Aerodynamics of airframe-engine integration of supersonic aircraft

AERODYNAMICS OF

AIRFRAME-ENGINE INTEGRATION

OF SUPERSONIC AIRCRAFT

by Mark R. Nichols

Lctngley Research Center

Lmzgley Station, Hampton, Va.

N A T I O N A L A E R O N A U T I C S A N D SPACE A D M I N I S T R A T I O N W A S H I N G T O N , D. C . A U G U S T 1 9 6 6 TECH LIBRARY KAFB, NM

1llllllllllllllIIIllllllllllHlllllllllllllll

..

0330230 NASA TN D-3390 AERODYNAMICS OF AIRFRAME-ENGINE INTEGRATION OF SUPERSONIC AIRCRAFT By M a r k R. Nichols Langley R e s e a r c h Center Langley Station, Hampton, Va.

N A T I O N A L A E R O N A U T ICs A N D SPACE ADMlN I S T R A T I O N For sole by the Clearinghouse for Federal Scientific and Technical Information

Springfield, Virginia 22151 - Price $2.00

AERODYNAMICS OF AIRFRAME-ENGINE INTEGRATION OF SUPERSONIC AIRCRAFT* By Mark R. Nichols Langley Research Center SUMMARY Some major aerodynamic problems involved in the design of the propulsion- system installations of advanced supersonic aircraft are reviewed from the airframe viewpoint.

Consideration is given to environmental factors such as flow fields, the airframe boundary layer, airframe-inle t and inlet- inlet shock interference, and jet- airframe interference .

New aerodynamic tools available to the designer are discussed and are used along with experimental data to analyze the drag minimization problem. Finally, detailed attention is given to the problem o f optimizing the propulsion system itself from the standpoint of inlet and nozzle selection, flow distortion effects, handling of the various secondary air- flows, and afterbody drag.

INTRODUCTION Manned aircraft have now been flying supersonically for nearly 19 years. Most of date has been accumulated in brief dashes by aircraft opti- the supersonic flight time to mized primarily for subsonic operation. The recent flights of the XB-7OA and the A-ll/SR-71, however, mark the advent of an entirely new class of vehicle optimized primarily for supersonic operation.

The propulsion systems of the new supersonic aircraft differ markedly from those of previous aircraft as a consequence of the much higher level of propulsion-system efficiency required. The designer cannot afford to let this efficiency slip more than several percent below the theoretical optimum, even at the expense of increased complex- ity and weight. At the same time, interactions between the airframe and the engine have become of even greater significance than previously. It is no longer possible to think in t e r m s of optimizing a propulsion package and then attaching it to a separately optimized airframe. The two parts must be designed together and fully integrated in all respects.

The purpose of this paper is to review major aerodynamic aspects of this integration problem.

*This material was originally presented at the 27th Meeting of the AGARD Propulsion and Energetics Panel, Paris, France, April 4, 1966.

It i

SYMBOLS area A cross-sectional area of two nacelle bases A2b area at cowl lip Alip cross-sectional area of two nacelles A2N cross-sectional area of four nacelles A4N area of excess flow Ax, projected boattail area AP,proj b wing span drag coefficient CD base drag coefficient CD,b wave drag coefficient cD, W boattail drag coefficient cD, P l i f t coefficient CL normal-force coefficient CN P - P,

pressure coefficient, -

CP q , base pressure coefficient CP,b resultant load coefficient (top-surface Cp minus bottom-surface Cp) CP,R d diameter D drag bypass drag secondary r a m drag spillage drag boattail drag gross thrust primary gross thrust distances as defined in figures l i f t Mach number local Mach number surface Mach number free-stream Mach number inlet mass- f low ratio static pressure total pressure average total pressure jet total pressure free-stream total pressure free-stream static pressure local dynamic pressure q , free-stream dynamic pressure R range S wing area jet temperature T j TS surface temperature T , free- stream temperature VA axial velocity VR rotational velocity rate primary weight-flow wP Wsec secondary weight-flow rate

rq) maximum secondary-to-primary weight-flow ratio

max a angle of attack nacelle toe-in angle; nacelle boattail angle P Sav average boattail angle 6 thickness of undisturbed boundary layer 6' thickness of boundary layer at base of strong shock 7 nozzle efficiency factor

e angle between local flow and line parallel to fuselage center line

A wing sweep angle - angle of sidewash peripheral angle DISCUSSION Environmental and Flow-Interference Considerations The performance of the propulsion system, as always, is critically sensitive to the environment. Items of concern are the characteristics of the airplane flow field, aero- dynamic interferences between the airframe and the propulsion package, and jet-airframe interferences.

Flow fields.- Pertinent features of the flaw about an isolated inclined body (see The angle 0 is the total angle between the local refs. 1 to 3) are illustrated in figure 1.

Ms is the surface flow and a line parallel to the center line of the body or fuselage, and Mach number. At an angle of attack of 12O, as shown by the plots of 0/a! and Ms/M,, the flow along the bottom of the body aft of the nose region is approximately alined with the local surface and the surface Mach number is decreased from the stream value. The flow across the body results in upflow angles along its sides greater than the angle of attack and in surface Mach numbers greater than the stream value. Accompanying this cross flow is a thickening of the boundary layer at the top of the body, the collection of this boundary layer into lobes along the aft top quadrants, and the formation of twin vor- tices which entrain much of the thickened boundary layer and then separate from the sur- face. These phenomena are qualitatively applicable to a strut-mounted nacelle and the section of the fuselage ahead of the wing.

The addition of a wing to the body breaks up the flow across the body and provides a broad sheltered region. Fig- ure 2 presents results of recent NASA A - A surveys of the flow about an arrow-wing 6- B A- A configuration at a longitudinal station typical of inlet location for a bomber or . 9 k transport. (Also see refs. 4, 5, and 6.)

3r The operating conditions chosen are representative of cruising flight for such aircraft. As is discussed subsequently, 0 90 180 the most desirable region for locating a i r #J, DEG inlets generally is beneath the wing-body, The contour plots of M i and t3Mt/aa Figure 1.- Pertinent features of flow about an isolated inclined bodv.

in figure 2 show that the local Mach number beneath the wing-body is 0 . 3 to 0.4 lower than that above the wing-body, and that this advantage of the low position increases rapidly with increasing angle of attack. To be alined with the local flow, a nacelle air intake located in the plane of symmetry would be placed approximately parallel to the body center line. If located under the wing just inboard of the tip, it would be toed in 5 ' with a 2' down tilt. The rate of change of total flow angularity with angle of attack (W/aor) is less than unity for the entire region beneath the wing and also above the fuselage. Hence, inlets located in these regions experience less severe flow angularity problems than inlets located in the free stream.

dis- Flow-field effects on inlet.- Estimated effects of the flow characteristics just cussed on the pressure recovery at M m = 3 of a typical high-performance mixed- compression axisymmetric inlet placed in several loca- M , 3.0 tions relative to the airframe a = 6O are presented in figure 3.

Experimental angle-of-attack characteristics of this inlet operating in the free stream,

ao 3

- aa aa obtained from reference 7, 1 .02 are shown by the dashed / curve. Experimental data relative to flow angularity effects on the performance

-.02 -

.75 -

- . 0 3 - ' of a wide variety of intake installations are available in Figure 2.- Flow field for an arrow-wing configuration.

the literature (e.g., refs. 3, 8, and 9), but generally are not suitable for the present purpose because of low design pressure recovery or other factors. It is clear UNDER WING that the optimum inlet loca- $ UNDER FUSELAGE tion is beneath the fuselage OVER FUSELAGE in the plane of symmetry.

\ An inlet beneath the wing \ \\<FREE STREAM provides comparable per- \ \ formance over an angle-of- \ \ attack range of about 5O in

% h A 6 ;I IO

a, DEG the vicinity of the design value, but suffers significant Figure 3.- Effects o f flow field on pressure recovery of a typical high- performance mixed-compression inlet.

M, = 3.

losses outside this range due to the change in sidewash beneath the wing. Surprisingly, an inlet located above the fuselage center line likewise can provide good performance up to cruise angles of attack, and thus it also can be considered as a candidate inlet type for cruise aircraft. Unfortunately, as the angle of attack is further increased, the perform- ance of an inlet located above the fuselage center line deteriorates at a rate even more rapid than does the performance of an inlet located in the free stream, due mainly to the rapid buildup of local Mach number. Its performance also is very sensitive to spanwise location because of rapidly varying flow-field characteristics.

Interference of adjacent surfaces with inlet.- It is well known that the boundary layer plays a major role in determining how close an inlet can be placed to an adjacent surface. As shown by the left-hand plot in figure 4, the pressure recovery of typical inlets (see refs. 3, 8, 10, and 11) increases so rapidly with increases in the lip location h/6 that it is almost always advantageous to locate the inlet entirely outside parameter the boundary layer. It is essential at the same time that the strut or diverter wedge between the inlet fairing and the surface be aerodynamically thin and have its leading edge (or leading edge of the boundary-layer separation plate, if used) located aft of the inlet plane.

The shock-boundary-layer interaction problem may necessitate greater wing- nacelle spacing than that required to optimize pressure recovery at the design condition.

Below the design flight Mach number, the nose shock of a center-body inlet passes ahead of the cowl lip and impinges on the adjacent surface. With center-body-nose included angles no greater than 25O or 30' and, for Mach numbers of 3 or less, the strength of this shock is not sufficient to STARTED I NLET UNSTARTED INLET separate a stable turbulent it boundary layer, although may cause some local dis- turbance and thickening.

When the inlet unstarts or buzzes, however, the strength of the shock front may be suf- I . O r ficient to cause boundary- layer separation down to flight Mach numbers approaching 1.3. When such separation occurs, a boundary-layer wedge is

formed which greatly thick -

ens the boundary layer at the Figure 4.- Role of boundary layer in determining location of inlet relative adjacent surfaces.

inlet station.

L .

The amount of boundary-layer thickening due to separation caused by shock impingement is a complicated function of Mach number, Reynolds number, and other fac- tors (e.g., see ref. 12). The right-hand plot in figure 4 presents results of an attempt to correlate experimental values of 6'/6 as a function of the free-stream Mach number.

The width of the band is indicative of the uncertainty that exists at the present time.

Additional research is needed to define the phenomena more exactly. However, the con- clusion can be drawn that the lip of an inlet of the type shown in figure 4 must be sepa- a distance several times greater than rated from the surface by 6 to avoid some of the separated boundary layer from entering the inlet and thereby complicating the restart process. An alternative solution might be the use of a horizontal-wedge inlet oriented so as to direct the ramp shocks away from the adjacent surface or a vertical-wedge inlet mounted beneath a boundary-layer separation plate. The recent flight research finding that an unstart of one side of the XB-7OA inlet frequently initiates an unstart of the other side raises questions concerning the soundness of such approaches. The time interval of 3 to 4 seconds observed between such successive unstarts indicates, however, that the problem in this particular instance probably is occasioned by airplane motions accompa- or other factors rather than by interactions between the left-hand nying the initial unstart and right-hand inlet shock systems. Thus, consideration of this general approach cannot be discarded at this time.

Mutual interference ~ of adjacent inlets.- - In a multiengine installation it is important to separate the inlets sufficiently so that the unstarted bow wave of one inlet will not enter and cause the unstarting of adjacent inlets.

Figure 5 presents a compilation of spacing requirements data, obtained mainly from references 13 and 14, for high- performance mixed-compression inlets with the forward inlet buzzing and the rearward inlet operating at high pressure recovery.

At & = 3 and with the inlets located abreast, separation of about lq inlet diam- e t e r s appears to be ample. How- ever, when the inlets are stag-

/

gered, as is desirable from the drag minimization viewpoint, the separation distance required increases markedly. In extreme

z ' I \

cases, some sort of separation

zb-2 z i

SHOCK ENTERS INLET plate, such as is discussed in reference 13, may be required to arrive at a practical configura- 1.5 2.0 2.5 3.0 Mal tion arrangement .

Two reservations con- Figure 5.- Spacing requirements for high-performance mixed-compression cerning the applicability of the inlets (forward inlet in steady-state buzz). & = 3.

, data in figure 5 for design purposes must be noted. First, the travel of the initial unstart bow wave and the buzz amplitude of the configuration being considered may differ from those for the models studied because of the different system dynamic characteristics.

Second, if the inlets under consideration are located close to an airplane surface, the nature of the mutual interaction between the two inlets will be modified by shock- boundary-layer interaction effects at the airplane surface. Reportedly, in extreme cases, the separation requirements may be nearly doubled. At the present time, the designer has no recourse but that of further experimental study to insure avoidance of trouble.

Inlet separation _. requirements.- Although there are installations such as that of the XB-70A in which two or more engines are fed by a single duct, the current trend is to separate the intake flows to the greatest extent possible. The Concorde reportedly uses a splitter plate to separate its two-dimensional inlet into two parts. In figure 6 is shown a similarly split three-dimensional inlet designed for a Mach 2.5 airplane. In both cases, the prime requirement of the plate is that it prevent the unstarting and buzz of one side of the inlet from affecting the other side.

Experimental studies were conducted by NASA (see ref. 15) to establish splitter- plate requirements of the nacelle arrangement illustrated in figure 6. As shown by the sketch at the lower left of the figure, the model used incorporated an inlet fully repre- sented on one side of the separation plate but replaced by rapid-response total-pressure probes on the other side. In the tests, buzz w a s initiated'in the ducted side of the plate by reducing the duct exit area. Oscillograph traces of the pressure fluctuations measured by the probes on the other side of the plate then were observed to determine whether or' not the flow on this side of the plate was affected. It w a s found that there was no meas- urable interaction so long as the inlet w a s located well away from the wing and so long as the splitter plate extended to the nose of the inlet center body and, in the plane of the inlet, was about 7 , percent wider than the inlet diameter. On the right-hand side of LI figure 6, a satisfactory split- ter plate and corresponding L oscillograph traces a r e shown in the top sketch.

When the same nacelle model was moved nearer the wing

E

(see middle sketch), exten- sive interaction occurred due / - l o . . r to shock-boundary- layer AP

- = 0.01(

interaction accompanied by Pt,av 0.10 SEC crossfeeding of disturbance pressures through the gap L-2614-15 between the plate and the Figure 6.- Splitter-plate requirements for twin-@ inlet. f& = 2.5.

wing. Sealing the gap between the separation plate and the wing eliminated this cross- feeding (see bottom sketch). Thus, a properly designed separation plate would appear capable of effectively isolating the two inlet flows provided the initial unstart shock front does not impose conditions more severe than those studied. Again, further experimental studies are needed.

Jet-airframe interference,- Basic jet phenomena (see refs. 16 and 17) are illus- trated in figure 7. At subsor& flight speeds and with a convergent nozzle o r an under- expanded convergent-divergent nozzle, the jet expands through a series of internal Mach diamonds that do not penetrate the jet boundary. The edge of the jet is a mixing region.

The jet acts as a weak sink causing an inflow of air from the external stream toward the jet axis. The jet itself is not a rigid body, but undergoes deflections due to aerodynamic forces at high angles of attack and in the presence of the ground. At supersonic speeds, inflow effects generally are negligible, but the expansion process within the jet produces an external shock system that SUBSONIC SUPERSONIC affects adjacent surfaces.

Jet deflection effects tend to INFLOW 7 at be more pronounced than subsonic speeds because the ratio of the jet dynamic pres- sure to the free-stream pres- sure is much lower.

Jet interference effects JET DEFLECTION JE? DEFLECTION greatly complicate the jobs of AT HIGH Q AND IN PRESENCE OF GROUND AT HIGH qw the aerodynamicist, the loads engineer, and the structural Figure 7.- Basic jet phenomena.

designer. An NASA jet- .32 HIGH TAIL effects investigation of a twin- JET ON AND OFF engine fighter model, reported in part in reference 18, affords a good illustration of typical problems encountered ‘ N .08 at subsonic speeds. In fig- ure 8 is plotted the normal- JET OFF force coefficient for the JET ON

- .08

horizontal tail, obtained by integration of pressure meas-

- .l6-

urements on the tail, as a 0 4 8 1 2 Q , DEG function of angle of attack.

For the high tail position, Figure 8.- Jet interference effects on tail normal-force ceofficient of a twin- engine fighter model. M, = 0.85; p t , j / b = 5.

there w a s no effect of jet operation on the tail forces. When the tail was moved to the low position, indicated by the dashed lines on the sketch in figure 8, the curve for CN as a function of a ! w a s shifted and rotated in a destabilizing direction, as would be expected as a consequence of the position of the wing wake and the change in downwash at the tail location. In addition, at this low tail position, operation of the jet produced both an addi- tional shift and an additional destabilizing rotation of the curve. Further important changes have been observed when the tail of an airplane of this general configuration is located close to the ground or when the tail is deflected sufficiently so that its leading edge penetrates the jet boundary.

Figure 9 presents temperature and static pressure distributions on the afterbody of the same model. For a jet pressure ratio of 5, which is applicable to some future high- pressure-ratio engines, the pressure-distribution curve exhibits a succession of peaks and valleys which are related to the expansion process within the jet and indicate jet attachment. For the lower jet pressure ratio of 2.7, which is typical for most present engines, the jet w a s no longer attached. Even at this operating condition, the temperature rises along the afterbody were of sufficient magnitude to eliminate the use of aluminum as a structural material. Likewise, although the static loadings imposed on the afterbody by jet operation were small in themselves, they were accompanied by pressure fluctua- tions of sufficient magnitude to affect the fatigue life of the structure. A number of con- figuration changes were investigated to explore means for reducing the severity of these adverse jet effects. The most successful was the introduction of secondary flow into the base areas between the jet nozzles and the afterbody (compare short- and long-dashed curves in fig. 9). The design solution adopted involved the use of significant quantities of secondary flow, high- SECONDARY FLOW temperature surface mate-

!%- BETWEEN NACELLES

pal AND BODY rials, and well-damped structural panels. Obviously

* 5.0 NO

a less costly design solution CP would be desirable.

- . 2 - Figure 10 shows how the nature of the jet interfer- ence problem changes as the flight speed progresses from the subsonic to the super- sonic regime. Incremental static loadings (sum of upper and lower surface pressure- coefficient change s) induced Figure 9.- Jet effects on afterbody of a twin-engine fighter model. & = 0.85; on a 60° delta wing by the jet a = 40.

I

exhaust of a nacelle located beneath the wing are plotted for flight Mach numbers of 0.90 and 1 . 0 5 . (See refs. 19 and 2 0 . ) At both Mach num- bers, the angle of attack was 5O and the jet pressure ratio was 5 . For the subsonic speed, both upper and lower surface pressures were affected by jet Mm=l.O 5 operation, with the greatest changes occurring on the lower surface in the .90 immediate vicinity of the jet exit. For the supersonic speed, in contrast, all 2 = 0 . 1 9 the pressure changes occasioned by jet b/2 operation occurred on the lower sur- Figure 10.- Jet-induced pressures on a 6@ delta wing. Exit face. A strong shock generated by the located 1 exit diameter below the wing at 1 = 0.41 and b/2 jet plume completely spanned the wing 69% chord; a = 5O; pt,j/p, = 5.

and produced significant changes in pitching moment, control effectiveness, drag, and wing loads. A s the Mach number w a s further increased, this shock swept back rapidly and caused further variations in inter- f e rence characteristics.

The jet interference effects considered in this discussion of figures 8 to 10 cannot For be predicted accurately for complex configurations by available theoretical methods.

the most part, the designer is forced to rely on empirical correlations or model tests.

The need for continued research in this area is evident.

Integration of Propulsion System With Airframe New analytical tools.- The development of machine computing techniques has pro- vided the designer with powerful new tools for optimizing the aerodynamics of the vehicle.

The increase in capability has been particularly great in the supersonic regime where efficient vehicles must be aerodynamically slender so that linear theory methods may be used.

In early applications of area-rule methods for the purpose of minimizing the zero- lift wave drag, it w a s necessary to obtain the many required area cuts by graphical or semigraphical methods. Now, most manufacturers have computer programs that elimi- nate this drudgery. (See ref. 21.) As indicated in the left part of figure 11, the engineer provides the computing machine with data defining the cross-sectional areas of the fuse- lage and nacelles as functions of fuselage station, data defining the airfoil coordinates of the wing and tail as functions of spanwise station, and coordinates defining where the for- ward points of components other than the fuselage are located relative to the fuselage.

The machine then takes over the task of calculating the areas' intercepted by "Mach planes", of making area distribution plots for each cutting plane, and finally of fairing the area . 010 curves and determining the MOD overall pressure drag. As shown in the right part of .005 figure 11, good correlation EXPER I MENTAL, has been obtained between computed and experimen- ILLUSTRATIVE A I RPLANE CD,W tally determined values for 0 .005 .010 COMPUTED, CD,W a large number of config- urations and for a wide Figure 11.- Zero-lift wave drag.

range of Mach numbers.

Another important application of the high-speed computer has been in the calcula- tion of the optimum wing surface shape for arbitrary planforms and loadings and, con- versely, in the calculation of pressure distributions on arbitrary wings.

(See refs. 22 and 23.) This capability has made practical the previously staggering task of calculating the drag due to lift, camber drag, pitching moments, aerodynamic loadings, and so forth, and has permitted the aerodynamicist to play an effective role in the preliminary design process. It also has cleared up some long-standing mysteries.

For example, although theory has

R 2 - REQ'D BY THEORY E7

long indicated that a properly I \ I REAL WING warped wing with a subsonic leading edge should provide sig- 0 - EXPERIMENT, REAL WING

w

nificantly lower drag due to lift THEORY - than its flat counterpart, the pre- THEORETICAL WING dicted gains were never fully realized in practice. As shown CD in the left-hand plot in figure 12, cD the experimental points fell well below the long-dashed curve cor- LOWER BOUND responding to flat-wing theory but cL still significantly above the solid curve predicted for the warped Figure 12.- Drag due to lift of real and theoretical wings. Design CL = 0.08; A = 700: & = 2.05.

-

I -

wing. The main reason for this discrepancy is that the theoretical model required infi- . nite slopes at the wing root (see left-hand sketch), whereas the model wing necessarily incorporated finite surface slopes in this region (see right-hand sketch). When the new machine computing techniques were utilized to calculate the pressure distribution on the real wing shape rather than on the previously used theoretical model, the discrepancy between theory and experiment essentially w a s eliminated. (See right-hand plot in fig. 12.) Good correlation between theory and experiment now is being obtained for a wide range of wing planforms.

Nacelle shape, size, and location effects.- The analytical tools just discussed have been used by Boeing (ref. 4 ) , NASA (ref. 2 4 ) , and others to study the effects of major nacelle design and installation variables on the increments in wave drag chargeable to nacelle installation. Boeing calculations presented in figure 13 show typical shape and size effects for nacelles placed on an arrow-wing airframe in a near-optimum location.

Wing l i f t coefficient is used as the abscissa of the nacelle wave-drag-increment plot inas- much as interference effects are a sensitive function of the Mach number and sidewash was found to increase significantly the drag beneath the wing. Boattailing of the nacelles due to the nacelles at cruise lift coefficients, as would be expected from consideration of its effect in reducing the local pressures on the aft part of the nacelles and on the adja- cent wing surfaces. Conversely, NASA calculations indicate that flaring the nacelle for- ward of the wing trailing edge has a pronounced favorable effect due mainly to the favor- Thus, the ideal shape for a nacelle would appear able interference effects on the wing.

to be a truncated cone rather than a cylinder as is frequently assumed.

Increasing the nacelle dimensions as much as 20 percent likewise w a s found to reduce the overall pressure drag at and above the usual values of cruise l i f t coefficient.

At first, this result appears inconsistent, but it can be explained readily by the large favorable interference effects. Confirmation of this general trend of reduction in drag ---____

- .OOIOL

I I 0 . I . 2 CL Figure l 3 . - Calculated effects of nacelle shape and size on wave drag. f & = 2.7.

with increase in nacelle size is provided by the NASA experimental data presented in fig- ure 14. The drag coefficients shown are the total external pressure plus skin-friction drag increments measured when four nacelles were added to a basic wing-body configura- tion. The large nacelles had higher drag than the small nacelles in the lower range of lift coefficient; however, the curves cross at a lift coefficient of about 0.1 so that the larger nacelles had lower drag in the cruise and climb ranges. These curves provide a dramatic illustration of the sensitivity of interference effects to the flow conditions beneath the wing. It is emphasized that these results are quantitatively applicable only to arrow-wing configurations with near-optimum nacelle locations. Interference effects would be expected to be different for other wing planforms and other nacelle locations.

.0030r ,-NACELLE DIMENSIONS INCREASED 25% < BASE CONFIGURATION; $=0.0128 ACD, 4 NACELLES \\ -.O03OL

I ~~ I

0 .I .E CL Figure 14.- Experimental effects of nacelle size on overall drag. M, = 2.6.

Caiculated effects of nacelle longitudinal location on the wave drag increment due to installation of nacelles on the arrow wings just discussed are presented in figure 15.

For convenience of interpretation, the wave drag increment is related BOEING to the calculated wave drag incre- ment for four isolated nacelles.

The abscissa of the curve is the distance of the lip of the most for- “D,W = ‘D,W -‘D,W, WINOBODY ward nacelle from the most for- BOEING - t & = 2.7 NASA - M m = 2.6 ward point in the wing notch ‘D.W. ISOLATED expressed in terms of the nacelle NACELLES diameter. The stagger of the

AcD’w 1 [

-I inboard and outboard nacelles and I their spanwise spacing were main- 8 tained constant in the study. At each longitudinal location, the Figure 15.- Calculated effects of nacelle location on wave drag.

basic wing-body was shaped to account for the presence of the nacelles. Hence, even for the most forward location which is the most unfavorable position considered (see left end of solid curve), addition of the nacelles to the wing-body was accomplished with no increase in total wave drag. The effect of moving the nacelles rearward from the most forward location was to provide sig- nificant further decreases in the wave drag. At the optimum location, the maximum addi- tional reduction is indicated to be approximately equal to the calculated total wave drag of fact that interference effects play a the isolated nacelles. These results emphasize the dominant role in determining the aircraft drag and cannot be neglected or ignored in the design process.

Integration of nacelle with wing.- In order to take maximum advantage of favorable interference effects, close attention must be paid to local details of the nacelle and the adjacent wing surfaces. The addition of the nacelles to the wing imposes overlapping regions of positive pressure on the wing lower surface. (See upper sketch in fig. 16.) As discussed in reference 24, the magnitude of these pressures can be estimated by the method of characteristics or the modified linear theory of Whitham. With the lift char- acteristics of the local surface defined, the wing camber surface then can be reflexed locally, as illustrated in the lower sketch in figure 16, to restore the wing load distribu- tion in this region to its design value. The effective result is a reduction, in the reflexed region, of the slopes of both the top and bottom wing surfaces relative to the stream These reductions in surface slope reduce the pressure drag associated with direction.

the reflexed part of the wing and, as shown in the data plot, produce major improvements to the drag polar.

CD REFLEXED

- -

UNREFLEXED O CL /REFLEXED

\. I

LUNREFLEXED SECTION A-A Figure 16.- Use of wing reflex to accommodate nacelle interference. & = 2.6.

It was noted previously that the nacelle drag increment is sensitive to changes in the wing lift coefficient due to accompanying changes in Mach number and sidewash beneath the wing. The drag, therefore, might be expected to be sensitive also to the toe- in angle of the nacelle. Figure 17 presents experimental results obtained by NASA at M , = 2 and CL = 0.16 for an arrow-wing configuration with simulated nacelles located at one nacelle radius below the wing. The experi- LOCAL FLOW mental results are shown to be in good agreement with

- O o 2 0 r I

calculations made in accord- ance with the theory outlined in reference 25. The drag increment of the two nacelles varied about seven airplane drag counts as the nacelle -4 4 8 toe-in angles were varied from -4' to 8O. In accord- A2N ance with the theory of ref- Figure 17.- Optimization of nacelle alinement. ,%& = 2; CL = 0 . 1 6 ; s r;: 0.007.

erence 25, minimum drag occurred when the nacelles were toed in at approximately one-half the local sidewash angle. This toe-in angle usually differs slightly from the toe-in angle necessary for alining the inlet precisely with the local flow, s o a compromise setting o r a bend in the nacelle may be desirable.

Propulsion System Optimization The design of the propulsion system itself is necessarily carried out in close coor- dination with the process of integrating it into the overall configuration. Figure 18 is included to provide an appreciation f o r the relative importance of the major installation considerations. Presented in this figure are range increments calculated for a typical Mach 2.7 transoceanic supersonic transport, with parameters varied in the cruise leg of the mission only. The performance of the airplane in other legs of the mission is affected similarly, of course, by changes in the same parameters, but to a different extent.

Small changes in inlet pressure recovery, propulsion package weight, lift-drag ratio, and nozzle gross-thrust coefficient all are shown to have significant effects on range capabil- ity, with their importance increasing in the order in which they are listed. As a result, propulsion systems designed for supersonic cruise aircraft (see fig. 19) typically incor- porate variable-geometry components and complex secondary-flow systems in order to optimize performance over the wide range of operating conditions encountered. This section of the paper includes a discussion of component characteristics from an installa- tion or systems viewpoint, the problem of optimizing the handling of the major secondary flows during supersonic operation, and afterbody drag.

NOZZLE GROSS-THRUST COEFFICIENT L I D LSION :E WT

--

- INLET PRESSURE

RECOVERY / - 2 w p / / I I L L . 1 I 1 1 I -300 Figure 18.- Performance sensitivity of propulsion-system installation for a typical supersonic transport. M = 2.7 cruise; 3500 n. mi. mission.

BYPASS VALVES THRUST

BYPASS-, f TO EJECTOR/REVERSER

INLET BLEED- Figure 19.- Features o f propulsion systems of supersonic cruise aircraft.

i a

Inlet characteristics.- Inlet pressure recovery continues to be a subject of major interest to the designer. However, as indicated in figure 18, other factors such as the L/D tend in many cases to be more nozzle gross-thrust coefficient and the airframe critical from the overall performance viewpoint. At the same time, inlet weight, starting characteristics, control characteristics, flow stability characteristics, and flow distortion characteristics frequently present more difficult design and developmental problems. A s a result, such factors frequently are assigned an importance equal to or greater than that of pressure recovery in the inlet selection process.

The three basic inlet types of principal interest for airplanes at Mach numbers from 2 . 5 to 3 are illustrated in figure 20. At the present time design applications of all three types sf inlets generally feature fixed cowl lips to minimize cowl weight. Variable throat area and variable supersonic compression geometry are incorporated to accom- modate starting, off-design operation, and control requirements. Shown below the sketches are typical plots of pressure recovery as a function of mass-flow ratio for the design Mach number. For each inlet type, the solid and dashed lines represent the regions of stable and unstable operation, respectively, and the circular symbol indicates a typical design operating point. The portion of the solid line to the left of the symbol is the stability margin either available o r required as the case may be.

The external-compression inlet is the shortest and lightest of the three inlet types.

Inasmuch as it has no internal contraction, it is inherently self-starting and can have a wide stability margin. Unfortunately, its peak pressure recovery is much lower than for the other types and, for the higher Mach number designs, the large external turning requirements necessitate rather large external lip angles with attending high cowl drag.

Hence, it is seldom considered for applications involving protracted operation at Mach numbers above about 2.2.

SIDE SPILLAGE EXTERNAL MIXED (USUALLY COMPRESSION COMPRESSION 2-DIMENSIONAL) STABILITY I MARGlN7T 1 ,,,,~, - pt,o" UNSTABLE// Pt, 0 ) 0 STABLE m - m a Figure 20.- High-performance inlet types.

The mixed-compression inlet affords the potential of very high peak pressure recovery and low external drag at the expense of increased complexity and weight. The terminal shock cannot be stabilized in the contracting region between the cowl lip and the inlet throat. Consequently, since the amount of internal contraction used is large, reduc- tions in mass-flow ratio below the value corresponding to peak pressure recovery cause a drastic drop in pressure recovery and initiate flow instability. Hence, the design oper- ating point must be located 3 to 5 percent below the peak pressure recovery point to insure the avoidance of unstarts due to external disturbances and to allow for the inability of the control system to hold the terminal shock very close to the throat position. In addi- tion, substantial variations in inlet geometry are required to reestablish the design oper- ating conditions should an unstart occur. In brief, this type of inlet affords the highest performance of the three basic types, but at the same time poses the most difficult flow stability and control problems.

Several American companies a r e studying so-called self-starting mixed- compression inlets. In one kind, the amount of external compression is increased at the expense of some increase in external cowl angle and the amount of internal contraction is reduced sufficiently to permit reestablishment of the supersonic flow following an unstart without the necessity for varying the throat geometry. In a second ki,id, the inlet bleed system is designed to provide a recirculation of the bleed flow when the normal shock moves ahead of the inlet throat. The flow enters the surface behind the shock where the static pressure is high and exits from the surface ahead of the shock where the static pressure is lower. Hopefully, this recirculation thickens the boundary layer at the lip station and thins it at the throat thereby reducing the effective internal contraction and permitting automatic restarting. If perfected successfully, either kind of self-starting inlet may provide a good compromise between the relatively tractable external- compression inlet and the higher performance but less easily controlled mixed- compression inlet.

The third basic type of inlet illustrated in figure 20 is essentially a two-dimensional internal-contraction inlet with sidewalls cut back to permit spillage during the starting process. It can be designed to provide high pressure recovery and low drag for the design operating condition, as well as a significant degree of subcritical stability. Its weight and sensitivity to flow angularity are expected to be problems, but these problems may be overbalanced for many applications by its desirable flow stability and restart characteristics.

Distortion and unsteadiness of inlet flow.- One of the major problems of the propulsion-system designer is that of maintaining a low level of distortion and unsteadi- ness of the flow at the engine throughout the operating envelope of the airplane. These phenomena, which are discussed in many references (e.g., refs. 3, 8, 9, and ll), are a complicated function of a very large number of variables, some of which are flow inclina- tion at the inlet, boundary layer of adjacent surfaces, external shock interference, vor- ticity and flow separation associated with subcritical inlet operation, inlet unstart and buzz, and inadequate mixing length in the inlet throat and in the subsonic diffuser. Other sources of distortion and unsteadiness include nonuniformity of air removed or inducted through the duct walls (e.g., through bypass or auxiliary take-off passages), inability of the control system to maintain the inlet terminal shock at a fixed station in the presence of atmospheric, engine, and control transients, and disturbances propagated forward through the engine and nacelle secondary-flow passages.

A detailed discussion of the many sources of flow distortion is beyond the scope of this paper. It is instructive, however, to take a look at the basic mechanism (see ref. 11) Consider in fig- through which distortion.of a steady flow affects engine performance.

ure 2 1 the front rotor-blade row of the engine, where VR is the rotational velocity, VA is the local blade angle of attack. If local is the local airflow axial velocity, and a!

total pressure losses reduce VA, CY increases in the direction of blade stall. On the other hand, if the local flow velocity is increased above the mean value, a! and thus the corresponding local pressure rise through the blading are reduced, with the reductions tending to promote choking further aft in the compressor. The importance of these effects depends on the magnitude and extent of the flow distortions involved as well as on the tolerance of the engine to such distortions. For a nonsensitive engine both small- magnitude radial distortions and small-magnitude circumferential distortions of small angular extent tend to wash out in succeeding blade rows. For greater flow distortion, or for a more sensitive engine, however, the distortion frequently carries completely Apt Pt, av -0. I O

30r /

-0.05 I /

Aa, ' / /

DEG

"1 A

+0.05 FOR WHOLE COMPRES v A

-20 t

-30 I I I

b I 2 3

M a 7 Figure 21.- Problem of allowable flow distortion.

1 1 1 1 1 1 1 1 1 1 I I I 1 1 l I 1 l 1 1 1 1 1 1 1 1 I I I I l l I I l l

through the compressor and produces hot spots in the combustor and turbine. Com- bustor flameout, compressor surge, and even structural damage to the engine may result.

Engine manufacturers frequently express their estimate of the allowable flow dis- tortion for an engine in terms of an area weighted deviation of the total pressure in local regions from the overall average. Effectively this way of specifying distortion tolerance is equivalent to assuming that the total pressure variation is indicative of the flow veloc- ity distribution and, hence, of the angle-of-attack distribution along the front blade row.

It is not always appreciated that this assumption is somewhat misleading. For example, the plot in figure 21 shows local blade angle-of-attack increments calculated for a Mach 3 turbojet for deviations of 5 and 10 percent in the local total pressure from the overall average. Even for 5-percent distortion, the local blade angle-of-attack changes are very large compared with the average available stall margin; also the magnitude of the local angle-of-attack change is a pronounced function of free-stream Mach number. Two con- clusions are evident. First, a distortion parameter is required that is more representa- tive of the actual flow phenomena and, second, the flow-distortion problem tends to become more critical at the higher Mach numbers.

tries to handle the flow distor- At the present time the propulsion-system engineer tion problem by placing the inlet in a favorable location, alining it carefully with the flow, allowing generous flow mixing lengths, conducting extensive model flow distribution tests, and so forth. Almost invariably these efforts are not sufficient and a cut-and-try devel- opment effort is required on the prototype. Extensive research is needed to better under- stand the flow distortion problem, to develop better flow distortion criteria, and to find ways of both improving the uniformity of the flaw delivered by the inlet and of reducing the flow distortion sensitivity of the engine.

Nozzle characteristics.- Calculated performance trades such as those presented in figure 18 show that a high level of nozzle efficiency is mandatory for supersonic cruise aircraft. Most current designs use various forms of the ejector nozzle such as those depicted in figure 22 because a certain amount of engine cooling airflow is always required.

The variable-flap BLOW-IN DOOR WITH converging- diverging nozzle VARIABLE FLAP' BLOW-IN DOOR 7 TERMINAL FAIRINGS 7 currently being utilized on the B-58 and XB-7OA airplanes is the most versatile form inas- much as both the internal and external shroud surfaces can J BLOW-IN DOOR^ FLOATING FLAP be varied to optimize its per- formance. (See refs. 26, 27, Figure 2 2 . - 'High-performance ejector nozzles.

and 2 8 . ) With no geometric restric- tions, the nozzle can be designed to I .o provide 98 to 98.5 percent of the ideal internal gross thrust for flight Mach numbers up to 3 . In an actual design, however, the drag of the external shroud surface must be considered and the use of "practical" linkage arrangements invariably leads to compromise of the internal shroud shape at off-design operating conditions. The net result,

.8 I

I --I illustrated in figure 23 for the case of 0 2 3 typical convergent- divergent ejector M , nozzles designed for a supersonic Figure 23.- Predicted maximum performance of convergent- divergent ejector nozzles designed for a typical supersonic transport, generally is a deterioration transport mission.

of performance in the off-design oper- ating conditions relative to the ideal value. Such losses tend to be large in the low super- sonic range where the thrust minus drag margin is a minimum. Hence, the various trades involved in the design of such a nozzle must be studied carefully. The exercise of design ingenuity frequently can provide important gains in overall mission performance.

22 is currently The "blow-in-door" nozzle portrayed in the middle sketch in figure being applied to a number of new aircraft. Design simplification is achieved and weight reductions obtained by the use of fixed geometry throughout most of the internal and exter- nal shroud surfaces. Blow-in doors, which are closed in high-speed operation but which open to provide increased secondary flow when internal shroud pressures a r e low, and floating flaps, which provide some variation in shroud exit area, are incorporated in the design to provide a level of nozzle performance approaching but obviously not quite equaling that of the variable-flap ejector nozzle. One problem that must be solved in each application of the blow-in-door nozzle is that of locating and proportioning the blow-in doors in such a manner that an acceptable distribution of blow-in-door flow will be obtained under all important flight conditions. Pressure distributions measured on the afterbody of a twin-engine fighter research model (fig. 24) indicate that this is sometimes a difficult task. For the particular model considered, the pressure distributions along different meridians of the afterbody differed significantly because of the presence of the fuselage and the tail surfaces. Further, these pressures changed with changes in flight condition and tail deflection and were everywhere lower than those measured for similar locations on a geometrically similar isolated nacelle. Hence, a major concern of the designer is the fact that operational characteristics estimated on the basis of test results for a simplified model may be misleading.

The nozzle shown in the right-hand sketch in figure 22 incorporates a further possible design simplification being studied currently by NASA.

FX J

Essentially, the aft part of the IATED :ELLE shroud and the floating flaps of the blow-in-door nozzle are

- . I 2 1

replaced by terminal fairings CP ref. 29), which provide a (see ventilated shroud afterbody.

At low internal pressure con-

- -.20 3 F 1 k

. -1

0 .5 I .o ditions, the external flow can x/ I enter the slots and prevent Figure 24.- P.ressure distributions measured on afterbody of twin-engine overexpansion of the internal fighter. Jets operating; M, = 1.2.

flow. Conversely, at high internal pressure conditions, the internal flow can expand outward through the slots and provide increased thrust by reacting against rearward sloping surfaces. Shock- boundary-layer interaction phenomena permit the surface pressures to be increased on the boattail areas ahead of the slots as well as on the terminal fairings only. Informa- tion obtained to date indicates that this type of nozzle can provide good off-design per- formance; however, its design-point performance again would not be expected to equal that of the variable-flap design.

From the performance viewpoint the jet nozzle is by far the most critical installa- tion component. All possible avenues for. further increasing its efficiency must be explored, including the possibility of exploiting favorable external-internal flow interfer- ence effects.

Handling of secondary nacelle airflows.- The basic nature of the inlet-engine air- ~- __ - flow matching process has been treated in so many references (e.g., refs. 3, 11, and 30) that it will not be discussed further herein. However, a typical airflow matching plot for a supersonic transport is shown in figure 25 to illustrate the magnitude of the secondary airflows that must be handled by the installation. From the viewpoint of optimization of the overall installation for high-speed operation, the airflows of principal interest are the inlet spillage flow, the overboard bypass flow, and the bypass flow to the ejector. The inlet bleed flow (see refs. 3, 7, 8, 9, 10, and 11) obviously also is subject to design opti- mization, but is so closely tied to the inlet performance question that it is considered to be a separate subject outside the scope of this paper.

CAPTURE AREA CAPTURE AREA REDUCTION DUE TO WING PRECOMPRESSION 1.0 I NLFT BLEED INLET SPILLAGE C A B I N COOLING,

A *8 - OVERBOARD

- LEAKAGE, ETC.

Alip

.6 -

-

.4 ENG I NE

I I

1.0 2.0 3 . 0 0 M c c l Figure 25.- Typical inlet-engine airflow matching plot.

For a center-body inlet, spillage occurs as the Mach number is reduced below the design value because of the movement of the center-body shock ahead of the cowl lip.

Figure 26 presents a calculation, obtained by the general method described in refer- ence 31, of the amount of and drag of the spillage flow for Mach 3 center-body inlets during supercritical operation, which is the condition necessary for minimizing the spill- age drag. (Also see refs. 3, 8 , 9, and 11.) In this figure, hs/A1ip is the ratio of the stream tube area of the spilled flow to the inlet capture area, and Dspil~/qzAxs is the this spilled flow, based on its stream tube area ahead of the center- drag coefficient of For low-angle center bodies body shock and on the corresponding dynamic pressure ql.

(5O wedge or 1 2 ; ' half-angle cone), the drag coefficient of this spilled flow is seen to be .3

(5)

CRITICAL, WITHOUT . 1 * 2 [ 1 2 TRANS LA T I ON 1/Z0 HALF-ANGLE CONE I I 0- I % J g 4ZA** - a .

-----=-- --

0 ' I I I I 1 2 3 Figure 26.- Drag of inlet spillage flow.

I 1 I 1 I small f o r flight Mach numbers down to 1.5 where 15 to 25 percent of the flow is bypassed around the inlet.

The amount of inlet spillage can be varied for inlet-engine matching purposes within certain limits by translating or changing the contours of the center body.

(Compare

results in fig. 26 for 5 O and loo wedge angles.) Attempts to increase the spillage flow

an excessive amount, however, cause a breakdown of the supersonic entrance flow and thereby result in greatly increased spillage drag. Cowl lip suction effects compensate for only a small amount of this increased drag, except possibly at transonic Mach num- bers. (See refs. 3, 8 , 9, and 11.)

Use of an overboard bypass has been found necessary in nearly all high-Mach- number installations for inlet-engine airflow matching and inlet airflow control. Fig- ure 27 presents calculations o f the momentum-plus-exit drag of the bypass flow of a typical installation based in part on information and data presented in references 32 to 35. The form of the drag coefficient plotted is identical to that used for the inlet- spillage-drag plot in figure 26, so that a direct comparison can be made.

is shown in figure 27 that a complete-expansion bypass nozzle with axial dis- It charge provides low drag throughout the entire Mach number range studied. The draw- backs of this type of bypass are the complexity of the nozzle variable geometry and con- trol requirements and the base or boattail drag that exists in cruise if the bypass flow is shut off or reduced a large amount below the design value. The axial sonic nozzle like- wise can provide good performance up to Mach numbers of about 2, but is significantly less efficient than the axial complete-expansion nozzle at higher speeds and is subject to the same cruise drag problem. When a "practical" flap nozzle of the type illustrated in

-

.5

-

.4 % J p . 3 - qzAxs - 2

-

-

.1 / A X I A L COMPLETE-EXPANSION NOZZLE

--------

n - I " 3 . 0 1.5 2 . 0 2.5 1.0 MZ Figure 27.- Drag of overboard bypass flow.

figure 27 is used, the drag is seen to be significantly higher than that associated with (Compare results in figs. 26 and 27.) Hence, except critical spillage flow by the inlet.

at transonic speeds, it usually is advantageous to maximize inlet critical spillage possibly flow and to restrict the overboard bypass flow to the minimum required for inlet control purposes or off-design operational requirements such as are occasioned by variations in atmospheric temperature, and so forth. The drag of the bypass flow is sufficiently high to warrant consideration of incorporation of special design features, such as rpm trim or high and low flowing, in the engine itself to minimize the amount of flow bypassed. More comprehensive design information on efficient overboard bypass arrangements would facilitate optimization of secondary-flow systems, which require very careful handling to avoid excessive installation drag.

As an alternate to the overboard bypass, air taken in by the inlet in excess of the engine requirement can be bypassed to the ejector so long as its pressure is high enough (See refs. 36 and 37.) As shown in figure 19, to satisfy ejector pumping requirements.

bypass valves usually are incorporated in the ducting to prevent reverse flow at some low-Mach-number operating conditions and to throttle the bypass flow to the desired value when excess upstream pressure is available.

The optimum amount of bypass flow to the ejector is established by engine cooling requirements, the ejector nozzle selected, and the performance trade with overboard bypass. The characteristics of the nozzle selected tend to play a dominant role in the

design trade-off. If a simplified nozzle - for example, a fixed shroud design (refs. 28

and 38) or a design incorporating shroud throat area variation only (refs. 26 and 28) - is

used, moderate or large amounts of secondary flow may yield large performance improve- ments over a wide range of operating conditions. These improvements result from bene- ficial aerodynamic effects on the expansion process of the primary flow. At the other extreme, little i f any performance gain would be expected for a highly sophisticated nozzle designed to provide essentially full expansion of the jet flow under all important operating conditions. Most practical nozzles f a l l between these two extremes.

The effect of secondary flow on the performance of a continuously variable ejector nozzle is difficult to determine accurately because of the large number of geometric vari- ables which must be optimized and because the differences studied are small compared with the total gross thrust. Nevertheless, inasmuch as the designer faces the problem of determining the effect of secondary flow, an attempt has been made to calculate this effect.

A study w a s conducted first to establish a realistic base-line variable-flap ejector for a Mach 3 afterburning turbojet operating over a typical supersonic transport mission profile. Restraints imposed were limitation of the design secondary flow to the minimum amount required for cooling, limitation of the flap lengths to values determined to be acceptable from the weights viewpoint in other studies, and limitation of the maximum nozzle-exit diameter to a value equal to the nacelle diameter. Inasmuch as the overall drag of a properly designed installation has been shown to be relatively insensitive to nacelle size, the basic nacelle diameter chosen was large enough to pass up to 30 per- cent of the secondary flow around the engine without difficulty. The data of reference 27 / then were corrected for temperature and real-gas effects and were used to estimate the I effects on performance of increasing the secondary flow.

Results of the ejector secondary-flow calculations are presented in figure 28 in a force coefficient which sums the effect of the secondary flow on nozzle gross terms of thrust with the ram drag of this flow. This force coefficient is directly comparable to the drag coefficients given in figures 26 and 27 and is plotted with thrust values (negative drag values) below the axis to maintain parallelism of figure format. Maximum attain- w se c

-

, as established by the

able values of the secondary-to-primary flow ratio

( WP )max

estimated secondary-flow pressure and the ejector pumping characteristics, are spotted along the curves.

For the maximum-dry-thrust operating condition, which is of critical importance is little or no drag penalty involved in in the supersonic cruise part of the mission, there increasing the amount of secondary flow handled by the ejector up to secondary-to- primary flow ratios as high as 30 percent. At the same time, however, little performance improvement is possible through increasing this flow because both the inlet spillage flow and the overboard bypass flow are negligible at the cruise Mach number (see fig. 25) unless nonstandard operating conditions are encountered.

For the maximum- augmentation operating condition, which applies to the climb and acceleration phases of the assumed supersonic transport mission, increases in the amount of second- ary flow to the ejector in excess of the minimum cooling requirement of about 4 percent is strongly bene- ficial. It both increases the net 2.01 .05 I thrust and reduces the amount of flow which must be disposed of by

t

inlet spillage or overboard bypass.

Thus, in this particular operating condition, bypassing excess inlet Figure 28.- Drag of bypass flow to continuously variable turbojet capture flow to the ejector provides ejector nozzle.

higher installation efficiency than can be obtained by either increasing the inlet critical flow spillage or increasing the amount of overboard bypass flow. It is unfortunate that the pumping characteristics of this and other typical ejectors prevent the handling of more than about 5 to 10 percent o f the basic engine flow in the range of Mach numbers from 1 to 1.5 where excess flow handling requirements are a maximum. As a result even when both inlet critical flow spillage and the bypass flow to the ejector are maximized, large amounts of excess flow still must be disposed of by overboard bypass, the least efficient method. (See fig. 25.)

A breakdown of the secondary-flow drag of a typical propulsion system designed for a supersonic transport is presented in figure 29. The drag increments shown correspond to the airflow breakdown presented as figure 25 and have been carefully optimized on the basis of the considerations outlined in the preceding discussion. It is indicated that, even when optimized, the secondary-flow-system drag (difference between boundary curves above and below axis) constitutes a signficant fraction of the airplane total drag at tran- sonic and low supersonic speeds for which the airplane thrust-minus-drag margin is a minimum. This fact emphasizes the need for conducting a careful analysis of the secondary-flow drag problem of each new engine installation.

The net secondary-flow-system drag is indicated in figure 29 to be slightly negative in the vicinity of M = 2.2 because of the highly favorable effects of secondary flow on ejector performance. This result is subject to question, with the most probable source of of the basic ejector data used in the calculations. These were e r r o r being the accuracy the best data available. It is obvious that more experimental work is needed to provide a PERCENT A I RPLANE (LESS LIP SUCTION) -6 .I - 1 .~ I 1 I I 0 1 2 M . W Figure 29.- Breakdown of secondary-flw drag of a typical propulsion system.

,Mil I I 1 1 1 1 1 m . 1 , I, 111 I I 111.1 I, I , 1 1 1 I1 1 1 1 1 1 . 1 1 1 . . 1 1 1 . 1 1 1 1 I.I.1 . 111111.1.11 1 1 1 1 1 1 I - L I . l 1111.11111.11111 1 1 1 1 1 1 1 . 1.111 I I I I I I I I I I I I I I I 1 1 1 m m 1 . -, range of reliable large-scale hot-flow ejector data for the use of the propulsion-system designer in making meaningful and accurate trade-off studies.

Afterbody drag.- The new aerodynamic tools discussed in a previous section obvi- ously are applicable to the design of the airplane afterbody. However, this part of the airplane operates in a boundary-layer and flow-field environment established by the air- plane forebody. In addition, it is subject to important airplane arrangement and balance constraints which tend to require the use of low fineness ratios and, occasionally, steep boattails and blunt bases. As a result, the drag of the afterbody tends to be dispropor- tionately high. Figure $0 shows a photograph of a carefully designed twin-engine fighter research model investigated in the Langley 16-foot transonic tunnel. Hydrogen-peroxide rocket motors were used to provide jet simulation. The model was broken at the parting line, with the external surfaces aft of this line including the nacelle afterbodies and the

tails supported on a separate balance. The afterbody drag - that is, the pressure and

viscous drag of all external airplane surfaces aft of the model parting line - amounted to

40 to 45 percent of the overall model drag, although the afterbody portion of the model comprised only one-third of the model length and had only one-third of the total wetted area. In less carefully optimized configurations, the drag of this part of the airplane has Thus, it is clear that been found to be even higher - as much as 55 percent of the total.

the shaping of components and the handling of the associated aerodynamic interferences for the afterbody are of even more critical importance than for the airplane forebody.

An investigation of the boattail-drag characteristics of a single-engine fighter model (ref. 39) affords a quantitative feeling for the impqrtance of afterbody interference .

effects for this familiar airplane con- figuration. Again, the model incor- porated a hydrogen-peroxide rocket motor for simulation of the airplane hot jet exhaust flow. As shown in 31, the afterbody surface shape figure w a s an irregular truncated cone with the vertical and horizontal tails attached directly thereto. Several boattail angles we re investigated.

The pressure drag of the boattail surface was determined for each case by integration of surface pres- sure measurements. A comparison, ‘-SUPPORT STRUT for a typical cruise operating condi- I .o tion, of the results of the experi- mental measurements with estimates L-2614-1 Figure 30.- Afterbody drag of fighter model.

.-. ~~ ... . . ..

made for equivalent conical afterbodies by utilizing the experimental data of ref- BOATTA I L PRESSURE DRAG erences 4 0 to 42 as a basis shows that r ESTIMATED the interference drag w a s several times

& >

as large as the basic boattail drag pre- DE ASURED

dicted without forebody and tail interfer- ence effects considered. The pressure drag of the tail surfaces likewise would be increased by the same interference flows. For this airplane, the boattail

-

drag was reduced to an acceptable level .0040 .0080 by increasing the afterbody fineness ratio B (i.e., reducing the average boattail angle Figure 31.- Afterbody interference drag in cruise for a single- from 15O to 6O). Additional reductions engine fighter model. M, = 0.90; a = @; pt,j/p.. = 3; could be made, if required, by thinning AB,proj = 0.092.

and staggering the tail surfaces.

The pressure drag of an isolated nacelle boattail can be calculated for the super- sonic speeds by linear theory (e.g., see refs. 21, 43, and 44) or characteristics theory.

A wide range of experimental data also exists (e.g., see refs. 2 1 and 45). Nacelle after- body shapes investigated in connection with the previously discussed study of the excess- flow optimization problem for the nacelle of a supersonic transport are shown in the top part of figure 32. The primary design trades considered were the amount of internal nozzle expansion provided and the external flap length. It is noted that the boattail angle

----- 1.00

.006 - -NO AUG.

MAX. AUGMENTATION

-I

+

CD,p .004- I 0 I .o 2.0 3 . 0 M a A m Figure 32.- Boattail drag for nacelle with typical ejector nozzle. s = 0.017.

w a s held at 00 for Mach numbers between 2 and 3 rather than introducing the outward flare required to permit full expansion of the jet flow (light extensions of curves). The reason for this choice was that the flare drag penalty (see lower part of figure) was found to be approximately an even trade-off with the gain in internal thrust obtained by incor- porating the flare; also the low augmentation cruise nozzle setting did not require such a flare to provide essentially full expansion of the jet flow. The longer flap length (Z/d = 1.00) is shown by the drag calculations to provide somewhat lower drag than the shorter flap length (Z/d = 0.75) in the high subsonic and transonic speed ranges. Avail- able flap weight data indicated, however, that this reduction in drag w a s counterbalanced for the supersonic transport mission considered by the associated increased weight.

Therefore, the shorter flap length was used for this application. A similar trade-off study must be made for each new installation. When strong flow interferences exist in the region of the nacelle boattail or when subsonic cruise performance is more important, the longer boattail flap is likely to be preferred.

Base drag.- Fixed external afterbody surfaces are used in many airplanes because of weight and complexity considerations. The XB-7OA (fig. 33) affords an example of this design approach for a multiengine pack-type propulsion installation. It is noted that the afterbody of this airplane incorporates only a moderate amount of boattailing so that large base areas exist around the nozzle shrouds for all operating conditions.

Although definitive base-drag data are not yet available for the airplane itself, pressure measurements obtained for a clustered-exit research model (see lower part of fig. 33 and refs. 46 and 47) indicate that the drag associated with the base areas of the engine pack will be significant at transonic and low supersonic Mach numbers. The *PHOTO OF XB-7OA XB-70A feature of disposing of avail- able low-energy airflows through the base areas is shown in numerous investigations (e.g., refs. 48 to 51) to provide a reduction in base drag. The dashed curve is an estimate, based on these references, of the base pressure for a bleed flaw equal to 3 percent of -.6 CLUSTERED-EXIT JET MODEL DATA the primary engine airflow. It is NO BASE BLEED emphasized that the drag reduction ESTIMATE WITH 3% BASE BLEED CP, b corresponding to the increase in base

-. 2

pressure shown, while significant, *COPYRIGHT, AVIATION WEEK & SPACE TECHNOLOGY 0.5 1.5 2.5 3.5 depends on the availability of large Mal quantities of energy-depleted air; also L-2614-28 the drag reduction attainable with base Figure 33.- Base drag for a multiengine pack-type propulsion installation.

bleed generally is much lower than that attainable with efficient boattailing. It should be noted that this off-design drag problem is not of crucial importance for the XB-7OA which is optimized for cruise at M = 3 and is not forced by sonic-boom considerations to per- form its transonic acceleration at high altitudes. However, an off-design drag increment of the magnitude indicated certainly would not be tolerable for a supersonic transport for which the sonic boom is a critical design factor.

The twin-engine fighter model discussed previously in connection with figures 8 and 9 provides a second example of base-drag problems occasioned by the use of fixed- geometry afterbody surfaces. In subsonic cruise operation the primary nozzles of this particular airplane are closed to the nonafterburning position, with annular bases being formed between the nozzles and the nacelle afterbody surfaces. In addition, airplane base areas exist between the nacelles and the fuselage. (See fig. 34.) Model test results from reference 18 presented in figure 34 indicate that the increment in drag coefficient asso- ciated with these base areas ranges from about 25 to 50 airplane drag counts. This level of base drag is sufficient to limit seriously the range and endurance capability of the air- plane. Obviously, incorporation of a variable-geometry nacelle afterbody surface in the design is one way of minimizing the performance penalty. Results presented in refer- erence 18 show that the base-drag penalty for this type of engine installation also can be reduced by modifying the fuselage shape in accordance with area-rule principles to increase the static pressures and reduce interference flow velocities in the vicinity of jet exits.

the The base-drag characteristics of blunt-based single-engine pods were studied in the course of a transonic wind-tunnel investigation of a four-nacelle research model.

Results obtained for an angle of attack of Oo are presented as figure 35. At high subsonic speeds, the base drag of the two inboard nacelles w a s found to be more than twice that of the two outboard nacelles.

This dramatic difference is traceable directly to the - O 0 6 E I zero-lift area distribution of the model which resulted in the formation of a region of low pressure in the vicinity CD. b .0°41 of the inboard nacelle bases.

-- "Eb -0.018

*O02t- A boattailed nacelle would be expected to experience sim- ilar drag penalties. Thus, I I I I I I L l

. 7 .. .9 1.1

0 '.5 the inboard nacelle location Ivlm studied would be ruled out Figure 34.- Base drag for a bin-engine fighter model with overhanging tail I".

for many present-day applications wherein flight efficiency in low-level dash - -006 operation is a primary , - 2 I N B O A R D A S E consideration.

A s would be inferred 2 O U T B O A R D from the previous discus- - ,002 sion of figures 1 to 3,

angle - of - attack effects

I I 0 ‘.8 .9 I .o 1.1 were found to be favorable M, on the flow-field environ- ment of the afterbodies of A2b Figure 35.- Base drag of blunt-based single-engine pods. a = 00; 7 = 0 . 0 0 7 3 .

the inboard nacelles. When the model angle of attack w a s increased to values representative for high-altitude cruise (40 to 60), the excessive base drag of these nacelles essentially w a s eliminated. In con- sequence, the inboard nacelle position studied is still competitive with other nacelle posi- tions for airplanes not required to execute protracted sea-level dashes.

The results in figures 30 to 35 stress the fact that the afterbody of the propulsion installation constitutes an important source of aircraft drag. In high-performance air- craft a strong effort must be made to keep the effective afterbody fineness ratio of the installation high, to minimize base areas, and to configure the aircraft so as to avoid adverse interference effects.

CONCLUDING REMARKS The present survey indicates that the advance of aircraft speeds into the region of Mach numbers from 2.5 to 3 has greatly increased both the difficulty and the importance of the airframe-engine integration process. External aerodynamic interference consider- ations enter into nearly all phases of the integration problem and have become of almost dominant importance. Recently developed analytical tools permit a rational approach to the treatment of wave-drag interference effects. Other external flow interferences such as shock-boundary-layer interaction effects during inlet unstarts, environmental effects on the performance of ejector nozzles utilizing auxiliary air inlets, and jet effects on aft- located airframe components, are less easily handled. Experimental study is required f o r each new installation.

From the performance viewpoint the jet nozzle is by far the most critical installa- All possible avenues for further increasing its efficiency must be tion component.

explored, including the possibility of exploiting favorable external-internal flow

.* p

k

interference effects. Even with existing concepts, additional experimental work is needed to provide a range of reliable large-scale hot-flow ejector data broad enough to permit the designer to make meaningful and accurate trade-off studies.

Additional research on other installation components also can be of significant benefit. For example, the perfection of a high-performance self-starting inlet would ease greatly the inlet control design problem. More comprehensive design information on efficient overboard bypass exit arrangements likewise would facilitate optimization of the secondary-flow systems. These systems require very careful handling to avoid excessive installation drag.

The flow distortion problem is a subject in its own right that requires careful consideration from both the installation and engine design viewpoints. This problem gets progressively more critical as either flight Mach number or engine performance increases. It is clear that a great deal of research effort will have to be devoted to finding ways of improving the uniformity and steadiness of the flow delivered to the engine and of reducing the sensitivity of the engine to distortion.

Langley Research Center, National Aeronautics and Space Administration, Langley Station, Hampton, Va., June 24, 1966.

REFERENCES 1. Hasel, Lowell E.; and Kouyoumjian, Walter L.: Investigation of Static Pressures and Boundary-Layer Characteristics on the Forward Parts of Nine Fuselages of Vari- ous Cross-Sectional Shapes at M, = 2.01. NACA RM L56113, 1957.

& Sons, 2. Liepmann, H. W.; and Roshko, A.: Elements of Gasdynamics. John Wiley Inc., c.1957.

3. Davis, Wallace F.; and Schemer, Richard Aerodynamic Principles for the Design of Jet-Engine Induction Systems. NACA RM A55F16, 1956.

4. Swan, Walter C.: A Discussion of Selected Aerodynamic Problems on Integration of Propulsion Systems With Airframe on Transport Aircraft. Aerodynamics of Power Plant Installation, Part 1, AGARDograph 103, Oct. 1965, pp. 23-68.

5. Reinhart, W. A.; and Tjonneland, E.: Inlet Flow Field Studies for the Supersonic preprint) 650198, SOC.Automotive Engrs., Apr. 1965. Transport.

6. Gaudet, L.; and Winter, K. G.: Preliminary Measurements of the Flow Field on the Leeside of a Delta Wing of Unit Aspect Ratio at a Mach Number of 2.6 and an Incidence of 150. Tech. Note No. Aero.2787, Brit. R.A.E., Sept. 1961.

7. Stitt, Leonard E.; and Salmi, Reino J.: Performance of a Mach 3.0 External-Internal- Compression Axisymmetric Inlet at Mach Numbers From 2.0 to 3.5. NASA TM X-145, 1960.

8. Faro, Ione D. V.: Supersonic Inlets. AGARDograph 102, May 1965.

9. Mossman, Emmet A.; Pfyl, Frank A.; and Lazzeroni, Frank A.: Fuselage Side

Inlets - A Study of Some Factors Affecting Their Performance and a Comparison

With Nose Inlets. NACA RM A55F29, 1956.

10. Samanich, N. E.; Barnett, D. 0.; and Salmi, R. J.: Effect of External Boundary Layer on Performance of h i s y m m e t r i c Inlet at Mach Numbers of 3.0 and 2.5. NASA TM X-49, 1959.

11. Fabri, J., ed.: Air Intake Problems in Supersonic Propulsion. AGARDograph No. 27, Pergamon Press, 1958.

12. Erdos, John; and Pallone, Adrian: Shock-Boundary Layer Interaction and Flow Separation. RAD-TR-61-23, Res. and Advanced Develop. Div., AVCO Corp., Aug. 15, 1961.

13. Motycka, D. L.; and Murphy, J. B.: Intel-to-Inlet Shock Interference Tests. NASA CR-264, 1965.

14. Moeckel, W. E.: Approximate Method for Predicting Form and Location of Detached Shock Waves Ahead of Plane or Axially Symmetric Bodies. NACA TN 1921, 1949.

15. Moseley, George W.; Peterson, John B., Jr.; and Braslow, Albert L.: An Investiga- tion of Splitter Plates for the Aerodynamic Separation of Twin Inlets at Mach 2.5.

NASA TN D-3385, 1966.

16. Love, Eugene S.; Grigsby, Carl E.; Lee, Louise P.; and Woodling, Mildred J.: Experimental and Theoretical Studies of Axisymmetric Free Jets. NASA TR R-6, RM L54L31 by Love and Grigsby, RM L55J14 by Love, 1959. (Supersedes NACA RM L56G18 by Love, Woodling, and Lee, and T N 4195 by Love and Lee.)

17. Kuchemann, Dietrich; and Weber, Johanna: Aerodynamics of Propulsion. McGraw- Hill Book Co., Inc., 1953.

18. FOSS, Willard E., Jr.; Runckel, Jack F.; and Lee, Edwin E., Jr.: Effects of Boattail Area Contouring and Simulated Turbojet Exhaust on the Loading and Fuselage-Tail Component Drag of a Twin-Engine Fighter-Type Airplane Model. NACA RM L58C04, 1958.

19. Swihart, John M.; and Crabill, Norman L.: Steady Loads Due to Jet Interference on at Transonic Speeds. NACA RM L57D24b7 1957.

Wings, Tails, and Fuselages 20. Lee, Edwin E., Jr.; and Swihart, John M.: Tabulated Pressure Data for a 60' Delta- Wing-Body- Tail Model With a Hot Jet Exhausting From a Pylon-Mounted Nacelle.

NACA RM L57J22, 1958.

21. Harris, Roy V., Jr.: An Analysis and Correlation of Aircraft Wave Drag. NASA TM X-947, 1964.

22. Carlson, Harry W.; and Middleton, Wilbur D.: A Numerical Method for the Design of Camber Surfaces of Supersonic Wings With Arbitrary Planforms. NASA TN D-2341, 1964.

23. Middleton, Wilbur D.; and Carlson, Harry W.: Numerical Method of Estimating and Optimizing Supersonic Aerodynamic Characteristics of Arbitrary Planform Wings.

J. Aircraft, vol. 2, no. 4, July-Aug. 1965, pp. 261-265.

24. Robins, A. Warner; Morris, Ode11 A.; and Harris, Roy V., Jr.: Recent Research Results in the Aerodynamics of Supersonic Vehicles. Paper No. 65-717, Am. Inst.

Aeron. Astronaut., Nov., 1965.

25. Landrum, Emma Jean: Effect of Nacelle Orientation on the Aerodynamic Character- of an Arrow Wing-Body Configuration at Mach Number 2.03. NASA istics TN D-3284, 1966.

Performance Characteristics of 26. Greathouse, William K.; and Beale, William T.: Several Divergent-Shroud Aircraft Ejectors. NACA RM E55G21a7 1955.

27. Stofan, Andrew J.; and Mihaloew, James R.: Performance of a Variable Divergent- Shroud Ejector Nozzle Designed for Flight Mach Numbers up to 3.0. NASA TM X-255, 1961.

28. Beheim, Milton A.: Off-Design Performance of Divergent Ejectors. NACA RM E58G10a, 1958.

29. Swihart, John M.; Norton, Harry T., Jr.; and Schmeer, James W.: Effect of Several Afterbody Modifications Including Terminal Fairings on the Drag of a Single -Engine Fighter Model with Hot-Jet Exhaust. NASA MEMO 10-29-58L, 1958.

30. Wyatt, &Marquis D.: An Analysis of Turbojet-Engine-Inlet Matching. NACA TN 3012, 1953.

Theoretical and Experimental Investigation of Additive Drag.

3 1. Sibulkin, Merwin: NACA Rept. 1187, 1954. (Supersedes NACA RM E51B13.)

32. Hearth, Dona1d.P.; and Connors, James F.: A Performance Analysis of Methods for Handling Excess Inlet Flow at Supersonic Speeds. NACA TN 4270, 1958.

33. Yeager, Richard A.; and Gertsma, Laurence W.: Performance of Several Inlet- Bypass Systems for Matching a Two-Dimensional Variable-Geometry Mach 3.0 at Mach 1.5 and 2.0. NASA MEMO 2-6-593, 1959.

Inlet 34. Vick, Allen R.: An Investigation to Determine the Discharge and Thrust Character- istics of Auxiliary-Air Outlets for a Stream Mach Number of 3.25. NASA TN D-1478, 1962.

35. Connors, James F.; and Wise, George A.: Performance of a Translating-Double- Cone Axisymmetric Inlet With Cowl Bypass at Mach Numbers From 2.0 to 3.5.

NACA RM E57H07b, 1957.

36. Hearth, Donald P.; Englert, Gerald W.; and Kowalski, Kenneth L.: Matching of Auxiliary Inlets to Secondary-Air Requirements of Aircraft Ejector Exhaust Nozzles. NACA RM E55D21, 1955.

37. Hearth, Donald P.: Use of Main-Inlet Bypass to Supply Ejector Exhaust Nozzle at Supersonic Speeds. NACA RM E56K08, 1957.

38. Hardy, Jean-Marie; and Delery, Jean: Possibilites Actuelles d fEtude Theorique d’Une Tuyere Supersonique a Double- Flux. Aerodynamics of Power Plant Instal- lation, Part 2, AGARDograph 103, Oct. 1965, pp. 445-478.

39. Norton, Harry T., Jr.; and Swihart, John M.: Effect of a Hot-Jet Exhaust on Pres- sure Distributions and External Drag of Several Afterbodies on a Single-Engine Airplane Model at Transonic Speeds. NACA RM L57J04, 1958.

I 40. Henry, Beverly Z., Jr.; and Cahn, Maurice S.: Preliminary Results of an Investiga- tion at Transonic Speeds To Determine the Effects of a Heated Propulsive Jet on the Drag Characteristics o f a Related Series of Afterbodies. NACA RM L55A24a, 1955.

41. Cubbage, James M., Jr.: Jet Effects on the Drag of Conical Afterbodies for Mach Numbers of 0.6 to 1.28. NACA RM L57B21, 1957.

42. Silhan, Frank V.; and Cubbage, James M., Jr.: Drag of Conical and Circular-Arc Boattail Afterbodies at Mach Numbers From 0.6 to 1.3. NACA RM L56K22, 1957.

43. Jack, John R.: Theoretical Pressure Distributions and Wave Drags for Conical Boattails. NACA TN 2972, 1953.

44. Fraenkel, L. E.: Curves for Estimating the Wave Drag of Some Bodies of Revolution, Based on Exact and Approximate Theories. C.P. No. 136, British A.R.C., 1953.

45. Cortright, Edgar M., Jr.; and Kochendorfer, Fred D.: Jet Effects on Flow Over Afterbodies in Supersonic Stream. NACA RM E53H25, 1953.

46. Swihart, John M.; and Nelson, William J.: Performance of Multiple Jet-Exit Instal- lations. NACA RM L58E01, 1958.

47. Norton, Harry T., Jr.; FOSS, Willard E., Jr.; and Swihart, John M.: An Investigation of Modified Clustered Jet-Exit Arrangements at Supersonic Speeds. NASA TM X-540, 1961.

48. Norton, Harry T., Jr.; and Keith, Arvid L., Jr.: Effect of Base Bleed and Terminal Fairings on the Performance of Exhaust-Nozzle-Afterbody Combinations at Mach Numbers of 1.93, 2.55, and 3.05. NASA TN D-539, 1960.

49. Cubbage, James M., Jr.: Jet Effects on Base and Afterbody Pressures of a Cylin- drical Afterbody at Transonic Speeds. NACA RM L56C21, 1956.

50. Beheim, Milton A.; Klann, John L.; and Yeager, Richard A.: Jet Effects on Annular Base Pressure and Temperature in a Supersonic Stream. NASA TR R-125, 1962.

51. Beheim, Milton A.: Flow in the Base Region of Axisymmetric and Two-Dimensional Configurations. NASA TR R-77, 1960.

NASA-Langley, 1966 L- 5009 “The aeronautical and space activities of the United States shall be

conducted so as to contribute . . . to the expansion of hzrman knowl-

edge of phenomena in the atmosphere and space. The Administration shall provide for the widest practicable and appropriate dissemination of information concerning its actiuities and the results thereof.” -NATIONAL AERONAUTICS AND SPACE A C T OF 1958

NASA SCIENTIFIC AND TECHNICAL PUBLICATIONS

TECHNICAL REPORTS: Scientific and technical information considered important, complete, and a lasting contribution to existing knowledge.

TECHNICAL NOTES: Information less broad in scope but nevertheless of importance as a contribution to existing knowledge.

TECHNICAL MEMORANDUMS: Information receiving limited distri- of preliminary data, security classification, or other reasons.

bution because CONTRACTOR REPORTS: Technical information generated in con- nection with a NASA contract or grant and released under NASA auspices.

TECHNICAL TRANSLATIONS: Information published in a foreign language considered to merit NASA distribution in English.

TECHNICAL REPRINTS: Information derived from NASA activities and initially published in the form of journal articles.

SPECIAL PUBLICATIONS: Information derived from or of value to NASA activities but not necessarily reporting the results .of individual NASA-programmed scientific efforts. Publications include conference proceedings, monographs, data compilations, handbooks, sourcebooks, and special bibliographies.

Details on the availability o f these publications may be obtained from: SCIENTIFIC AND TECHNICAL INFORMATION DIVISION .

NATIONAL AERONAUTICS AND SPACE ADMINISTRATION Washington, D.C. PO546