Copyright 2004 by DLR-SART. Published by the American Institute of Aeronautics and Astronautics, Inc. with permission. Released to IAF/IAA/AIAA to publish in all forms.

1

IAC-04-V.8.07

Search for Technically Viable Options to Improve RLV by Variable Wings

Martin Sippel, Josef Klevanski, Uta Atanassov Space Launcher Systems Analysis (SART), DLR, Cologne, Germany

A systematic study screening the possible variable wing options for a reusable booster or first stage with a separation Mach number of about 6 is described. This investigation is motivated by a list of severe obstacles of vertically launched winged vehicles which should be reduced or avoided. Based on a generic vehicle with circular fuselage the propellant combinations LOX/LH2 and LOX/RP1 are analyzed for low- and shoulder wing positions, and high aspect ratio vs. delta-type shape. Variability options are checked for swing-wing, rotational, foldable lay-out. To reduce the technical risk and complexity, it will be assumed that only two wing positions exist: deployed and undeployed. During ascent the aerodynamic lifting device is in its stored configuration, and the deployment is subsequently realized during the stage's ballistic phase outside the atmosphere and prior to reentry. In the study's first step the tanks are sized, the overall vehicle is configured, and wings and stabilizers are dimensioned and arranged. Calculating mass, center of gravity and aerodynamic data sets, enables to perform a flight dynamic assessment for ascent and descent. Once this preliminary work is completed non feasible options can be scrapped. In a second more detailed loop, the most promising combinations are investigated further. This includes dynamic simulations of atmospheric re-entry, fly-back, and improved aerodynamic analysis with local heat flow assessment. The assessment addresses questions of technical sanity, preliminary sizing and performance issues. Finally, the most promising configuration is integrated into a long-term, strategic scenario of reusable stages as part of a broader scenario for options in future European launcher architecture

Nomenclature

D Drag N L Lift N M Mach-number - T Thrust N W weight N m mass kg q dynamic pressure Pa v velocity m/s α angle of attack - γ flight path angle - δ deflection angle - η control surface defection angle -

Subscripts, Abbreviations

AOA Angle of Attack CAE Computer Aided Engineering GLOW Gross Lift-Off Mass GTO Geostationary Transfer Orbit LEO Low Earth Orbit LFBB Liquid Fly-Back Booster LH2 Liquid Hydrogen LOX Liquid Oxygen MECO Main Engine Cut Off NPSP Net Positive Suction Pressure RFS Reusable First Stage SHLL Super Heavy Lift Launcher SSO Solar Synchronous Orbit TSTO Two Stage to Orbit TVC Thrust Vector Control cog center of gravity sep separation

1 INTRODUCTION

The most prominent example of a variable wing application in a proposed future reusable stage is Baikal. This first stage or booster of the Angara launcher, designed for fly-back to the launch site, is equipped with a wing to be rotated around a pivot located at the top of its fuselage. An extensive DLR study on reusable first or booster stages probed different sizes and design constraints of RLV vehicles [1]. This investigation within the system studies of the German future launcher technology research program ASTRA is based on a reusable booster stage dedicated for near term application with an existing expendable core. Analysis shows that such a winged fly-back booster in connection with the unchanged Ariane 5 expendable core stage is technically feasible and is a strong competitor to other proposed reusable and advanced expendable launchers. Realizing the fact that a single launch system application alone might not be sufficient to justify the development of a reusable stage, the options for continuous operation of such stages or of their derivatives in a timeframe of at least 50 years have been looked upon. These investigations showed a high interest in evolving reusable booster stages into various space transportation systems performing different tasks. However, the above mentioned investigations and preliminary RLV designs also revealed severe operational constraints acting on RLV during ascent due to their wings [1].

2

1.1 Problems related to wings on RLV In fact, the wing, as an eminent part for reusability, causes serious problems during the ascent phase: At least three major obstacles can be identified. Undesired but unavoidable strong aerodynamic moments arise from unsymmetrical launchers with a winged stage. One example of this type of asymmetrical winged RLV is the proposed French EVEREST concept [4], for which these effects have been carefully taken into account [5]. A launcher investigated in reference [1] is another projected asymmetric configuration with a winged Reusable First Stage (RFS) and an expendable belly attached core (see Figure 1). The two stages are mounted and operated in parallel up to RFS separation.

Figure 1: Asymmetric launcher configuration with extended wing on RLV (at top) and expendable stage (at bottom)

A critical problem of the RFS with parallel-operating upper stage arises from the unsymmetrical thrust load and the cog-movement perpendicular to the flight direction. Most important is to achieve a moment balance at each instant of the trajectory which can be fulfilled by thrust vector control, though this automatically increases the angle of attack. This undesired but unavoidable behavior is especially critical for the winged stage, as the wing under AOA creates strong aerodynamic moments. A flight dynamics simulation of the ascent trajectory found the resulting moments to be at or beyond TVC limits. The maximum engine deflection in normal (z-) direction reaches more than 6.5° (Figure 2). These deflections would have been much smaller for a conventional, fully expendable launcher because the disturbing aerodynamic effects of the airfoils would disappear. Actually, the wing causes serious problems during the ascent phase to which a conventional ELV without wings will not be exposed. A technical solution could be found by reducing the undesirable effects caused by the wing during ascent flight. Namely, a variable wing which can be folded or retracted and thus is able to avoid or reduce negative impacts caused by aerodynamic forces encountered during ascent, is such a solution. Figure 2 demonstrates that an up-folding wing (like for carrier airplanes) considerably improves the situation. Preliminary aerodynamic analysis performed of the adapted RFS configuration and succeeding flight dynamics simulation, using the DLR tool cac [10], show an acceptable ascent trajectory control (Figure 2), if the RFS engines are pre-inclined by about 6 degrees. Note that the folded airfoil is only one of several options for

the technical design. The search for the most suitable variable wing set-up is systematically addressed in this paper.

-18

-16

-14

-12

-10

-8

-6

-4

-2

0

0 50 100 150time [s]

eng

ine

def

lect

ion

[deg

]

fixed wing

folded wing

Figure 2: Required TVC angle of asymmetric RFS launcher with different wing configurations during ascent flight

Furthermore, configurational reasons limit the possible stage arrangements of multiple winged RLV to form super heavy lift launchers (SHLL) or fully reusable TSTO. A SHLL investigated in [1] consists of a central core stage, five circumferentially attached LFBB, and a small re-ignitable injection stage. The central core has a diameter of 10 m, similar to the Saturn V second stage S-II. The circumference is slightly above 62 m, allowing the integration of five LFBB boosters, only if some kind of variable and retractable wing is used. Figure 3 illustrates the need of a variable wing due to geometric constraints. Note that the depicted foldable-wing is only intended to demonstrate that a variable wing fits within the available geometry, but no technical solution of a mechanism should be proposed.

Figure 3: Schematic view of SHLL launcher with five attached RLV as booster in foldable-wing configuration

Symmetrical launchers in a dual winged booster configuration would also benefit if disturbances arising from the wing could be marginalized in ascent flight. Although they might not be subject to geometrical limitations or strong aerodynamic moments in nominal flight, the increased, sometimes critical, perturbations caused by winds and gusts when compared with a fully expendable system lacking any wings, can not be ignored.

3

The controllability during the ascent flight of the ASTRA reference LFBB configuration with expendable core has been investigated in a dynamic simulation [3]. Control is performed by thrust vector gimballing of all available rocket engines of the two boosters. The most critical instant during the ascent flight has been identified to occur at about 30 s after lift-off. The maximal obtained deflection of the booster engines takes an extreme value for a wind gust of 10 m/s at that time. The required nozzle deflection angle of all six booster engines to counter the perturbation is shown in Figure 4. The simulation shows that the most demanding deflection value of approximately –3 degrees stays well below the limits of 5.5°. The remaining maneuvering margin is however reduced. The presented results prove the controllability of the examined winged launcher configurations during the mated ascent until booster separation, if all booster engines are available for two dimensional thrust vector control. However, in the case of an engine failure or TVC anomaly the existing margin might become critical. In such an undesirable case a retracted variable wing would add robustness to the launch vehicle. The operational constraints due to high wind conditions might be eased for a variable wing launcher, improving its launch readiness over conventional types.

-3.5-3

-2.5-2

-1.5-1

-0.50

0.5

0 25 50 75 100 125 150

time [s]

TV

C [d

eg]

Figure 4: Thrust vector deflection angle δTVC as function of ascent flight time for an LFBB configuration subject to wind and gust [3]

The three examples discussed in this section describe the main advantages of variable wing for reusable launch vehicles:

• better controllability during ascent due to reduced wind and gust influence and improved low eigenfrequencies,

• easing of geometrical constraints for multiple booster stage arrangements, and

• enabling asymmetrical winged configurations which might be attractive for partially reusable small launchers.

Additionally, the drag of a wing reduces payload performance. However, it has to be acknowledged that the negative impact of drag on vertical launchers’ payload mass is quite small. On the other hand, any large movable structure on a launcher may lead to severe drawbacks:

• potentially increased weight due to mechanisms allowing for movement,

• a strong demand on the reliability of the mechanisms due to the risk of RLV loss in the case of a malfunctioning, and

• concerns about the aerothermal behavior of the structure due to possibly unavoidable breaches or gaps,

1.2 Already Realized or Proposed Variable Wings on Space Transportation Vehicles Although, technical design of a variable wing is still in an embryonic stage it is instructive to first have a look at already realized or proposed designs for space transportation vehicles. One recent and probably the most prominent example of a variable wing application in a proposed future reusable stage is Baikal. This first stage or booster of the Angara launcher, designed for fly-back to the launch site, is equipped with a wing to be rotated around a pivot located at the top of its fuselage. The propellant choice of this Krunichev-designed stage is LOX-Kerosene. Baikal has been presented at the Le Bourget airshow in June 2001 as a full scale technology mockup. Some derivatives of increased size in the role as a reusable Ariane 5 booster are currently under investigation at TSNIIMASH dubbed Bargouzin of which the LOX-Kerosene versions maintain the variable wing [6]. Another proposal of the 80s was a future evolvement of the Energia-launcher (GK-175). The boosters were to become reusable stages with forward swept variable wings [7]. A very interesting example of a variable wing application in the history of space flight is the orbital stage of the Soviet Spiral vehicle. Its re-entry flight demonstrator Bor successfully flew four times from 1982 through 1984 (Kosmos-1374, Kosmos-1445, Kosmos-1517, Kosmos-1614) into a high inclination orbit before reentering the atmosphere and being recovered at sea. The outboard wings are folded upward in hypersonic high angle of attack flight (see Figure 5) to enhance stability and to reduce thermal loads on the leading edges. The wings are subsequently deployed to a horizontal position in subsonic and landing configuration to increase L/D [8], [9].

Figure 5: Bor-3 (top) and Bor-4 (bottom) re-entry flight demonstrators with foldable wing [8]

4

2 STUDY LOGIC AND CONSTRAINTS

DLR-SART has studied a broad range of reusable booster stages (see for some examples ref. [2]) in the last five years. The experience from the German ASTRA research program which saw SART in the design lead for a Liquid Fly-back Booster with Ariane 5 core stage delivered a large amount of valuable data for this investigation on variable wings. Nevertheless, to obtain plausible results which are transferable to various different RLV designs, the study has not been carried out with existing ASTRA booster concepts and their specific constraints.

2.1 Constraints and requirements on the variable wing RLV stages First of all, any requirement on the attachment of a specific upper stage or orbiter has been intentionally excluded. But, mounting of the RLV to another stage should be possible by three attachment points without any restriction on their exact position. However, these points have to be always on the lower side of the RLV where the landing gears are also to be integrated as the upper side is reserved for vertical fins which prohibit any close attachment to another stage. Two different classes of generic vehicles have been defined:

• a booster or first stage using LOX-LH2 as propellant, and

• another booster or first stage using LOX-Kerosene as propellant

A propulsion system investigation is not in the focus of this paper. Therefore, all engine specific impulse, mass flow, and engine weight are set to fixed values based on rocket engine data of previous studies [1], [11]. While a gas generator cycle is used for LH2, staged combustion is employed for the kerosene booster. The cycle choice has been influenced by the earlier work but has no significant impact on this study. For comparability reasons the amount of propellant is fixed that all vehicles deliver a similar total impulse. The nominal ascent propellant of the LOX-LH2 propelled booster is 167500 kg and that of the denser but less performing LOX-RP reaches 200000 kg. Tank mixture ratio is fixed at 5.9 and 2.72 respectively. Including residuals, reserves and necessary engine start-up propellants the tanks have to be capable of at least loading the amount listed in Table 1 at normal boiling point conditions for cryogenics and 300 K for kerosene.

[kg] LOX Fuel

LOX/RP-1 150839.4 55455.6

LOX/LH2 148267.1 25130.0 Table 1: Fixed amount of propellant to be loaded

Consequently, all investigated concepts are similar in tank volume due to the fixed propellant loading. Although, a different number of tanks, design solutions (integral vs. non-integral), and altered arrangements inside the fuselage are accepted. Based on a generic vehicle with circular fuselage the propellant combinations LOX/LH2 and LOX/RP1 are analyzed for low- and shoulder wing positions, and high

aspect ratio vs. delta-type shape. Variability options are checked for swing-wing, rotational, foldable lay-out. To reduce the technical risk and complexity, it will be assumed that only two wing positions exist: deployed and undeployed. During ascent the aerodynamic lifting device is in its stored configuration, and the deployment is subsequently realized during the stage's ballistic phase outside the atmosphere and prior to reentry. Afterwards the wing's configuration and attitude remains unchanged until landing. The total RLV stage length and fuselage as well as the outer diameter of integral tanks are freely chosen. A suitable number of air-breathing engines, for fly-back, have to be installed on the vehicle. An option allows also the integration of non-integral tanks if required by the overall lay-out of fuselage and the wing attachment frames. The main wing's shape is strongly dependent on the type of variable wing. No particular airfoil is chosen but the relative chord thickness is about 12%. The wing’s surface and span is another important free parameter. The engines are protected on the lower side by a body flap, which could be used for aerodynamic trimming and control. Two vertical fins which are attached to the upper part of the fuselage and inclined at 45 deg. are also used for trimming. The latter configuration is chosen for its relative compactness, though flight dynamic drawbacks exist. Subsystem masses are based on data from previous ASTRA studies. This choice enhances credibility but restricts in no way the catholicity of the results. The maximum mechanical loads during re-entry shall not exceed nz= 3.5 g and maximum dynamic pressure is 60 kPa. The maximum heat flux is not explicitly restricted, but is assessed critical if it should approach, locally, 100 kW/m2. A variable wing drive mechanism or necessary bearing will not be designed in this preliminary study. To take care for a potential weight penalty an additional mass of 1000 kg is added for all variable wing configurations. This add-on mass should also cover additional farings or support structures probably required at the wing's free end.

2.2 Study logic The complete flow chart of the investigation on variable wings and preliminary evaluation and selection is shown in Figure 6. In the study's first step the tanks are sized, the overall vehicle is configured, and wings and stabilizers are dimensioned and arranged. Tank sizing is performed with pmp, a DLR tool under development for rapid tank, feedline, and pressurization system analysis, design and corresponding mass estimation. In the next step a stage mass is estimated and the corresponding center of gravity (cog) position is calculated using the program stsm. This DLR tool is designed for investigation of single and multiple stage space transportation systems (STS) mass (M) estimation and analysis. The amount of necessary stsm input is restricted to the absolute minimum to support the early evaluation process at the pre-design phase. The latest version stsm 1.3 is able to include sub-component data

5

sets from structural design at any degree of detail for implementation in stage mass analyses. To obtain reliable results for cog which is a prerequisite to check the flyability at a later stage, the geometric arrangement of all major components has to be known. One option to achieve this purpose is the imple-mentation of a commercial CAE program. However, this approach is quite time consuming and found to be unsuitable for this early selection of the most promising variable wing configurations. As an alternative a simple spreadsheet tool is used, in which the outer mold lines of tanks, fuselage and wings are plotted in 2D. All

subcomponent cog-positions as used in stsm are marked in these plots with red diamond shaped symbols (see Figure 7 and Figure 8 for example), and hence can be easily checked on their appropriate location. The latter becomes highly important because the most suitable configuration is found using an iterative process (see Figure 6) and relocation of some components during this procedure is not uncommon. The plots will always show if components would be wrongly positioned outside the vehicle or inside tanks, thus the calculated stage center of gravity would not be realistic. This straightforward tactic proved to be highly efficient for the current study.

program task change input output

pmp tank sizing # tanks geometric arrangement tank diameter cog of each tanktank position volume of each tank

surface of each tank

stsm stage mass and cog analysis wing data stage massesgeometric arrangement of components stage cog drycog of each tank and other components stage cog operationalvolume of each tank

cac aerodynamic data main geometric data wing cd(M,α)main geometric data fins / canards cl(M,α)main geometric data fuselage cm(M,α)

consistency check geometry-cog

cact trim analysis cd(M,α) flap deflectionscl(M,α)cm(M,α)stage cog drystage cog operational

flight evelope determination

check on flyability

check on technical feasibility

RFD return flight dynamics simulationprevious stage data fly-back propellant

consolidated inert separation massVariable wing design ranking

-

-

-

+

+

+

?

?

?

poor or unfeasible

design solution

promising or acceptable design solution

Figure 6: Flow chart for the study logic on variable wing design options

An important and powerful capability of stsm is the calculation of center of gravity movement and pitching inertia change along the flight trajectory’s history. Any record for the depletion of tanks can be modeled by including fluid mass and inertia sets for further flight dynamic analysis. The aerodynamic configuration is found and preliminary data sets are generated with the DLR program CAC 2.24. CAC is a powerful tool for extremely fast preliminary analysis of launcher's and hypersonic vehicle's aerodynamics. Although based on no more than very simple basic geometry neglecting any interference, it is possible to obtain aerodynamic

coefficient data in good agreement with wind tunnel measurements or results of much more time consuming CFD-calculations [10]. It is important to select a suitable homologous geometry. For these reasons CAC is very efficient in identifying and also in quantifying a general trend, but is not well suited to study slight differences between more or less similar configurations. Calculating mass, center of gravity and aerodynamic data sets, enables to perform a flight dynamic assessment for descent and fly-back. This is done with the DLR tool CACT which performs a pre-trimming based on the cog history and the calculated efficiencies of all identified trimming devices (flaps, canards, etc.).

6

Although, the program always finds a numerical solution to the problem, this might be beyond actual deflection performances due to aerodynamic nonlinearities. Acceptable and save boundaries comparable to existing flaps have to be manually verified. At this point of the study the flight envelope is determined and flyability of the configuration in a typical re-entry and fly-back mission has to be checked. In any case of non-acceptable behavior or insufficient robustness against perturbations, a redesign or rearrangement of components like tanks or wing is unavoidable. Once this preliminary work is completed, unfeasible options, from a flight dynamics point of view, can be scrapped. Still the technical feasibility of the variable wing attachment to the fuselage frame has to be ensured. In some cases it turns out that integration of the main landing gear is severely restricted by the drive mechanism of the wing. This might be one criterion to discard a configuration type, if no better arrangement can be found. In a final step, the actual amount of fly-back propellant is estimated based on a full integration of a typical trajectory. Separation conditions of the reusable stage are kept constant, since no data on the core stage is available. The fly-back analysis is important for the evaluation process as the wing span and hence achievable L/D ratio strongly influences this mass item. The total inert mass at separation is always an important criterion in launch vehicle effectiveness. For each propellant combination two of the most promising combinations are investigated further. This includes improved aerodynamic analysis with local heat flow assessment and preparation for a first structural mechanic sizing of the wing, preliminary choices for the necessary mechanisms and mass estimation foreseen for a later stage of the variable wing investigations.

3 RESULTS OF THE STUDY ON VARIABLE WING OPTIONS

3.1 Investigated Configurations The study on variable wing options has so far investigated into more than 16 different lay-outs in total. A rigid nomenclature had to be established to keep control over the design process of the different configurations. In continuation of the recent ASTRA LFBB configuration series, dubbed “Y”, all RLV stages of the variable wing study are under the “Z”-label. A 10-series represents the LOX-RP propellant combination, while a 20-series stands for LOX-LH2. Two conventional fixed wing types have also been included in this investigation to be able to assess the advantages or drawbacks of variable wings on the same common basis. The nomenclature is as follows: Z xy-z with x=1 for LOX-RP, x=2 for LOX-LH2, y=1 for a high wing position on top of the fuselage, y=2 for a low wing position at the bottom of the fuselage, and z=1 for a fixed straight wing for reference,

z=2 for a fixed delta-type wing for reference, z=3 for a variable swept wing as used in some military aircraft, z=4 for a variable rotated wing as proposed for Baikal, z=5 for a variable folded wing as used in carrier aircraft. An overview of all the investigated configurations with their respective Z-names is provided in Table 2. Four options possible by this permutation are rejected before any analysis is performed. These are the fixed delta-wing type on top of the fuselage as these would be very unusual for a winged launcher and no potential interest could be identified. The others are a foldable wing in the high position which would increase total height of the stage, complicating the wing’s stored arrangement support without offering any advantage

3.2 Facts detected during the iterative design process Some configurations encountered severe obstacles during the iterative design process, which excluded the early preferred internal arrangement. In general, these difficulties are found to be considerably less severe with the propellant combination LOX-kerosene. The smaller tank volume demand allows for a much easier placement of tanks and wing attachment frames than with LOX-LH2. All hydro-carbon RLV-stages have an integral LOX-tank at the front and a separate, non-integral RP-tank at the aft (see example in Figure 7). The forward placement of liquid oxygen is the classical arrangement found in Baikal, Saturn V, and in most launchers’ first stages. The mixture ratio of 2.72 favorably enables the positioning of the wing attachment on top or below the non-integral kerosene tank at about 65 % of stage length, closely behind the cog. Although the fuselage and tank diameter is not a fixed value, all configurations are designed with 3.8 m. Therefore, the major part of the Z-10-series alterations focused on the wing design, its effective aerodynamic area, and its aspect ratio. From an aerodynamic point of view it seems attractive to increase wing span and hence aspect ratio to achieve relatively high L/D-ratio. This seems to be more attractive for a variable wing than for a conventional type, because interference restrictions during launch can be minimized by the retracted wing. However, the investigation shows that the maximum trimmed L/D does not exceed 9 but is strongly penalized by a high wing mass. The structural disadvantage more than outbalances the reduced fuel consumption for fly-back. The wing’s inner chord length should approach at least 4 m so that the maximum wing thickness is not below 0.5 m. The high bending loads during re-entry would otherwise demand unacceptable strengthening and would generate a weight disadvantage. Another important factor is the wing loading during re-entry deceleration conditions, which determines its minimum effective area. Maximum lift in the range of linear lift curve slope can be increased with angle-of-attack. Without more detailed aerodynamic analyses, the AOA is restricted to not more than 40 degrees in hypersonic flight. In the case of variable sweep wings (Z 1y-3) the two devices have to be mounted on top of each

7

other and swept forward in stored position to provide sufficient wing area (see Figure 7). Taking the load and flyability considerations into account, a feasible solution has been found for all LOX-RP Z-10-series RLV stages. However, if technical feasibility is checked for some reservations exist for the lower mounted sweep and rotation types Z 12-3 and Z 12-4. First, integration of the main landing gear is quite complicated. Another problem of this lower wing position is related to the fact that the wing has to be stored on the belly and in-between the RLV stage and the attached second or core stage. The minimum distance between the stages would increase by up to 1 m, which is undesirable for the stability of the whole attached launch configuration and will increase structural mass. Further, the possible attachment points are severely restricted, strongly reducing flexibility which was one of the major intentions for a variable wing booster stage. In conclusion the promising design solutions are the high position sweep (Z 11-3), rotation (Z 11-4), and the low wing position fold configuration Z 12-5. The latter is not effected by the other’s Z 12 disadvantages because the wing is stored in upward position during launch and the fold mechanism is assumed to be outboard of the fuselage connection.

21.495

-5

0

5

0 5 10 15 20 25 30 35

-15

-10

-5

0

5

10

15

0 5 10 15 20 25 30 35

Figure 7: Schematic sketch of variable sweep wing configuration Z 11-3 showing airframe (blue) and tanks (magenta), wing (blue) in stored and extended position, and cog of major components (red diamond shapes), all dimensions in m

The main difficulty of the Z-20-series of LOX-LH2 RLV stages is tied to the low density of liquid hydrogen requiring very large tank volume. Including the fly-back propellant this amounts to at least 400 m3. Despite an increase in tank diameter to 5.4 m it is found extremely hard to place the LOX-tank in the forward position without subdividing the LH2 on several tanks. The ASTRA study demonstrates that a viable solution with

oxygen on the top exists [1]. However, it requires canards and an overall arrangement which is not well suited for this variation of generic configurations. Therefore, all Z-20-series have the LOX tank in the aft close to the engines (see example in Figure 8) which is a disadvantage concerning their NPSP requirements due to reduced hydrostatic pressure. On the other hand this arrangement is not technically unfeasible and thus is maintained for the current study. At the beginning of the iteration it was tried to implement a forward attached large span wing, swept-back approximately 45° after deployment. This classical airliner-like configuration would have offered a sub-sonic L/D exceeding 10. But it turned out not to be trimmable in the complete flight regime. Therefore, the only acceptable wing position is starting at about 60% of fuselage length. The attachment is foreseen in all viable cases at the intertank ring structure between forward LH2 and aft LOX-tank. Both main tanks are assumed to be integral. Fly-back propulsion of the Z-20-series is also using hydrogen because the inert mass at separation can be significantly reduced [1], although on the downside requiring additional volume. A separate non-integral tank might be attractive for its storage. An examination to place this tank in the intertank structure close to the stage's cog was not successful and it showed up to be more suitable for trimming if implemented in the nose section. It is further assumed that the turbofans are installed directly above this tank, thus sufficient ample space has to be reserved. The overall length of the stage has been kept as short as possible for flight dynamic reasons. The demand to reduce the destabilizing influence of the nose at high AOA evolved into a relatively blunt design which delivered good volume efficiency but reduced L/D. The larger tank diameter compared to the Z-10-series eased storage of the retracted wings. The wings of the Z 2y-3 configurations have to be swept forward in storage due to the location of their attachment point and otherwise encountering interference problems with the vertical stabilizers. But it is not necessary to pack them one above each other as for the kerosene RLV. An impressive total span of 38 m is reachable for Z 21-3 and Z 22-3 with deployed wing. The span of the rotational wing is restricted by its attachment close to the stabilizers. An asymmetric pivot point is able to slightly increase this value. As for the Z-10-series, a feasible solution has been found for all LOX-LH2 Z-20-series RLV stages if load and flyability considerations are taken into account. Similar reservations exist for the lower mounted sweep and rotation types Z 22-3 and Z 22-4 if technical feasibility is checked. These concerns are even stronger than for the kerosene boosters because due to the integral LOX-tank behind the wing it seems to be almost impossible to integrate the main landing gear and the stage attachment points. Two technical options exist in theory for the gear: A very flat design inside the movable wing or a cut-out allowing landing gear extraction through the extended wing. Both have to be evaluated as heavy and high risk designs. Additionally, the minimum distance between the stages would increase by up to 1.1 m, which is undesirable for the stability of the whole attached launch configuration and

8

would increase structural mass. The Z 22-3 and Z 22-4 have both to be assessed as almost practically unfeasible. The promising design solutions are similar to the kerosene stages: The high position sweep (Z 21-3), rotation (Z 21-4), and the low wing position fold configuration Z 22-5. Figure 8 shows a schematic drawing of a possible fold-design. As can be seen from the lower sketch, it is desirable to have a sideward or end support for the wing in stored position. The same support will be required for variable sweep and rotate designs during launch. However, it seems to be technically more demanding in the upward fold case -5, because similar tight packaging to the fuselage will not be possible as for the -3 and -4 designs. This problem has to be addressed in a more detailed structural lay-out and analysis in the future. At the current state of the investigation no additional mass penalty has been included for Z 22-5.

23.782

-5

0

5

10

0 5 10 15 20 25 30 35 40

0

-5

0

5

-15 -10 -5 0 5 10 15

Figure 8: Schematic sketch of variable fold wing configuration Z 22-5 showing airframe (blue) and tanks (magenta), wing (blue) in stored and extended position, and cog of major components (red diamond shapes), all dimensions in m

fixed (reference) fixed (reference) variable variable variable

propellant wing position

straight, high aspect ratio Delta-type sweep rotation fold

LOX/RP-1 high wing Z 11-1 - Z 11-3 Z 11-4 -

LOX/RP-1 low wing Z 12-1 Z 12-2 Z 12-3 Z 12-4 Z 12-5

LOX/LH2 high wing Z 21-1 - Z 21-3 Z 21-4 -

LOX/LH2 low wing Z 22-1 Z 22-2 Z 22-3 Z 22-4 Z 22-5 Table 2: Nomenclature of investigated RLV boosters in variable wing study

Reference

fixed fixed fixed variable variable variable variable variable LOX/RP-1

Z 11-1 Z 12-1 Z 12-2 Z 11-3 Z 12-3 Z 11-4 Z 12-4 Z 12-5

rel. extended span to ref. 147.87% 147.87% 100.00% 158.51% 158.51% 89.89% 89.89% 147.87%

rel. retracted span - - - 21.81% 22.15% 22.49% 22.49% 25.18%

rel. retracted span to ref. - - - 34.57% 35.11% 20.21% 20.21% 37.23%

rel. stage distance to ref. 100.00% 100.00% 100.00% 100.00% 398.51% 100.00% 249.25% 119.40%

rel. L/D to reference 127.27% 127.27% 100.00% 145.45% 154.55% 100.00% 101.82% 136.36%

rel. dry mass 105.08% 105.08% 100.00% 114.43% 114.43% 103.67% 105.21% 109.83%

rel. sep. inert mass 99.41% 99.11% 100.00% 105.89% 105.46% 101.22% 103.46% 102.54% Table 3: Relative behavior of LOX-RP variable wing RLV compared to fixed delta wing RLV, best cases are highlighted in blue, worst cases in orange boxes, if appropriate

Reference

fixed fixed fixed variable variable variable variable variable LOX/LH2

Z 21-1 Z 22-1 Z 22-2 Z 21-3 Z 22-3 Z 21-4 Z 22-4 Z 22-5

rel. extended span to ref. 114.45% 114.45% 100.00% 181.21% 181.21% 95.37% 95.37% 114.45%

rel. retracted span - - - 19.21% 19.21% 28.75% 28.75% 31.25%

rel. retracted span to ref. - - - 34.81% 34.81% 27.42% 27.42% 35.77%

rel. stage distance to ref. 100.00% 100.00% 100.00% 100.00% 328.36% 100.00% 328.36% 100.00%

rel. L/D to reference 87.27% 103.64% 100.00% 110.91% 130.91% 78.18% 92.73% 109.09%

rel. dry mass 90.08% 90.08% 100.00% 97.72% 97.72% 90.37% 90.37% 92.75%

rel. sep. inert mass 90.99% 90.16% 100.00% 96.59% 95.70% 92.21% 91.30% 92.63%

Table 4: Relative behavior of LOX-LH2 variable wing RLV compared to fixed delta wing RLV, best cases are highlighted in blue, worst cases in orange boxes, if appropriate

9

3.3 Evaluation and ranking of variable wing designs After finishing the iterative design process and finding at least in principle technically feasible solutions, the next task is to evaluate the different concepts by quantifying the differences. The main criteria should be geometric efficiency and vehicle mass. The geometry effect is most important, since it is a strong justification to introduce variable wings for RLV. Two data sets should be compared to each other:

• the relative span of the retracted wing related to the span of a fixed delta-type (reference), and

• the relative distance to a belly attached upper stage related to a reference configuration.

The former should be considerably smaller than for the fixed wing to offer a significant advantage. The latter can not be further improved, but should not considerably exceed the distance required by the reference vehicle. The RLV dry mass and the inert or separation mass are two other criteria of paramount importance in launch vehicle assessment. These data are directly related to payload performance and hence cost effectiveness of a reusable system. The first takes into account the impact of wing lay-out and additional mechanisms or strengthening. The second furthermore considers the fly-back propellant mass which is influenced by cruise L/D. A full re-entry and propelled fly-back simulation is performed considering a quasi-optimal trajectory [12] to obtain a reliable propellant mass. The aerodynamic and flight dynamic improvement of a retracted wing during the launcher's ascent is another vital criterion. However, due to the limited capability of the preliminary aerodynamic analysis to distinguish between slight geometric differences, a comparison between the variable wing concepts is not very meaningful. A first quantified estimation relevant for all types is obtained from Figure 2. Perturbations by wind, gusts or undesired angle of attack can be reduced with a retracted variable wing by at least 50% compared to the conventional fixed type. All the geometry and mass data of Table 3 and Table 4 is normalized by the figures of their respective reference RLV-configurations with fixed delta-wing. This approach simplifies the evaluation process. For retracted span and L/D-ratio best case values are highlighted in a blue box, to identify the strongest improvement. The interstage distance and the mass data are marked with an orange box if this type experiences the strongest degradation or disadvantage. Table 3 compares the LOX-kerosene propelled Z-10-series of RLV-stages. The span can be reduced during launch by at least more than 60% and almost 80 % for the rotated wing. The low attachment Z 12 is penalized by an up to fourfold increase in stage distance. The corresponding obstacles have been discussed in section 3.2. L/D is mostly improved with the large span variable sweep wing. The L/D difference between lower and upper attachment is consistent with the experience from aircraft design. However, the exact value is also due to CAC modeling and should not be interpreted too far

reaching. The L/D of the rotated wing (Z 1y-4) is close to the reference which is influenced by the relatively small wing, limited in size by its integration constraints. The long and thin wings of Z 1y-3 cause the heaviest weight which is compensated in part by lower fly-back fuel consumption. Hence, the inert mass of all fixed and variable wing kerosene stages is within a narrow band of 6% increase at most. Such a small difference is within the estimation inaccuracy of the preliminary analysis. Table 4 compares the LOX-LH2 propelled Z-20-series of RLV-stages. The span can again be retracted during launch by at least more than 60% and more than 70 % for the rotated wing. The smaller improvement of the rotation type compared to the corresponding kerosene-boosters is to be explained by the larger fuselage diameter and a small part of the leading edge still reaching beyond the vehicle's body in its retracted position. The low attachment Z 22 vehicles have to accept a more than threefold increase in stage distance. L/D is reduced in comparison to the reference stage in case of the rotated wing because of its relatively small size. Its dimensions are strongly limited by integration constraints of pivot point and fin position. The best situation is once more found for the relatively large span of the variable sweep wing. The differences in aerodynamic quality between Z 21 and Z 22 should be interpreted cautiously. For the LOX-LH2 RLV stages the reference type turns out to be the heaviest. This, on a first look surprising result, is due to mainly two facts. The large tank volume of the cryogenic booster severely restricts attachment of a delta wing. Therefore, the LOX-tank of Z 22-2 is the only one of non-integral form of all Z-20-series and additionally canards have to be added for trimming. Generally, a delta wing is heavier than a straight wing and the latter's aerodynamic acceptability for the complete mission has to be verified. The inert mass of all variable wing hydrogen stages is slightly below the reference. The small difference is, however, once again not significant. The high position rotational wing concepts Z x1-4 seem to be very promising from their mass and compactness figures. However, the relatively small wing which can not be further enlarged might be a concern due to its landing speed and stall velocity. The high position sweep wing concepts Z x1-3 will provide excellent subsonic flight characteristics but numbers indicate that they are slightly heavier than their counterparts. The up-folding wing Z x2-5 is quite attractive but is somewhat less compact and might require a more complicated support structure. The above mentioned stages are promising concepts but it turns out to be impossible to give a clear ranking based only on the preliminary analyses of generic RLV configurations. In fact more detailed analysis and the specific requirements of a defined launcher with its mission will be necessary for the final choice.

4 FUTURE PERSPECTIVE OF VARIABLE WING CONFIGURATIONS

4.1 Proposed Next Investigation Steps The configurations selected for more detailed analysis of the variable wing options in the future are:

10

• Z 11-3, • Z 11-4, • Z 21-3, and • Z 22-5

They are feasible and promising concepts and represent both propellant combinations and all three types of variable wing: -3: sweep, -4: rotate, -5: fold. The next step should be a preliminary mechanical sizing and analysis of the interface fuselage-wing and of the variable wing drive mechanism. Afterwards more detailed data on reliability and technical feasibility of the required devices are available. Additionally, one will obtain the mass of wing, attachment point and mechanism, supporting an evaluation process between the remaining concepts. Any detailed design process requires relevant load data, which in sufficient quality is already available. The flight dynamic re-entry and fly-back simulation delivers such information. Acceleration levels, dynamic pressure, Mach, altitude, α-histories identify the most critical flight conditions. One such point is the maximum deceleration close to the maximum normal load nz and also maximum heat-flow. It is possible to perform a preliminary lay-out at the most critical static load by merging these three conditions in one single design point. This situation is found in the hypersonic regime at (depending on the configuration) approximately M=5.2 in an altitude close to 30 km, and α about 20 to 25 degrees. Any structural analysis will benefit from a pressure and heat flow distribution on fuselage and wing structure. The aerodynamic design tool CAC has now to be replaced with a more elaborate but still fast code. The DLR program HOTSOSE is a surface inclination Newton-scheme enhanced by compressible boundary-layer theory such as Fay-Riddel-method for heat-flow estimation. Two examples of this analysis are shown in the following figures. Figure 9 depicts the pressure distribution of the variable sweep type Z 21-3. Figure 10 shows the heat-flow on the wing of the foldable type Z 22-5. A maximum value of about 62 kW/m2 at the leading edges is acceptable.

Figure 9: Pressure distribution cp on Z 21-3 wing and fuselage section at maximum re-entry load, M=5.26, altitude= 30 km

Figure 10: Heat-flow distribution on Z 22-5 wing and fuselage section at maximum re-entry load, M=5.2, altitude= 30 km

More detailed aerodynamic analysis should be further used to more precisely evaluate the relatively small differences of the various concepts. Aerodynamic performances in the complete flight regime including ascent is to be recalculated to check the preliminary results of CAC.

4.2 Application of variable wing stages in a future scenario The final section of the paper describes a long-term, strategic scenario of a fly-back booster vehicle. The general idea is the gradual evolution of the LFBB outlined in [1] into at least three space transportation systems performing different operational tasks. If the reusable booster can support satellite launches from the lower end to the very high upper end efficiently with virtually the same type of vehicle, production can be surged to numbers otherwise not realistically achievable by a reusable stage. In combination with further operational synergies considerable cost reductions should be reached. Starting with the heterogeneous expendable launcher family of Vega, Soyuz, and Ariane to be operated from Kourou within the next few years, a reusable and potentially common element can be introduced with the LFBBs replacing the EAP of the Ariane 5 ECB. Assuming an operational capability of the LFBB in combination with the expendable Ariane 5 core stage at mid of next decade, one can imagine an evolution of the reusable booster stage as shown in Figure 11. In a next step, the reusable boosters extend their application as a reusable first stage (RFS) in the class of small and medium size launchers like Vega and Soyuz. In a parallel burn, asymmetric configuration, the aero-dynamic moments of the wing are critical for ascent control of the launcher. Flight dynamic simulations prove that retractable airfoils significantly improve the situation. The second direction of evolution is the upper segment of a super heavy-lift launcher (SHLL). Payload should be close to the capability of the famous Saturn V and Energia boosters to support an ambitious space flight program like manned Mars missions. No showstoppers

11

could be found for this large launcher, but the boosters require variable wings for integration reasons. Eventually, the partially reusable system with Ariane 5 core might evolve into an RLV TSTO still relying on the (upgraded) LFBB as the first stage element if its wings can be retracted.

It has to be stated that a flexible scenario for a reusable booster stage as in Figure 11 is only sustainable if a variable wing configuration is implemented. A decision to design such a type of launcher with movable wing technology should be drawn early in the development process to avoid costly redesign or risky technical compromises.

2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031

Vega RFS + expendable upper stageLight - Weight

Soyuz (ST)

Medium - Weight

Ariane 5 + LFBB

Ariane 5 ECB TSTO (LFBB + Orbiter etc.)

Ariane 5 class Weight

5 LFBB + heavy (600 Mg) expendable upper stage

Heavy/super Heavy- Weight

Figure 11: Application scenario for variable wing RLV as future launchers for Europe based on similar booster stage

5 CONCLUSION

The wing of an RLV causes serious problems during the ascent phase. At least three major obstacles can be identified. Undesired but unavoidable strong aero-dynamic moments arise from unsymmetrical launchers with a winged stage. Limitations in available space for the arrangement of multiple winged boosters are found for a large launcher. RLV experience considerably more critical perturbations caused by winds and gusts compared with a fully expendable system lacking wings Therefore, a variable wing for a reusable launch vehicle offers the following potential advantages:

• better controllability during ascent due to reduced wind and gust influence and improved low eigenfrequencies,

• relaxing of geometrical constraints for multiple booster stage arrangements, and

• enabling asymmetrical winged configurations which might be attractive for partially reusable small launchers.

On the other hand, the drawbacks of a large movable structure on a launcher have to be recognized. The advantages and critical points have been investigated and evaluated on a preliminary basis. In this study two different classes of generic vehicles have been defined:

• a booster or first stage using LOX-LH2 as propellant, and

• another booster or first stage using LOX-Kerosene as propellant

These generic vehicles are combined with three main technical options for the design of the variable wing: swing-wing, rotational, and foldable lay-out. All investigated concepts are similar in tank volume due to a fixed propellant loading. The concepts may however possess a different number of tanks, different design solutions (integral vs. non-integral) and altered arrangements inside the fuselage are accepted. The tank, vehicle mass, and wing sizing is performed in an iterative loop. Data consistency is assured by consistency checks. Flyability and trimmability of all RLV configurations is verified during the design process and a rearrangement of components is performed if necessary. Taking the load and flyability considerations into account, a feasible solution has been found for all RLV stages. However, the design study quickly revealed that a low attachment of the variable wing implies serious concerns about its technical feasibility. In general, difficulties in finding a suitable arrangement of tanks and wings are more severe with the propellant combination LOX-LH2. The smaller tank volume demand of kerosene allows for a much easier placement of tanks and wing attachment frames than with LOX-LH2. Based on the preliminary analysis of this study, it appears that the introduction of a variable wing might not lead to a significant mass disadvantage. The high position rotational-wing concepts seem to be very promising due to their mass and compactness figures. However, their relatively small wing which can not be further enlarged might be a concern in terms of landing speed and stall velocity. The high position swept-wing concepts provide excellent subsonic flight characteristics but numbers indicate that they are slightly heavier than

12

their counterparts. The up-folding-wing is quite attractive but is somewhat less compact and might require a more complicated support structure. Although some promising concepts are identified, it turns out to be impossible to give a clear ranking based only on the preliminary analyses of generic RLV configurations. In fact, more detailed analyses and the specific requirements of a defined launcher with its mission will be necessary for the final choice. The next steps in the investigation of variable wing RLVs should be a preliminary mechanical sizing and analysis of the interface fuselage-wing and of the variable wing drive mechanism. Afterwards more detailed data on reliability and technical feasibility as well as the mass based on structural sizing of the required devices are available. More detailed aerodynamic analysis in the complete flight regime including ascent should be performed to more precisely evaluate the relatively small differences of the various concepts. An application of variable wing RLV booster stages in a long-term launcher scenario is described. At least three space transportation systems performing different operational tasks from the lower end to the very high upper end of payload capability can be identified for such a fly-back booster. As an RLV would be able to replace a whole pallet of boosters and first stages with virtually the same type of vehicle, production can be surged to numbers otherwise not realistically achievable by a reusable stage. In combination with further operational synergies considerable cost reductions can be envisioned. It has to be stated that the presented flexible scenario for a reusable booster stage is only sustainable if some kind of a variable wing configuration is implemented. A decision to design such a technology in a launcher stage should be drawn early in the development process to avoid costly redesign or risky technical compromises.

6 REFERENCES

1. Sippel, M.; Manfletti, Ch.; Burkhardt, H.: Long-

Term / Strategic Scenario for Reusable Booster Stages, IAC-03-V.4.02, 2003

2. Sippel, M.; Atanassov, U.; Klevanski, J.; Schmid,

V.: First Stage Design Variations of Partially Reusable Launch Vehicles, J. Spacecraft, V.39, No.4, pp. 571-579, July-August 2002

3. Sippel, M.; Klevanski, J.; Burkhardt, H.; Eggers,

Th.; Bozic, O.; Langholf, Ph.; Rittweger, A: Progress in the Design of a Reusable Launch Vehicle Stage, AIAA-2002-5220, September 2002

4. Iranzo-Greus, D.; Deneu, F.; Le-Couls, O.; Bonnal,

C.; Prel, Y.; Guedron, S.: Selection and Design Process of TSTO Configurations, IAC-03-V.4.05, 2003

5. Iranzo-Greus, D.; Cerf, M.; Duranté, N.; Leonard, C.; Bonnal, C.: TSTO RLV Mission Analysis and Complete Trajectory Optimization Strategy, 5th International Conference on Launcher Technology, S17.3, Madrid November 2003

6. Guédron, S.; Prel, Y.; Bonnal, Ch.; Rojo, I.: RLV

CONCEPTS AND EXPERIMENTAL VEHICLE SYSTEM STUDIES: CURRENT STATUS, IAC-03-V.6.05, 2003

7. Губанов, Б.И.: Триумф и трагедия "Энергии",

Вариант "Энергия-2" или ГК-175 , to be found at http://www.buran.ru/htm/41-3.htm

8. NN: Авиационно-космическая система

СПИРАЛЬ: подробности, to be found at http://www.buran.ru/htm/aviaspi2.htm

9. Lukashevich, V, P.: Predecessor of Shuttle and

Buran, Spiral orbital aircraft programme, http://www.buran.ru/htm/str126.htm

10. Klevanski, J.; Sippel, M.: Beschreibung des

Programms zur aerodynamischen Voranalyse CAC Version 2, SART TN-004-2003, DLR-IB 647-2003/04, March 2003

11. Burkhardt, H.; Sippel, M.; Herbertz, A.; Klevanski,

J.: Effects of the Choice Between Kerosene and Methane on Size and Performance of Reusable Liquid Booster Stages, AIAA-2003-5122, July 2003

12. Klevanski, J.; Sippel, M.: Quasi-optimal Control

for the Reentry and Return Flight of an RLV, 5th International Conference on Launcher Technology, Missions, Control and Avionics, S5.1, November 2003

Further updated information concerning the SART space transportation concepts is available at: http://www.dlr.de/SART