenge New DesignConcepts for High Speed Air Transport,H. Sobieczky ed., CISM 366, Springer Wien NewYork, 1997,pp. 31-51 Janes - All the Worlds Aircraft, 1977178 Janes - All the Worlds AircraR, 1995/96 A. Van der Velden,D. vonReith: Multi-Disciplinary SCT Designat Deutsche Aerospace Airbus Proceedings of the 7th EuropeanAerospaceConference EAC94 The Supersonic Transport of Second Generation,Toulouse, 25-27October1994,paper3.61 C. Frantzen: Introduction to Regulatory Aspects of Supersonic Transports Proceedingsof the European Symposiumon Future Supersonic Hypersonic Transportation systems, Strasbourg, November6-8, 1989,paperII, 3.1 J. Mertens:Certification of Supersonic Civil Transports New DesignConcepts for High Speed Air Transport,H.

5-16

7 8

10

11

12

13

14

15

16

17

18

19

20

Sobieczky ed., CISM 366, Springer Wien NewYork, 1997, pp. 97-103 J. Mertens: Sonic Boom Overwater Issues and Past Test Data, Review and Recommendations DA-ReportDA-010-93 /EF-1971,Bremen 6.10.1993 M. Goldring: A Second Generation Supersonic Transport, the Lessons from Concorde Proceedings of the 7th European Aerospace Conference EAC94 The Supersonic Transport of Second Generation, Toulouse, 25-27 October 1994, paper 2.3 1 J.M. Swihart: Prospects for a Second Generation High Speed Civil Transport Proceedings of the 7th European Aerospace Conference EAC94 The Supersonic Transport of Second Generation, Toulouse, 25-27 October 1994, paper 2.33 A.R. Sebass: The Prospects for Commercial Supersonic Transport New Design Concepts for High Speed Air Transport, H. Sobieczky ed., CISM 366, Springer Wien NewYork, 1997, pp. l-12 A. Van der Velden: Aircraft Economy for Design Tradeoffs New Design Concepts for High Speed Air Transport, H. Sobieczky ed., CISM 366, Springer Wien NewYork, 1997, pp. 13-30 U. Michel: How to Satisfy the Takeoff Noise Requirements for a Supersonic Transport AIAA-paper ALAA-87-2726, AIAA 11th Aeroacoustics Conference, Oct. 19-2 1, 1987, Palo Alto, CA, USA R. Barreau, T. Renard (Reporters): BRITE EURAM Program Supersonic Flow Phenomena, Final Report Subtask 1.3 Preliminary Aeroelastic Investigation of Supersonic Transport Aircraft Configuration J, Mertens: Laminar Flow for Supersonic Transports New Design Concepts for High Speed Air Transport, H. Sobieczky ed., CISM 366, Springer Wien NewYork, 1997, pp. 275-290 G.S. Dulikravich: Multidisciplinary Inverse Design and Optimization (MIDO) New Design Concepts for High Speed Air Transport, H. Sobieczky ed., CISM 366, Springer Wien NewYork, 1997, pp. 223-236 A. Van der Velden: Multi-Disciplinary Supersonic Transport Design New Design Concepts for High Speed Air Transport, H. Sobieczky ed., CISM 366, Springer Wien NewYork, 1997, pp. 251-273 A. Van der Velden: Tools for Applied Engineering Optimization VKI lecture series in Optimum Design Methods in Aerodynamics, AGARD R 803, April 1994 J. Mertens: Required Aerodynamic Technologies New Design Concepts for High Speed Air Transport, H. Sobieczky ed., CISM 366, Springer Wien NewYork, 1997, pp. 69-96 M.J. Mann, H.W. Carlson: Aerodynamic Design of Supersonic Cruise Wings with a Calibrated Linearized Theory Journal ofAircraft, 31, 1, Jan.-Feb. 1994, pp. 3540 A. Van der Velden: Aerodynamic Design and Synthesis of the Oblique Flying Wing Supersonic Transport PhD-thesis Stanford University, Dept. Aero Astro SUDDAR 621, Univ. Microfilms no. DA9234183, June 1992

21 R.H. Engler, K. Hartmann, B. Schulze: Aerodynamic Assessment of an Optical Pressure Measurement System (OPMS) by Comparison with Conventional Pressure Measurements in a High Speed Wind Tunnel Paper, presented at ICIASF 91, Washington D.D., 8 pages 22 A. Vollan, L. Alati: A new Optical Pressure Measurement System Paper, presented at ICIASF 91, Washington D.D., 7 pages 23 B.G. McLachlan, J.H. Bell, H. Park, R.A. Kennelly, J.A. Schreiner, S.C. Smith, J.M. Strong: Pressure-Sensitive Paint Measurements on a Supersonic High-Sweep Oblique Wing Model Journal of Aircraft, 32,2, March-April 1995,217-227 24 C.E. Willert: A Comparison of Several Particle Image Velocimetry Systems DGLR-Bericht 94-04 Stromungen mit Ablosung, Erlangen, 4.-7.10.1994,266-271 25 H. Lienhart, T. Bohnert: Grenzschichtmessungen an einem Laminarfltigelprotil mit einem Laser-DopplerAnemometer DGLR-Bericht 92-07 Stromungen mit Ablosung, KolnPot-z, lo.-12.11.1992,471476 26 A. Van der Velden: Aerodynamic Shape Optimization VKI lecture series in Optimum Design Methods in Aerodynamics, AGARD R 803, April 1994 27 A Van der Velden: Supersonic Aircraft Shape Optimization New Design Concepts for High Speed Air Transport, H. Sobieczky ed., CISM 366, Springer Wien NewYork, 1997, pp. 237-250 28 H. Sobieczky: Geomew Generator for CFD and Applied Aerodynamics New Design Concepts for High Speed Air Transport, H. Sobieczky ed., CISM 366, Springer Wien NewYork, 1997, pp. 137-157 29 G.S. Dulikravich: Combined Optimization and Inverse Design of 3-D Aerodynamic Shapes New Design Concepts for High Speed Air Transport, H. Sobieczky ed., CISM 366, Springer Wien NewYork, 1997, pp. 189-200 30 Boeing Commercial Airplane Company: Application of Laminar Flow Control to Supersonic Transport Configurations. NASA Contract Report 181917, July 1990 3 1 J. Mertens: Laminar Flow for Supersonic Transports in J. Szodruch (ed.): Proceedings of the First European Forum on Laminar Flow Technology, Hamburg, 16.18.3.1992, DGLR-Report 92-06