Document
' ,d_ f
N88-23741
WIND TUNNEL TESTING OF LOW-DRAG AIRFUILS W. D. Harvey, R. J. McGhee, and C. D. Harris NASA Langley Research Center Hampton, Virginia 23665 ABSTRACT Results are presented for the measured performance recently obtained on several airfoil concepts designed to achieve low drag by maintaining extensive regions of laminar flow without compromising high-lift perform- ance. The wind tunnel results extend from subsonic to transonic speeds and include boundary-layer control through shaping and suction. The research was conducted in the NASA Langley 8-Ft. Transonic Pressure Tunnel (TPT) and Low Turbulence Pressure Tunnel (LTPT) which have been developed for testing such low-drag airfoils. Emphasis is placed on identifying some of the major factors influencing the anticipated performance of low-drag airfoils.
INTRODUCTION Application of laminar flow concepts to aircraft design depends on fabrication, materials, and ease of maintaining laminar flow. The bene- fits of laminar flow are measured by achievement of very low drag which depends on the total wetted surface that is maintained laminar under various flight conditions. Performance at off-design conditions and sur- face maintenance tolerances are also of importance. Successful laminar flow application may cause significant changes in the trend of future aircraft design.
Whereas wing loadings on recent aircraft designs have been increasing, a laminar flow airplane will generally have a lower wing loading than a turbulent one. This effect occurs because of the type of pressure distribution required to yield the insensitivity to surface con- ditions and provides for long runs of laminar flow. Large laminar flow airplanes (transports) will almost surely operate at high altitudes to minimize Reynolds number effects and thus maximize performance.
Considerable basic research and technology, with and without boundary- layer control, is available (refs. 1-12) and believed suitable for design and construction of an aircraft wing to achieve laminar flow with reasonable success at subsonic speeds. Interest in this capability has been renewed by the inflight and wind tunnel test results obtained on several aircraft to establish the existence of natural laminar flow (NLF) on recent production-quality general-aviation airframe surfaces in typical operating environments (refs. 13-15). These results were based primarily on flow visualization (sublimating chemicals) techniques to define transition location and provide increased knowledge and under- standing for present day aircraft. However, many of the wings investi- gated incorporated turbulent airfoil sections and were not designed to achieve laminar flow.
: *_CEDG_G PAGE BLANK NOT FU.NED .... ,La_inariz'ation has proven to be an inherently difficult boundary- layer stability problem to analyze and control due to influences of various local and external disturbances. This difficulty becomes more acute when sweep effects are included at high speeds. For this reason, a good understanding of the various stability theories along with advanced design technology will be required for the development and certification of future high performance aircraft with laminar flow aerodynamics. The emergence of advanced design codes, boundary-layer stability analysis methods, composite materials, and new fabrication technology can substan- tially alleviate previous laminarization concerns and encourage aerody- nm_cists to design better airfoils with higher lift-to-drag ratios.
The Airfoil Aerodynamics _ranch at Langley Research Center is currently involved in utilizing these emerging technologies to develop low-drag airfoils over a wide range of conditions. One such effort is directed toward developing natural laminar flow (NLF) airfoils for general- aviation applications which combine the high maximum lift capability of new NASA high-lift airfoils (refs. 16-18) with the low-drag characteris- tics of the NACA 6-series airfoils. A major design goal of these airfoils is to avoid degradation of high-lift performance characteristics if the flow becomes fully turbulent. Another effort is directed toward research on large-scale swept laminar flow control (LFC) airfoils at transonic speeds to evaluate the compatibility of suction laminarization and supercritical technology at conditions which are typical of high- performance transport aircraft (refs. 6, 19).
The purpose of this paper is to develop a better understanding of the wind tunnel testing environment and its influences on the measured performance of several advanced low-drag airfoil concepts designed to achieve extensive regions of laminar flow. The wind tunnel results extend from subsonic to transonic speeds and include boundary-layer control. The low-speed research was conducted in the Langley Low Turbu- lence Pressure Tunnel (LTPT), and the transonic research was conducted in the NASA Langley 8-Ft. Transonic Pressure Tunnel. These tunnels were developed or modified for testing low-drag airfoils.
SYMBOLS airfoil chord C section profile-drag coefficient c d section lift coefficient c£ C section pitching-moment coefficient at quarter-chord In C pressure coefficient, (p-p_)/q_ P C * pressure coefficient for local sonic velocity P suction coefficient CQ h height LID lift-to-drag-ratio, c£/c d free-stream Mach number static pressure P P rms pressure fluctuation q dynamic pressure R unit Reynolds number R Reynolds number based on chord c Reynolds number based on transition location (Rx) tr t section maximum thickness U_V velocity U rms velocity fluctuations chordwise and spanwise coordinate system x,y O_ angle of attack flap deflection angle, degrees f A leading-edge sweep, degrees wavelength Subscripts: B balance c corrected max maximum suction s tot total tr transition wake w OO free-stream conditions LOW-D_AG CHARACTERIST[CS The drag due to friction on a current transport aircraft at cruise conditions with turbulent boundary layers is approximately 60% of the total drag. Induced drag accounts for most of the balance. The friction drag approaches nearly 90% of the total drag for submersible vehicles.
It is clear, then, that there is room for performance improvements in either case by reducing the drag.
In principle, the most promising approach towards achieving signifi- cant drag reduction is through the stabilization and maintenance of the laminar boundary layer as long as possible such that most of the friction drag remains at the laminar rather than the turbulent level. It is expected (ref. 20) that techniques involving local flow manipulators may soon be available for reduction of the turbulent friction drag of regions of the aircraft that are not laminarized. However, such tech- niques are not anticipated to give drag reduction levels comparable to that of maintaining laminar flow. These techniques will not be discussed herein and only pre-transition concepts are considered.
Past and present wind tunnel research and development and wing-glove flight testing have established pressure and friction as the two major sources of aerodynamic drag. The most effective approach of reducing drag is by geometric shaping (passive) and minimization of wetted area (active), respectively. These approaches have provided a means of main- taining lamillar flow over extensive lengths with subsequent low drag.
Passive Method - Geometric Shaping C(mtrol The passive or natural laminar flow (NLF) approach involves stabil- izing laminar boundary layers by producing a favorable pressure gradient through geometric shaping and requires no active system for control. The exploitation of favorable pressure gradient can be traced back to the development of the NACA 0-series airfoils and sailplane airfoils as well as more recent airfoils developed by Somers (ref. 16) and Viken (ref.
17).
If flow can be maintained laminar over the entire favorable pressure gradient region, it will either undergo transition just beyond the pres- sure minimum or else proceed to laminar separation with subsequent tran- sition to turbulent flow. Which of these flow processes occurs will depend on several factors that include the geometric shape, angle of attack, local Reynolds number, and surface conditions. These combined factors can also produce a hysteresis effect in the lift performance that is often observed for low Reynolds number airfoils (ref. 21). Thus the major objective is to shape the airfoil contour to have as extensive a region of _avorable pressure gradient as possible to ensure laminar flow followed by an appropriate recovery in the adverse pressure gradient region for maintaining attached flow. This becomes more difficult to accomplish the more rearward the favorable pressure gradient is retained. As one approaches transonic speeds, shaping becomes more important in order to minimize pressure peaks in the nose region and
shock formation in the rear adverse pressure gradient (ref. 22). In
addition, inherent instabilities due to boundary-layer crossflow at the
leading edge of swept wings and in the rear pressure rise regions become
very difficult to control passively (refs. 5-8, 23-24).
Active Method - Suction Control
A detailed discussion and summaryreview of a large number of
suction control (LFC) investigations, including both wind tunnels and
flight results, have been presented by Pfenninger (ref. 5). In general,
large reductions in friction and profile drag were achieved with LFC as
comparedwith turbulent flow.
Active approaches usually depend on both shaping and mass transfer
through local suction or blowing concepts. This concept appears to be
the most attractive way of laminarization for low drag, especially when
sweep is required at the higher speeds. Flight experience has shown that
on swept wings the transition location is considerably further forward
than on unswept wings as reported earlier (refs. 7-10) and recently by
Holmes et al. (ref. 13). Earlier transition on swept wings is probably
caused by unstable boundary-layer profiles in the direction normal to the
potential streamlines that create a crossflow in the immediate leading-
edge region and rear pressure rise regions (refs. 6, 23-24). These
crossflow instabilities are less responsive to suction control than
Tollmien-Schlichting instabilities which develop in the streamwise
direction or constant pressure regions. Weakly amplified oblique
Tollmien-Schlichting waves can superimpose on crossflow disturbances
causing distortion of the crossflow vortices that are stretched and
converged downstream. The resulting nonlinear interaction of different
disturbance modeswill cause the less stable crossflow vortices to grow
considerably faster than predicted by linearized stability theory. It is
anticipated that this interaction can be minimized by designing swept
low-drag wings so that crossflow is only critical over a small percentage
of the chord. In the nose region, this may be accomplished by reducing
both the sweep angle and nose radius to acceptable design values. In the
aft region, control of adverse pressure gradient should be the objective.
The boundary-layer development and stability limits of these cross-
flow profiles, as well as the Tollmien-Schlichting instabilities, and the
boundary-layer air which must be removed to stabilize either can be
calculated by numerousavailable theories (refs. 6, 23-27). However,
these methods require arbitrary choice of the growth limitation of the
disturbances or transition location as input to the theory. Thus, these
methods should serve only as a guide in the design process.
Because one of the key elements to the successful achievement of
very low drag with or without boundary-layer control is the question of
surface tolerance, it is important to recognize that no easing of
tolerances is afforded by boundary-layer suction or shaping if both the
speed and unit Reynolds numberincrease (refs. 5, 7, 8, 14, 28, 29). In
attainment of low drag by NLF, success depends on surface shape and
ability to control smoothness. Similarly, for suction surfaces, the boundary-layer stability Reynolds number is held to below limiting values by keeping the boundary layer thin. However, thin boundary layers are inherently developed by increasing Reynolds number and suction and require surfaces with correspondingly smaller roughness and waviness (refs. 5-6, 19).
The turbulent boundary-layer flow over the fuselage of an aircraft can spread from the wing juncture along the attachment line causing contamination. This effect will increase with sweep. Such leading-edge contamination can be avoided by keeping the critical momentum thickness Reynolds number below I00 (ref. 30). This may be accomplished by applying a fence for shielding the inboard turbulent boundary layer from spreading, or reducing sweep angle and leading-edge radius (ref. 5).
Steep pressure gradients due to shock waves can cause separation of the boundary layer and substantial increases in drag. Earlier efforts (ref. 5) and recent in-house analysis and tests (ref. 19) suggest that suction laminarization appears basically feasible in regions of weak shocks at transonic conditions. Apparently, the pressure rise which a laminar boundary layer with suction can sustain in regions of shock interaction decreases with length Reynolds number, unless the upstream boundary-layer thickness is reduced by appropriate suction (ref. 31). In summary, the above discussed effects (sweep, disturbances, contamination, shocks, etc.) impose design challenges to maintaining extensive laminar flow and low drag.
LAMINARIZATION ASPECTS Some of the major factors known to affect transition on low-drag airfoils are surface roughness, waviness, pressure gradient, Reynolds number, suction-induced disturbances, crossflow instability, and wind tunnel or flight environment. A prerequisite for [aminarization is a surface finish compatible with the boundary-layer thickness for which the investigation is undertaken. Three-dimensional surface-induced disturb- ances become primary sources for distortion of growth disturbances in the absence of sweep-induced crossflow effects. However, in comparison with small scale experiments in low turbulence wind tunnels, somewhat increased two- and three-dimensional surface roughness seems permissible in flight (refs. 5, 7, 8, 14, 28, 29). Thus, conclusions from low-drag experiments in wind tunnels often result in misleading and/or unduly pessimistic views about surface roughness or waviness requirements.
Wind tunnel turbulence and noise influence the transition process, and the isolation of these effects requires the total elimination or control of the other known factors (refs. 32-34). The objective of achieving very low wind tunnel disturbance levels approaching anticipated flight simulation levels becomes increasingly difficult as one moves from subsonic to transonic speeds or increases Reynolds number at a given
speed. The characteristic disturbances increase in proportion to the
tunnel speed or pressure level. Thus, the ability to simulate a free air
environment diminishes with the existence and increased level of stream
turbulence, radiated sound from the wall boundary layer or drive system, diffuser flow separation disturbances, mechanical vibrations, etc.
Previous investigations in wind tunnels and flight have clearly shown
that the maximum transition Reynolds numbers obtained with and without
suction (ref. 19) on simple and complex geometries critically depend on
the characteristic disturbance level and broadband frequency present.
Figure I summarizes a large quantity of experimental data from previous
investigations (refs. 19, 34) that show the effect of disturbance level
on transition Reynolds number. The data indicate that low disturbance
levels are required (u/u << 0.1%) for maximum transition Reynolds numbers
in wind tunnels. However, there may be limitations in the ability of
facilities to achieve diminished disturbance levels and scales compared
with flight.
Laminarizing the flow on subscale airfoil models in wind tunnels is
generally a more difficult aerodynamic problem than on full scale wing
surfaces in flight as previously discussed. In particular, the achieve-
ment of moderately high chord Reynolds number simulation on practical
size models in most wind tunnels requires testing at high unit Reynolds
numberswhere characteristic tunnel disturbances dominate, causing early
transition. Laminar separation without reattachment may occur at very
low Reynolds numbers causing difficulty in measuring airfoil performance
(ref. 21). Wind tunnel testing at high unit Reynolds numbers adversely
influences the surface tolerance critieria for both NLF and LFC and will
strongly affect the suction surface and metering system design; physical
dimensions are frequently so small that practical fabrication tolerances
for certain model features becomedifficult to accomplish (refs. 6,
19). Thus, wind tunnel selection is very important for low-drag
testing. The major objective is to be able to test large chord and
aspect ratio models to reduce scale effects and to have good flow
quality.
Establishing the lift performance of low-drag or even turbulent
airfoils is very important. Lift performance can be influenced in wind
tunnels by large adverse pressure gradients induced by the airfoil at
high angles of attack which carl cause sidewall juncture boundary-layer
separation. Obviously, this separated flow can influence the pressure
distribution on the wing and spread across the airfoil span, causing both
loss of laminar flow and lift performance. This influence can be
compoundedby the addition of leading- and trailing-edge high-lift
devices on the airfoil during wind tunnel testing. Thus, consideration
must be given to model aspect ratio and to sidewall boundary-layer
control for high-lift testing.
FACILITIES FORLOW-DRAG TESTING
Low-Turbulence Pressure Tunnel (LTPT)
The Langley Low-Turbulence Pressure Tunnel (LTPT) is a single-return
closed-circuit tunnel which can be operated at pressures from near-vacuum
to i0 atmospheres. The test section is rectangular in shape (3 feet wide
and 7.5 feet in height and length) and the contraction ratio is 17.6:1.
The LTPT is capable of testing at Mach numbers from 0.05 to 0.50 and unit
Reynolds numbers from O.ixlO v to 15x10 6 per foot. This tunnel has provi-
sions for removal of the sidewall boundary layer by meansof a closed-
loop suction system mounted inside the pressure chamber. This system
utilizes slotted vertical sidewalls just ahead of the model test section,
and the removed air is reinjected through an annular slot downstreamof
the test section. A flow control system allows the flow and pressure
requirements to be varied as dictated by tunnel operation. This system
can be used to provide boundary-layer control (BLC) for airfoil research.
A BLCsystem for high-lift airfoil testing is also available. This
system utilizes compresseddry air and involves tangential blowing from
slots located on the sidewall mounting endplates. Flowmeters can be used
to monitor the amount of air blown into the tunnel. An automatically
controlled vent valve is utilized to remove the air injected into the
tunnel by this system. A high-lift model support and force balance
system is provided to handle both single-element and multiple-element
airfoils.
The measured turbulence level of the LTPTis very low due to the
large contraction ratio and tile nine fine-mesh antiturbulence screens.
This excellent flow quality facility is particularly suitable for testing low-drag airfoils. Recent flow quality measurements in the LTPT indicate that the velocity fluctuations in the test section range from 0.025 percent at Mach 0.05 to 0.30 percent at Mach 0.20 at the highest unit Reynolds number (refs. 35, 30).
The drive system is a 2000-horsepower direct-current motor with power supplied from a motor-generator set. The t,Jnnel stagnation temper- ature is controlled by a heat exchanger which provides both heating and cooling using steam injectors and modulated valves that control the flow volume of water through a set of coils. A complete description and cali- bration of the tunnel are reported in reference 37.
8-Ft. Transonic Pressure Tunnel (8-Ft. TPT) The Langley 8-Foot Transonic Pressure Funnel is a closed-circuit single-return variable density continuous-flow wind tunnel with a contraction ratio of 20:[. The test section walls are slotted (5 percent porosity) top and bottom, with solid sidewalls fitted with windows for schlieren flow visualization. In 1981 the facility was modified for flow quality improvements and reconfigured for low-drag testing of a
large-chord swept laminar-flow-control airfoil at transonic speeds
(ref. 19). A honeycomband five screens were permanently installed in
the settling chamber to suppress the turbulence level in the test
section. A contoured liner was installed on all four walls of the test
section to simulate interference-free flow about an infinite yawed wing.
This contoured liner produces a contraction ratio of 25:1 and covers
existing floor and ceiling slots. An adjustable sonic throat is also
located at the end of the test section to block upstream propagation of
diffuser noise.
The combination of honeycomb,screens, and choke provides a very low
disturbance level (p/p _ 0.05%) in the test region at transonic speeds.
Except for the honeycomband screens, the modifications are reversible.
In the current configuration, the stagnation pressure can be varied from
about 0.25 to 1.25 atmospheres up to a Mach number of less than 0.85 with
the transonic slots closed by the liner. The stagnation temperature is
controlled by a water-cooled radiator upstream of the settling chamber.
Tunnel air can be dried by a dryer using silica gel desiccant to prevent
fogging due to expansion in the high-speed nozzle.
RESULTS ANDDISCUSSION
Past and present wind tunnel and in-flight testing has shown that
the maintenance of extensive regions of laminar flow by the use of NLF or
LFCapproaches can provide significant drag reduction for improved air-
craft performance. The following discussion is a review of recent wind
tunnel tests of advanced design concepts (refs. 17, 19, 38) for low-drag
airfoils and someof the effects that the wind tunnel environment has on
those results. Also, it is intended to identify influences that are
known to affect performance results obtained in wind tunnels that, if not
taken into account, can cause concern and rejection of such low-drag air-
foils for future application. Several of the 2-D/3-D designed airfoil
configurations (refs. 17, 38, 39) shown herein were discussed in detail
earlier in this workshop along with discussions of integrated trailing-
edge flap designs (ref. 40) for the medium- to high-speed NLF airfoil
designs. Thus, no detailed discussion of design concepts will be
presented here, only background information, experimental verification, and factors that influence overall results.
Tunnel Flow Quality
Aside from other factors known to affect transition on low-drag
airfoils, the maximum transition Reynolds number, with or without
boundary-layer control, critically depends on the characteristic disturb-
ance level and broadband frequencies generated in wind tunnels utilized
for testing (fig. l). An example of this effect is shown in figure 2 for
several airfoils recently tested in the two NASA Langley wind tunnels
developed and used for low drag research. Both facilities operate above
and below atmospheric pressure, providing a wide range of Reynolds
numbers and Mach numbers from 0.05 < M_ < 0.85 as currently configured.
The present flow quality values were measuredwith conventional hot-wire
and acoustic probe techniques (ref. 33). Transition location on the
airfoils was measuredusing surface thin film gages (ref. 41) and is a
routine requirement for assessmentof laminar patterns or state of the
local boundary layer.
For either the NLF or LFC airfoil results in figure 2, the measured
logarithm of transition-length Reynolds numbervaries inversely in
proportion to the logarithm of the tunnel disturbance level. As
expected, the indicated levels for (Rx) tr with LFC applications are
significantly higher than those without suction control as can be seen by
comparing past (ref. 12) and present results obtained in the LTPT
(M _ 0.20). It should be noted that the extensive lengths of laminar
flow measuredon the new NASA airfoils (figure 2) generally agree with
expectations and that this achievement may in part be attributed to
excellent flow quality. Thus, the achievement of high transition
Reynolds numbers for low-drag testing may not be possible if acceptable
flow quality cannot be realized in test facilities. The selection of
suitable facilities would, of course, imply the need for measuredand
documentedflow quality for assessment.
Low-SpeedAirfoils - Surface Tolerances
Figure 3 illustrates the airfoil shape and near design velocity
distribution over both surfaces and represents a concept developed by
industry for long endurance operation requiring high L/D. This configur-
ation was shaped to provide a velocity profile or favorable pressure
gradient suitable for maintaining lacunar flow back to x/c _ 0.30 on the
upper and x/c " 0.75 on the lower surfaces at a chord Reynolds number of
14x10 b with zero sweep. This geometry type and velocity distribution are
not entirely unfamiliar to today's aerodynamicists in that they resemble
those which may be found on sailplanes, low-speed aircraft, and business
jets that have utilized NLF for drag reduction and improved performances.
A model of the long endurance airfoil concept was constructed of
metal with a 2.7 foot chord and aspect ratio near 1 and instrumented with
pressure orifices. The photograph in figure 4 shows the model removed
from the tunnel and is a view of the underside leading-edge region illus-
trating the removable metal cover plate located at near mid-span for
access to internal instrumentation and leads. The model was initially
tested as received with only minor cleaning of the surface with diluted
alcohol.
Wake-rake drag measurementswere obtained in the LTPTat one model
chord length downstreamof the trailing edge at spanwise stations of
y/c = 0 (mid-span) and y/c = 0.325 (10.4 inches from mid-span). The
measured surface pressures were integrated to obtain airfoil section lift
coefficients. Performance of the low endurance airfoil is summarizedin
figure 5 for [_ = 0.I and R c = 14xi0 6 with and witilOUt fixed transi-
tion. The results clearly indicate that the drag _evels obtained
at _d-span (y/c = 0.0) with free transition were extremely high, indica-
tive of only small lengths of laminar flow. This result is supported by a comparison of the drag levels obtained at mid-span with free or fixed transition on the upper and lower surfaces at x = 0.03c.
Visual inspection of the model prior to this initial test revealed that the lower surface may have had adverse roughness effects due to the model cover plate. Therefore, the model was sanded on both surfaces with number 600 carborundum paper and thoroughly cleaned with diluted alcohol in an effort to eliminate the suspected cause for loss of laminar flow.
Further precaution was taken by also lightly wiping the surface with a special cloth (tack rag) to remove lint and dust settlement on the hori- zontally mounted model. Drag measurements were then made at the spanwise station y/c = 0.325 to minimize possible lower-surface cover-plate disturbance effects. Figure 5 shows that a drag coefficient of about 0.0055 was measured at a lift coefficient of about 1.2 (L/D = 218), signifying a large gain in the extent of laminar flow for the improved surface conditions. Upon completion of the test, surface waviness measurements were made at both spanwise stations using a surface dial indicator with fixed legs on a solid base spaced 2 inches apart. The resolution of the dial indicator was determined to be 0.0005 inches. The measured waviness on the long endurance airfoil indicated possible exces- sive waviness at the mid-span station on both surfaces that would be unacceptable for wind tunnel models and low-drag tests. For example, several waves with height-to-wavelength ratio of h/X = 0.003 were measured on both surfaces near x/c = 0.15.
Previous research by Carmichael (refs. 28, 29) on low-drag airfoils with and without sweep has provided an empirical expression that repre- sents local allowable waviness for single waves. Carmichael further suggested that one could estimate tolerances for closely spaced multiple waves by multiplying the single wave expression by a factor of I/3.
Since multiple waves were present on the long endurance model surface, the measured h/% = 0.003 for single waves was reduced by this factor and compared to Carmichael's empirical expression for several Reynolds numbers and constant sweep angle of 30 ° (figure 6). Carmichael has shown that only a small reduction in allowable multiple waviness exists for a swept wing compared to unswept wings at low speeds. Also, included in figure 6 are measured values of allowable waviness obtained for other low-drag airfoil models tested in the NASA Langley 8-Ft. TPT and LTPT and subsequently discussed herein. The present results clearly indicate the need for tight control of fabrication tolerances on low-drag wind tunnel models due to scale effects. These tolerances, however, may be relaxed for full scale aircraft surfaces, as suggested in a recent review by Holmes et al. (ref. 14), since scale effects are greatly reduced both _ wing chord and low unit Reynolds numbers at cruise conditions (R/ft < _i0_).
The extent of laminar fl_w at mid-span oD the long endurance air- foil model at R c = 14x10 _ (R/ft = 5.2xi0 b) was probably influenced by increased tunnel turbulence level associated with high unit Reynolds number, in addition to sensitivity of surface conditions.
Low-SpeedAirfoils - Shaping
Recent experimental performance results have been obtained in the
LTPT for a low-speed NLF airfoil (ref. 18), designated NLF(1)-O414F, and details of the design features are given by Viken (ref. 38). In general, the design objective for this airfoil was to obtain very low cruise drag coefficients by selective shaping of the contour to provide a favorable pressure gradient with extensive laminar flow regions over both upper and lower surfaces and high maximum lift. The design conditions were cg = 0.43, R c = 10x106, M® < 0.40, and thickness ratio of t/c = 0.14. Figure 7 illustrates the calculated design pressure distri- bution and airfoil section shape. A simple trailing-edge flap having a length equal to 0.125c was incorporated to substantially increase the low-drag cg range. As can be seen from the design favorable pressure gradient (figure 7), laminar flow is anticipated over both upper and lower surfaces rearward to x/c = 0.70. Furthermore, design considera- tions were given to the achievement of gentle stall characteristics and to maintaining an acceptable lift performance if the airfoil becomes turbulent.
Representative airfoil section data obtained with wake rake and sur- face pressures are presented in figure 8 for the design Reynolds number with and without fixed transition in the indicated nose regions. For the free transition case, laminar flow was measured on both surfaces back to x/c = 0.70 and will be subsequently shown and discussed. Figure 8 shows that a minimum drag coefficient of about 0.0027 was measured at the design lift of about 0.40. This corresponds to L/D = 160 with zero flap deflection. Furthermore, a value of [.8 for the maximum lift coefficient was obtained at an angle of attack of [8 ° while the pitching moment remained relatively constant. However, of major significance is the fact that fixing transition near the leading edge had only a very small effect on the lift performance and (C£)ma x value at the expense of drag increase. This finding is believed to be a very important improvement over the previous NACA b-series airfoils which have adverse stall charac- teristics. In other words, this new NLF airfoil design can also be classified as a very good turbulent flow airfoil in terms of lift performance as well as drag level and pitching moment.
Transition location on the airfoil upper surface was determined by using small thin-film gages that were glued to the model surface at several chordwise and spanwise locations and spaced to eliminate inter- ference effects from one another. These instruments basically operate on the same principle as hot-wire anemometry with overheat ratio set for the sensitivity required for the detection of the state of the local boundary layer where they are placed (ref. 41). This is accomplished by utilizing characteristic behavior of the gages for detection of local changes in heating due to shear stresses of either a laminar, transitional, or tur- bulent boundary layer. These local changes are recorded as variation of rms output signals with time and require a sufficient number of these gages to be spatially located on the surface to properly identify patterns.
It is essential that each thin-film gage experience a known laminar, transitional, and turbulent flow output signal for a given investigation in order to reference and properly interpret results. This may either be accomplished by starting at sufficiently low Reynolds number test condi- tions where the output signals for all gages are known to be laminar or by locating a reference gage where it always senses a known turbulent or laminar boundary layer for comparison. Caution should be used when these conventional type gages are located in separated flow zones as to inter- pretation of results. The output signals in such a zone may indicate similar signals to those for turbulent attached flow.
Figure 9 shows example results taken by J. P. Stack (NASA Langley) from surface mounted thin-film gages on an NLF airfoil model in the LTPT. The gage located at x/c = 0.40 and in a known laminar flow region indicates a time-dependent low-level rms output signal with essentially no deviations above or below the mean. As the laminar boundary layer approaches its stability limit, laminar-to-turbulent bursts are locally detected with elapsed time as indicated for x/c = 0.5 location. The flow becomes progressively unstable downstream (or with increased Reynolds number) until peak transition occurs with a higher rms level and very random signal with time as seen for x/c = 0.6. Once the flow goes through transition to fully turbulent flow (x/c = 0.7), the output signal remains high, but the deviations above and below the mean become more consistent. From these type signals obtained over a series of test conditions, the extent of laminar flow could be determined. Figure i0 shows the measured upper surface transition location on the NLF(1)-0414F airfoil with lift coefficient for constant Rc's. The results confirm that the existence of laminar flow was maintained rearward to x/c = 0.70 at design c_ = 0.40 and R c = 10El0 6. The corresponding wake-rake drag measurements (fig. 8) with free transition support the thin-film results. However, as the lift coefficient is increased above design c_, transition gradually moves forward, and at c£ = 0.50, (x/C)tr = 0.50.
The successful verification of this airfoil's performance is attributed to holding very tight surface tolerances during fabrication and obtaining test results in a wind tunnel with good flow quality. The fabricated and measured surface waviness was held to h/% = 1/3000 for single waves and is shown in figure 6 (h/_ = 0.00033 at _/c = 0.056) to be well below other data and that allowable for multiple waves at design R c = lOxlO 6.
The measured drag variation with lift for different fla_ deflections from -I0 ° < 6f < 20 ° is summarized in figure ii at R c = 6x10 V and M_ = 0.07. The results indicate that very low drag values can be maintained over a lift coefficient range from U < c£ < 1.0. These results were obtained with a simple flap of 0.125c length and offer the potential for long runs of laminar flow over a wide c% range. Because the NLF(1)-0414F airfoil was shaped (fig. 7) for long regions of accelerated flow necessary to achieve laminarization followed by a rather steep pressure gradient downstream, a laminar separation bubble was anticipated beyond the pressure minimum at low Reynolds numbers. Such bubbles have inherently unstable characteristics that generally cause transition to rapidly move forward with significant lift losses. The existence of a laminar separation bubble on either surface was detected by comparison of measured and predicted pressure distributions (not shown) and associated drag level increase from lift-drag polars. These comparisons clearly indicated that a laminar bubble which existed aft of x/c = 0.70 for R c < 3xi0 was nonexistent for R c > 4x10 v.
It is well known that a significant reduction in drag can be realized if flow can be kept attached. One method for reducing the drag associated with tile presence of a laminar bubble is to force boundary- layer transition to occur ahead of the bubble causing the flow to remain attached (refs. 42-44). This can be accomplished by the use of turbula- tors (2D-3D trips, spoilers, sound, passive or active blowing and suction, etc.). Of course, one must account for the 'device drag" of the turbulator used. Tests were conducted to evaluate the effectiveness of several turbulators on the NLF(1)-O414F airfoil in the LTPT. Figure 12 illustrates the effect of using simple 2D strips of commercial tape placed at x = 0.68c on both upper and lower surfaces of the m_del to force transition and eliminate separation bubbles (R c = 3x10 v, H_ < 0.2, and 6f = 0°). A tape 0.012 inch thick and 0.25 inch wide was used. The results (figure 12) clearly show a measured reduction in drag coefficient of about 0.0010 at c£ = 0.40 with the turbulator tape. Apparently, only a small amount of induced energy by turbulators is required to force transition and attachment of laminar separation bubbles.
The effectiveness of turbulators strongly depends both on their geometry and location since they function like trips or roughness which scale with local boundary-layer properties. Other turbulator devices (not shown) were tested aside from the tape and found to be effective.
For example, small vortex generators of h/c = 0.25, and 15 ° leading-edge sweep with respect to flow direction, were spaced (Ay/h = 8) along the model span at x/c = O.60 and x/c = 0.70. Results from using these devices proved very effective in forcing transition and bubble attachment but produced undesirable drag penalty. Probably the most effective turbulator, in terms of both forcing laminar bubble attachment and reducing wake drag with no apparent device drag, was a spanwise row of holes (ref. 42) ahead of x/c = 0.70, The holes of diameter d = 0.01)18 inches, located 0.25 inches apart, were drilled through from the top to bottom surface. Since the design pressure distribution (figure 7) gener- ated a pressure differential across the upper and lower surfaces, passive suction and blowing occur, providing a method for energizing the boundary layer that was sufficient to be effective in these tests. In summary, such devices and techniques as described above appear very promising and economically feasible for application and control of laminar separation bubble attachment with subsequent drag reduction.
High-Speed Airfoils - Shaping It is well known that the subsonic cruise speeds of high performance aircraft are limited by the onset of the transonic drag rise and that the use of wing sweepback delays this onset (ref. 22). Another method for increasing the cruise Mach number is through the use of geometric shaping
which delays the drag rise Mach number. The first airfoils developed in
the U.S.A. to delay drag rise were the NACAl-series (ref. 45). These
airfoils were designed to delay the Mach numbers at which supersonic flow
first develops locally on the airfoil. These airfoils have significantly
higher drag-rise Mach numbers than the earlier NACAfour-digit series;
however, the low-speed high-lift characteristics are much poorer than
those of the earlier airfoils. The NACA 6-series airfoils also provided
increased critical speeds with improved drag-rise characteristics com-
pared to the four-digit series but also have small degraded low-speed
characteristics. Such airfoils or their derivatives have been used on
many first generation subsonic jet aircraft.
The first airfoils designed to purposely delay drag rise by
improving the supercritical flow over the upper surface were the "peaky"
airfoils. These airfoil shapes generate an isentropic recompression of the supersonic flow on the forward airfoil region and provide some delay in drag rise but also have degraded low-speed characteristics compared to the NACA o-series airfoils. Whitcomb's research efforts (ref. 22) led to designs which allowed the recompression to move far rearward on the air- foil at transonic speeds and resulted in significant delays in drag-rise Mach number without degrading low-speed characteristics.
Based on the encouraging results obtained by geometric shaping to achieve extensive laminar flow on both surfaces of the low-speed NLF(1)-O414F airfoil, effort has been recently directed towards extension of the concepts to higher speed NLF airfoils. Details of the two-dimensional design concepts have been given by Viken (ref. 38) alon_ with wing body integration by Waggoner (ref. 39) and integrated trailing-edge flap design by Morgan (ref. 40). One of the more promising high-speed NLF airfoil concepts has been fabricated and tested in the NASA Langley LTPT and 6x28-inch transonic tunnel (TT) complex to investigate its low-speed high-lift and drag-rise characteristics. This NASA high-speed natural laminar flow airfoil is designated HSNLF(I)-0213. The airfoil was designed for a lift coefficient of 0.25, Mach number of 0.70, chord Reynolds number of R c = 10xl06, and t/c = 0.134. This particular design was for essentially zero sweep.
The HSNLF(1)-0213 airfoil design pressure distribution and section shape are shown in figure 13. Geometric shaping was expected to provide laminar flow rearward to x/c = _.55 on the upper and x/c = 0.70 on the lower surfaces up to R c < lOxl0 . In general, the bottom side of the nose was slightly modified from the NLF(1)-O414F to minimize off-design pressure peaks on the lower surface, and upper surface aft camber was reduced to minimize the possibility of turbulent separation.
Results obtained (not shown) from tests in the 6x28-inch TT indicated good agreement between measured and predicted pressure profiles and that drag ris_ occurred at about M_ = 0.72 for c£ " 0.2, and design Rc = lOxlO_. The level before drag rise, with and without fixed transition at x/c = 0.05, was about 20 to 30 percent below that of a good turbulent airfoil. It should be noted that the turbulence and noise levels are believed to be high in the 6x28-inch TT which operates as a I03
blowdown facility. Thus, poor flow quality contributed to the inability
to achieve extensive laminar flow over the model, especially at the
higher Keynolds numbers. However, the drag rise characteristics can
still be approximated. Surface contour accuracy was measured for the
6-inch chord steel model fabricated and found to be acceptable.
A second HSNLF(;)-0213 airfoil model, fabricated with fiberglass
external surfaces, had a 2-foot chord and no flap. This model was tested
in the LTPTfor low-speed performance evaluation. An example of the
results is shown in figure 14 for Moo = 0.108 and R c = 4x10 _ with and
without fixed transition at x/c = 0.05. Figure 14 shows that a minimum
drag coefficient of about 0.0038 was measuredat c£ = I).2 or L/D _ 53.
The results indicate that the airfoil displayed trailing-edge-type stall
characteristics, and a value of (C%)ma x _ 1.55 was obtained for
R c = 4xll) b. It is also apparent that fixing transition had only small
effects on the lift performance. Figure 15 shows the effect of Mach
number and Reynolds numberon the maximum lift performance. While the
(C£)ma x increases with increasing R c for constant Machnumber as
expected, there is a small effect of Mach numberon the maximum lift.
For example, results at z_ = 0.i are consistently 0.05 higher than
results at _ = 0.2 over the Reynolds number range tested. Thus, one cannot simulate Reynolds number effects on (C%)ma x by increasing Hach number at the same time Reynolds number is increased _ince they have opposing influences.
Boundary-layer transition locations were also obtained by J. P.
Stack (NASA Langley) on both surfaces of the HSNLF(1)-0213 airfoil in the LTPT using surface-mounted thin-film gages. A summary of the transition locations on the upper and lower surfaces compared with predictions from the Eppler theory (refs. 46-47) is shown _n figure l_ for _ = 0 ° and chord Reynolds number from 3.0xI06 to 9x10-. The data clearly indicate that lamiuar flow was maintained rearward to about x/c _ 0.5 and x/c _ 0.70 on the upper and lower surface, respectiw_]y, up to R c = 8xi0 b, before any forward movement of transition was measured.
Low Reynolds Number Airfoils For airfoils designed to operate at low Reynolds numbers (R e < 500xlOB), the existence of a laminar separation bubble and turbulent separation significantly increase the drag and decrease the lift, both of which contribute to low lift-to-drag ratios. This phenome- non has previously been extensively investigated and discussed (refs. 21, 48-51). Increasing the Reynolds number will reduce the length of the laminar separation bubble and extent of turbulent separation. At posi- tive incidence, the boundary layer, which is laminar along the airfoil's upper-forward surface, separates at the downstream adverse pressure- gradient recovery region. It then quickly undergoes transition to turbulent flow in the separated shear layer.
Depending on the Reynolds number and the severity of the adverse
pressure gradient, this separated turbulent boundary layer may or may not
reattach to the airfoil surface. If reattachment occurs, the turbulent
boundary layer may then separate again near the trailing edge. If the
Reynolds number is sufficiently low such that reattachment does not
occur, increasing the Reynolds number to somecritical value will cause
reattachment that corresponds to a dramatic increase in L/D. Thus, the
possible existence of both laminar and turbulent separation should be
considered in the design and wind tunnel testing of airfoils in the low
Reynolds number regime. For example, such airfoils are typical of those
on current RPVs, sailplanes, and general aviation aircraft canards.
Mangalamand Pfenninger (ref. 52) have recently designed a low
Reynolds numberairfoil and tested it in the 12"x18°' open-circuit tunnel
at the NASA Langley 8-Ft. TPT complex at low speeds. The airfoil section
shape and an example of the measuredan_ predicted pressure distributions
are shown in figure 17 for R c _ lOOxl0_, c£ = 1.0, and _ = 4°.
Basically, the airfoil was shaped to have moderate negative camber in the
nose region to reduce pressure peaks at off-design and attached flow.
The forward lower surface cusp is due to combined leading-edge thickness
and camber and requirement for increased mid-chord thickness for struc-
tural strength. This airfoil is designated LRN(1)-I007 and represents
about a 40%increase in t/c and an appreciable increase in c£/c d at
design (fig. 18) above previous similar airfoils (ref. 51).
Except for the upper surface aft region, good agreement is shown
between the measuredand predicted pressures using the Eppler theory
(ref. 46) for the smooth model. The measureddata indicate a long
separation bubble in the rear upper-surface pressure-rise region, with
reattachment near the trailing edge. Mangalamand Pfenninger (ref. 52)
concluded from these results that at low Reynolds number the laminar
boundary layer is highly stable and a number of trips are required in
several locations along the chord to promote transition. Applications of
the 2-D spanwise trips (ref. 52) eliminated the laminar separation bubble
and provided about a 25%increase in lift-drag ratio. Subsequent flow
visualization photographs were obtained of the model flow field in the
same tunnel, using smokewire techniques*, and are shown in figure 19
illustrating tile occurrence of laminar separation with incidence angle
for R c = 40,000. Figure 19 shows attached flow over most of the air-
foil surface at _ = 3 ° . However, for _ = 18 ° , separation occurs at the
leading edge and never reattaches. Similarly, for a = -12 ° , the lower surface separates without reattachment. It can be seen from these photo- graphs that the measurement and verification of the performance of low Reynolds number airfoils with separated flow become highly questionable.
*Smoke wire technique was developed and results obtained by Amir Bar-Sever and Dr. S. Mangalam, under contract to AAB, NASA Langley.
High-Lift Testing - Sidewall Effects Depending upon model aspect ratio and facility utilized, testing of airfoils at high angles of attack can result in severe tunnel sidewall interference effects. Large adverse pressure gradients can be induced by the airfoil at incidence that cause the oncoming tunnel sidewall boundary layer to separate, spread downstream and spanwise, and result in a large decrease in airfoil lift. The following results and discussion attempt to illustrate this influence on high-lift performance for low-drag or turbulent airfoils.
Figure 20 shows lift performance results for the same single-element airfoil teste R in two different NASA Langley facilities at M= = 0.30 and R c _ 6x10-. The models tested in the LTPT and 6x28-inch TT had aspect ratios of 1.5 and 1.0, respectively. The results indicate that severe sidewall interference effects occur in the 6x28-inch TT for angles of attack greater than about i0 °. This resulted in measured maximum lift coefficient for this airfoil in the 6x28-inch TT that was about 17 percent lower than in the LTPT. Figure 21 shows photographs of oil flow patterns obtained in the 6x28-inch TT with the airfoil having an aspect ratio of 1.0 for 8 ° < a < 12 ° . A complex secondary flow field that nearly dominates the entire model span is seen to develop due to sidewall interference as model incidence is increased. Separated flow occurs on either side of mid-span for _ = 12 ° causing drastic lift loss (fig. 20).
To ensure that models experience uniform, two-dimensional, interference- free air flow when testing multi-element airfoils, some feasible concept is required for tunnel sidewall boundary-layer control (BLC).
The LTPT has recently been modified to incorporate a BLC system (ref. 37) which includes both upstream sidewall suction slots ahead of the model and tangential blowing slots located on the same tunnel walls near the model juncture region. This system provides a means for reducing the oncoming boundary-layer thickness as well as energizing the boundary layer locally around the model for maintaining attached flow as incidence is increased.
The effects of sidewall BLC on the lift performance of a multi- element airfoil using tangential slot blowing near the nose and flap regions are shown in figure 22. Slot blowing is applied until the lift at mid-span (obtained from surface pressure data) is approximately matched with the lift near the tunnel sidewalls. The results shown in figure 22 indicate large differences in lift coefficient at high inci- dence between the mid-span and near wall regions without blowing compared to the results with blowing. With blowing, the measured lift values at mid-span and near wall regions are essentially the same. Thus, one should use caution in conducting high-lift performance testing
or verification to select a facility that accommodateslarge-aspect-ratio
models or that has sidewall control for reducing interference effects to
insure meaningful results.
In addition to the interference effects produced by separation of
the boundary layer on the vertical sidewalls, corrections are required to
account for model and wake blockage and interference due to test section
floor and ceiling constraints on streamline curvature. Tests are planned
in the LTPTto evaluate these interference effects by testing high-lift
models with chord lengths of I and 2 feet. Lift coefficients up to about
4.0 are expected from these models.
High-Speed Airfoils - Shaping and Suction
The concept of combining geometric shaping and suction laminariza-
tion on airfoils to achieve very low drag dates back to the late 1930's
(refs. 4-5). The basic feasibility of achieving full chord laminar flow
with very low drag on swept nonsupercritical LFC wings was pioneered by
Pfenninger (ref. 5) with suction applied through many closely spaced
surface slots on the wings. Results were obtained on large chord wing
sections (modified 66012) of 30° sweep and t/c _ 0.12 in three different
wind tunnels. These studies confirmed earlier beliefs that results were
dependent on the characteristic turbulent and acoustic disturbance levels
in each facility. Since this research demonstrated the potential for
significant drag reduction through application of LFC, an interest in
evaluating the feasibility of combined suction laminarization and super-
critical airfoil technology at conditions which are typical of high-
performance transports has been generated. Therefore, a large chord
(c = 7.07 ft.) swept supercritical LFC airfoil with suction slots has
been designed, constructed, and recently tested in the NASA Langley
8-Ft. TPT. This NASA airfoil is designated SCLFC(1)-O513F. Details of
the airfoil and suction system design along with the test setup have been
reported (ref. 19). Requirements for this test also included modifica-
tion of the wind tunnel to achieve the desired flow quality and test sec-
tion wall contouring to simulate free air flow about an infinitely yawed
model at transonic speeds.
Figure 23 shows the design pressure distribution for the swept
supercritical LFC airfoil. Attempts were madeto minimize suction lami-
narization through a highly tailored pressure distribution and choices of
leading-edge sweep, chord Reynolds number, and crossflow Reynolds number
(ref. 6). Depending on geometry, boundary-layer instabilities that can
occur on swept wings are leading-edge instability, Tollmien-Schlichting
tangential instability, crossflow instability, and Taylor-Goertler insta-
bility due to surface concave curvature. These instability regions are
shown in figure 23 for the LFC airfoil and indicate where combined
shaping and suction were applied for control. A rather large supercriti-
cal zone (aspect ratio _ 0.37) exists over the upper surface flat-
pressure region followed by a steep rear pressure rise. The lower
surface is seen to be heavily loaded in the fore and aft regions with a
small supercritical zone in between.
Figure 24 shows the measured and designed chordwise pressure distri- butions for two chord Reynolds numbers at M_ = 0.82 on the LFC airfoil that are generally in good agreement. Essentially shock-free flow was obtained for the results shown. The slightly overall higher velocities on the upper surface and chordwise deviations from design are attributed to classical problems associated with transonic wind tunnel testing, wall interference, and model deformation under design air loads. The velocity field between the upper surface-tunnel wall channel (supersonic bubble zone) was higher than predicted due to the contoured liner wall and inability to completely account for three-dimensional boundary-layer displacement thickness effects in the design analysis. Measured coordi- nate deviations from design, obtained with a dial indicator under applied simulated design airloads, were about 0.O03-inch on the model forward upper surface at mid-span. This deviation corresponds to h/X = 0.0015 for multiple waves and is shown in figure _ to meet allowable criteria based on earlier results at low speeds but well above the projected goal for supercritical airfoils (ref. 19). The data in figure 24 indicate that flow separation occurs on the lower surface at about g/c = 0.80 when the Reynolds number is increased from R c = 10xlO 6 to 20xi0 6. Since transition correspondingly moved rapidly forward on the lower surface, the flow into the trailing-edge cusp apparently was unable to sustain the adverse pressure gradient. This separated flow changes the local effec- tive area distribution of the test section resulting in a slightly higher freestream b_ch number and increased sensitivity to local surface condi- tions and pressure variations.
The measured chordwise suction coefficient (CQ) distribution required to maintain full-chord laminar flow over both surfaces at the design Mach number and R c = 10xl0 6 is shown in figure 25 compared to prediction. The measured required suction was higher than the theory over most of the upper and lower surfaces. The higher suction require- ments were attributed to the previously discussed higher than anticipated velocities and surface pressure irregularities, the higher suction con- trol required to overcome cross-flow instabilities associated with the steep pressure gradients on the upper and lower surface nose and aft regions, and the minimization of centrifugal Taylor-Goertler type boundary-layer instabilities and interactions with crossflow in the concave regions of the lower surfaces.
A summary of the measured transition locations on the LFC airfoil upper surface for several Mach numbers is shown in figure 26. These results were obtained from a grid of flush-mounted surface thin-film gages to detect the state of the local boundary layer. Full chord lami- b 1 nat flow was maintained on both surfaces up to R c = t0x10 for al Mach numbers. As Reynolds number was increased for constant Mach number, transition moved gradually forward on the upper surface. The Reynolds number at which this forward movement began was dependent on Mach number. It was concluded that suction laminarization over a large super- critical zone is feasible to high chord Reynolds numbers even under non- ideal surface conditions on a swept LFC airfoil at high-lift conditions.
Analysis of both spanwise pressure distributions and transition pat- terns revealed that the flow over the yawed wing was not two-dimensional at M= < 0.80. It is believed that the resulting spanwise gradients influenced suction requirements and laminarization at the lower speeds.
The total drag at Moo = 0.40 and 0.82 for R c = 10xl0 6 with full chord laminar flow is seen in figure 27 to be equal to about (Cd)to t = 0.0030. This represents about 60-percent drag reduction as compared to an equivalent turbulent airfoil drag level of about Cdw = 0.0080 and a lift-to-drag ratio of about 180 based on design c£. Total drag is the sum of measured wake rake drag (Cdw) at mid- span and the suction drag (Cds) penalty required to maintain laminar flow. The suction required to maintain full chord LFC was some- what higher than anticipated (figure 25) and the contribution of suction drag penalty was about 40-percent for the upper and 60-percent for the lower surface. The increase in wake drag for M_ > 0.70 was attributed to the formation of a weak shock wave at the leading-edge region as the supersonic bubble began to develop. As the bubble develops (0.78 < Moo < 0.80), full chord laminar flow still exists, but periodic turbulent hursts occur over the upper surface causing an increase in wake drag. As Mach number is increased to 0.82, the supersonic zone spreads rearward to about 80-percent chord, the bursts disappear, and the wake drag returns to near its subsonic level (figure 27). It is concluded that the basic phenomenon of applying suction laminarization over an extensive supercritical zone is feasible up to high chord Reynolds numbers as demonstrated on a swept LFC airfoil at high lift conditions.
The major difficulty or influence in achieving this result was that of overcoming the classic "non-ideal" wind tunnel test environment and hardware tolerances.
CONCLUDING REMARKS Laminarization through passive or active methods is a boundary-layer stability problem which has been proven to be difficult to analyze, control, and verify. This is especially true as one moves from low to high speeds where swept-back wings and higher lift-to-drag ratios are desirable for improved aircraft performance.
In an effort to simulate flight conditions on models in wind tunnels, we need to better understand the environment and its influences on high-lift and low-drag testing. Several factors influencing the per- formance of low-drag airfoils have been identified which are primarily involved with overcoming the classic "nonideal '° wind tunnel test environ- ment and hardware tolerances.
NASA Langley has recently developed several advanced low-drag air- foil concepts with and without boundary-layer suction control for achiev- ing extensive laminar flow with very low drag. Verification of the anticipated performance of these concepts through wind tunnel testing, from subsonic to transonic speeds, has shown significant improvements in lift-to-drag ratio over previous airfoils designed for low drag. Proba- bly the most significant result at subsonic speeds is that the lift performance for these lower drag airfoil concepts is not degraded with fully turbulent flow over the airfoil surface. This provides a factor of safety in aircraft operation, should laminar flow be lost due to contami- nation. Suction laminarization over a large supercr[tical zone has been shown to be feasible to high chord Reynolds numbers even under non-ideal surface conditions on a swept LFC airfoil at high lift.
REFERENCES I • Jacobs, Eastman N.: "Preliminary Report on Laminar Flow Airfoils Method Adopted for Airfoil and Boundary Layer Investigations", NASA WR L-345, June 1939.
o Abbot, Ira H.; and yon Doehhoff, Albert E.: "Theory of Wing Sections", New York, Dover Publications. 1959.
.
Pfenninger, W.: 'Investigations on Reduction of Friction on Wings, in Particular by Means of Boundary Layer Suction." NASA Technical Memorandum 181, August 1947.
.
Lachmann, G. V. (Editor): Boundary Layer and Flow Control, Vol. 2, Pergamon Press, 1961.
• Pfenninger, W.: Laminar Flow Control, Laminarization.
AGARD-R-654, Von Karmon Institute, Belgium, 1977.
.
Pfenninger, W.; Reed, H. L,; and Dagenhart, J. R.: "Design Considerations of Advanced Supercritical Low Drag Suction Airfoils." Viscous Flow Drag Reduction, Vol. 72, Progress in Astronautics and Aeronautics, 1980.
# Gray, W.E.: "Transition in Flight on a Laminar-Flow Wing of Low Waviness (King Cobra)". R.A.E. Report N. 2364, March 1950.
_e Gray, W. E.; and Fullam, P. W. J.: "Comparison of Flight and Wind Tunnel Measurements of Transition on a Highly Finished Wing (King Cobra)"• R.A.E. Report No. 2383, June 1950.
• Pfenninger, W.; and Growth, E.: "Low Drag Boundary Layer Suction Experiments in Flight on a Wing Glove of an F-94A Airplane with Suction Through a Large Number of Fine Slots". Boundary Layer and Flow Control, Vol. 2, G. V. Lachmann, ed., [961, pp. 981-999.
I0.
Fowell, L. R.; and Ant,nat.s, P. P.: "Laminar Flow Control Flight Test Results, Some Results from the X-21A Program". Part 2, Recent Developments in Boundary Layer Research, Part IV, May 1905. AGARDograph 97, pp. 1-76.
I[.
Loft[n, L. K. Jr.; and Burrows, D. L.: "Investigations Relating to the Extension of Laminar Flow by Means of Boundary-Layer Suction Through Slots". NACA TN 1961, 1949.
12.
Braslow, A. L.; Burrows, D. L.; Tetervin, N.; and Visconti, F.: "Experimental and Theoretical Studies of Area Suction for the Control of the Laminar Boundary Layer on an NACA 64A010 Airfoil".
NACA Rept. 1025, 1951.
"Natural 13.
Holmes, Bruce J.; Obara, Clifford, J.; and Yip, Long P.: NASA Laminar Flow Experiments on Modern Airplane Surfaces."
TP-225b, June 1984.
[4.
IIolmes, Bruce J.; Obara, Clifford J.; Martin, Glenn L.; and Domack, Christopher S.: "Manufacturing Tolerance for Natural Laminar Flow Airplane Surfaces." SAE Paper 850863, April 1985.
15.
Croom, C. C.; and Holmes, B. J.: "Flight Evaluation of an Insect Contamination Protection System for Laminar Flow Wings." SAE Paper 850860, April 1985.
16.
Somers, Dan M.: "Design and Experimental Results for a Flapped Natural Laminar Flow Airfoil for General Aviation Applications."
NASA Technical Paper, 1965.
17.
Viken, Jeffrey K.: "Aerodynamic Design Considerations and Theoretical Results for a High Reynolds Number NLF Airfoil."
Master's Thesis. George Washington University. January 1983.
18.
McGhee, Robert J.; Viken, Jeffrey K.; Pfenninger, W.; Beasley, William D.; and Harvey, William D.: "Experimental Results for a Flapped NLF Airfoil with High Lift/Drag Ratio." NASA TM-85788, May 1984.
19.
Harvey, W. D.; and Pride, J. D.: "The NASA Langley Laminar Flow Control Airfoil Experiment." AIAA Paper No. 82-0567, March 1982.
20.
Reshotko, Eli: "Control of Boundary Layer Transition."
AIAA-85-0562, 1985.
21.
Mueller, T.J.: "The Influence of Laminar Separation and Transition on Low Reynolds Number Airfoil Hysteresis." AIAA-84-1617, 1984.
22.
Whitcomb, R. T.: "Review of NASA Supercritical Airfoils."
ICAS Paper No. 74-I0, Aug. 1974.
23.
Dagenhart, J. R.: "Amplified Crossflow Disturbances in the Laminar Boundary Layer on Swept Wings with Suction." NASA TP-1902, November [981.
24.
Ei-Hady, Nabil M.: "HADY-I, a FORTRAN Program for the Compressible Stability Analysis of Three-Dimensional Boundary Layers." NASA CR-3467, 1981.
25.
Srokowski, A. J.; and Orszag, S.A.: "Mass Flow Requirements for
LFCWing Design." AIAA Paper No. 77-1222, Aug. 1977.
26.
EI-Hady, N.M.: "On the Stability of Three-Dimensional Compressible
Nonparallel Boundary Layers." AIAA Paper No. 80-1374, July 1980.
27.
Mack, L.M.: "On the Stability of the Boundary Layer on a Transonic
Swept Wing." AIAA Paper No. 79-0264, Jan. 1979.
28.
Carmichael, B. H.: "Surface Waviness Criteria for Swept and Unswept
Laminar Suction Wings." NOR-59-438, 1959.
29.
Carmichael, B. H.: "Surface Imperfection Experiments on a Swept
Laminar Suction Wing." Northrop Rept. 8LC-124, NOR-59-454, 1959.
30.
Poll, D. I.A.: "Leading Edge Transition on Swept Wings." AGARD
CP-224, pp. 21-I-21-11, 1977.
31.
Ram, R. B.; Vemuru, C. S.; and Harvey, W. D.: "Hybrid Approach to
Steady Transonic Normal Shock-Compressible Laminar Boundary Layer
Interactions over Airfoils with Suction." AIAA 85-0522, 1985.
32.
Pate, S. R.: "Effects of Wind Tunnel Disturbances on Boundary-Layer
Transition with Emphasis on Radiated Noise: A Review." AIAA
Paper No. 80-0431, March 1980.
33.
Owen, F. K.; Stainback, P. C.; and Harvey, W. D.: "An Evaluation of
Factors Affecting the Flow Quality in Wind Tunnels." AGARD
Conference Proceedings No. 348 on Wind Tunnels and Testing
Techniques, July 1979.
34.
Harvey, W. 0.; and Bobbitt, P. J.: "SomeAnomalies Between Wind
Tunnel and Flight Transition Results." AIAA Paper 81-1225, 1981.
35.
Harvey, W. D.; Stainback, P. C.; and Owen, F. K.: "An Evaluation
and Assessmentof Flow Quality in Selected NASA Wind Tunnels."
NASA TM-85659,August 1983.
36.
Stainback, P. Calvin; and Owen, F. Kevin: "Dynamic Flow Quality
Measurementsin the Langley Low Turbulence Pressure Tunnel."
AIAA-84-0621, 1984.
37.
McGhee,Robert J.; Beasley, William D.; and Foster, Jean M.:
"Recent Modifications and Calibration of the Langley
Low-Turbulence Pressure Tunnel." NASA TP-2328, July 1984.
38.
Viken, J. K.: "Boundary Layer Stability and Airfoil Design." NASA
CP-2413, 1986.
39.
Waggoner, E. G.: "Computational Wing Design Studies Relating to
Natural Laminar Flow." NASA CP-2413, 1986.
40.
Morgan, Harry L.: "High-Lift Flaps for Natural Laminar Flow
Airfoils." NASA CP-2413, 1986.
41.
Dagenhart, J. Ray; and Stack, John P.: "Boundary-Layer Transition
Detection Using Flush-Mounted Hot-Film Gagesand Semiconductor
Dynamic Pressure Transducers." AIAA Paper No. 82-0593, March
1982.
42.
Horstmann, K. H.; and Quast, A.: "Reduction of Section Drag by
Blowing Through Rowsof Holes in Areas of Laminar Separation
Bubbles," Technical Soaring, Vol. VII, No. i, September 1981.
43.
Pfenninger, W.: "Investigation on Reductions of Friction on Wings
in Particular by Meansof of Boundary Layer Suction. NACA TM
1181, August 1947. (Translated from the German"Untersuchungen
uber Reibungsverminderungenan Tragflugeln, insbesondere mit Hilfe
yon Grenzschtabsaugung," Mitteilungen aus demInst. fur
Aerodynamik, Nr. 13, Zurich, 1946).
44.
Champine, R.; and Blanchard, W.: "Airplane Model Experiments with
Wedge-Shaped Strips Mounted on the Upper Wing Surface." Soaring,
vol. 18-19, 1955.)
45.
Stack, John.: "Tests of Airfoils Designed to Delay the
Compressibility Bubble. NACA TN-976, 1944.
46.
Eppler, Richard; and Somers, Dan M.: "A Computer Program for the
Design and Analysis of Low-SpeedAirfoils," NASA TM-80210, August
1980.
47.
Eppler, Richard; and Somers, Dan M.: "Low SpeedAirfoil Design and
Analysis." Proceedings of the NASA AdvancedTechnology Airfoil
Research Conference, NASA CP-2045, Part I, March 1978.
48.
Mueller, T. J.; and Pohlen, L. J.: "Boundary Layer Characteristics
of the Miley Airfoil at Low Reynolds Numbers," AIAA-83-1795, July
1983.
49.
Miley, S.J.: "On the Design of Airfoil for Low Reynolds Numbers,
Proceedings of the Second International Symposium on the
Technology and Science of Low-SpeedMotorless Flight, The Soaring
Society of America, Inc., September 1974.
50.
Lissaman, P. B. S.: Low Reynolds NumberAirfoils. Annual Review of
Fluid Mechanics, Vol. 15, 1983.
51.
Pfenninger, W.: Experimental Investigation of an Airfoil with High
Lift-To-Drag Ratios at Low Reynolds Numbers. Northrop BLC-Report
84 (NAI 560188), 1956.
52.
Mangalam,S.; and Pfenninger, W.: "Wind-Tunnel Tests on a High
Performance Low-Reynolds NumberAirfoil." AIAA-84-0628, 1984.
-_i_'- X-2I flight LFC Wind tunnel LFC wings _///_ / _ //- F-15 flight cone Rx)tr I06 cones and flat plates _X L__ _k,l i _, l .01 .i 1.0 I0 Turbulence level, % Figure I.- Influence of turbulence level on transition Reynolds number.
lO8 \ Data Exp Model \ NASA/LRC LTPT NLF(1)-O414F \ © \ NASA/LRC LTPT Long endurance NASA/LRC LTPT LFC Flat plate NASA/LTPT(ref 12) LFC airfoil (Braslow) .\ • M=.6 _l_l M= • o\ _ © NASA/LRC LTPT HSNLF(1)-0213 M--.82 o_ LFC • NASA/LRC 8-ft Swept LFC airfoil (Rx) tr - '_<I,\/M--'.2\\ NLF /'_ \ M _.2 -" I I .I0 1.0 .01 51u % or #Ip % disturbance level on measured maximum Figure 2.- Influence of tunnel transition Reynolds number for low-drag airfoils in two NASA Langley tunnels.
ORIGINAL pAG_ _. IS 2. 0 -- 1.5 V/Voo I. 0 .5 L I .5 1.0 x/c Figure 3.- Velocity distribution for long endurance airfoils.
Figure 4.- Long endurance laminar-flow airfoil tested in the Low- Turbulence Pressure Tunnel.
"---'--,}1 nitial model with waviness, etc.
• 020 Improved model surface lower surface at 0.03c (ylc =0.0) .Olc -- / Fixed transition upper and /z..
Cdw .012 - Free transition (y/c =0.01
k
a -- Free transition (y/c =0.325) • 008 -
/
1 I l • 004 .4 .8 1.2 1.6 2.0 ct Figure 5.- Long endurance laminar flow airf£il results from LTPT. M = 0.I0, R c = 14 x I0".
.5 I F59000cos 2A- ; A = 30 ° Model Data Exp. A .010 Norair tunnel 30° airfoil 6 - series i; U. of Michigan 30° O (ref 28, 29) [] NASA/ARC 12' PWT 30° NASAIIRC ITPT 0° D_"...... _ Long endurance • NASA/[RC I_TPT 0° NLF(1) - 0414 • NASA/I-RC8-ft 23 ° _.__ SCLFC(1)-0513 A .00, -,, SUPa._'o it_cal _\ \ ,0001 I \ I .01 . I0 1.0 X/c Permissible amplitude ratio (h/x) of multiple waves for Figure 6.- airfoils in wind tunnels.
-I. 0 -.5 0 - Cp °5 - 1.0 - 1,5 - Figure 7.- Calculated pressure dis- tribution for NLF(1)-O414F airfoil at design conditions, co = 0.43, M = 0.40; R = 10.0 x I06_.
2.0 -- _mq 1.6-
F
mY 1.2- _on .8 -- cI .4 [] Fixedat O.075c Fixed at O.OlOc E L I I I I I 1 1 [ I -. 4 -8 -4 4 8 12 16 20 -.2 -.I 0 .004 .008.012 .016 Q, deg cd cm Figure 8.- Section data f_pr NLF(1)-O414F airfoil.
R c = I0 x I0 '), M = 0.I0.
_[ _-2 0 M =.047 R =3 × 106 c RMS NLF ai rfoil output + .6 O__ : f Peak transition xlc +_0 _ __!:_L_ fTurbu'entjLamburst + 0 _- ! _- Laminar/Turb burst + .4 O_ _- " '_-_'_'_'--'_ _ Laminar I 0.1sec j Time Figure 9.- Example oscillograph recorder traces of thin-film gage RMS output signals with Lime for state of boundary-layer detection.
.8- ( .7 R x 10 -6 0 3.0 [] 6,0 © 8.0 c[ .6 _- _ A 10.0 _7 _5 .4 •50 .55 .60 .65 .70 (x/C)tr Figure 10.- Upper-surface transition loca- tion as determined from thin-film data for NLF(1)-O414F airfoil.
5f,deg
©-I0.0
-5.0
.010
- _0.0 A5.0 _ [h 10.0 <_
•008
12.5 /
_ a 15.o /
c d .006 _k o17.5 /
.004
•002
I I I I I 1 I • .4 -.2 0 .2 .4 .6 .8 1.0 cI Figure ii.- Effects of flap deflection _or NLF(1)-O414F airfoil Rc = 6.1 x I0 , M = 0.07.
• 010 - Turbulator tape 0 Off (,_ .OO8 cd .OO6 .OO4 J 0 .2 .4 .6 .8 cI Figure 12.- Effect of turbulator tape on section characteristics
for 0414F airfoil oRto. =
3.0 x 10o; M < 0.20; _f = -I.0 - -.5 D C P .5 -- 1.0 1.5 -- Figure 13,- Design pressure distribu- tion for HSNLF(1)-0213 airfoil, M = 0.70; c_ = 0,20, 1.6 -- 1.2- .8 -- (c[) B ,4 -- [] Fixed at O.05c m I I i L I L J -,4 .2 0 4 8 12 16 20 0 .004 .008 .012 .016.020 -.2 0 -8 --4 Cdw (Cm)B a, deg Figure 14,- Section data for HSNLF(1)-0213 airfoil in LTPT, M = 0,168; Rc = 4 x lOP; 6f = 0 °.
1 20 ], 8 -- 1o6 --
(C[)ma x
1,4- 2O 1.2 I I I I J 5 6 I 8 9 xlO° Reynolds number, R C Figure 15.- Effect of Mach number and Reynolds number on maximum lift coeffi- cient performance for NASA HSNLF(1)-0213 airfoil; af = 0 °.
.8- (i>--0 .... O-- O-X_ 0 .6 - Eppler theory (lower surface (xiC)tr .4 • 1 1 I _A .2 2 4 6 8 10 x 106 R C Figure 16.- Transition location for HSNLF(1)-0213 airfoil; M < 0.20, C& ----- 0 0 .
-1.4 _r tleory with B.L (ref46) -1.0 u --.,,_ 0 0 -. 6 Experimental J
C
-. 2 (smooth airfoil) "_ 0
P
C a= 4.0 ° R = 101262 .6 C t/c = O.0728 c[ =I.0 1.0 .6 .8 1.0 0 .2 .4 x/c Figure 17.- Measured and predicted pressure distri- bution for LRN(1)-IO{17 airfoil.
Eppler theory (ref 46) 0 Smooth airfoil • 2-D trips at.25c, .32c, and .53c 80- Pfenninger (ref 51) - 0 0 6O 5O __ ill,jill 0 c[#c d _ _<Z2" l//" 0 0 /../<0" o 2O Smoke tunnel -_ = 0.1% U I I I I (_ I _±__H J 2 4 5 6 7 8 9 I0 20 R x 10_ C Figure 18.- Maximum lift/drag variation with Reynolds number on LRN(1)-IO07 airfoil.
(a) _ = 3°.
ORIGINAL (b) o_ = 18 °.
Figure 19.- Smoke-wire flow visualization photographs for LRN(1)-IO07 airfoil. Rc = 40,000.
ORIGINAL PA_ IS
OE POOR q_JAUI"Y
(c) _ = -12 °.
Figure 19.- Concluded.
OF POOR _U_'_'
2.0 -- m 1.6 cI AR 1.5 .4 1.0 ._ [] 6 x 28 in 1-r L_ [ I i J -°4 -4 0 4 8 12 16 20 ac, deg Figure 20,- Tunnel sidewall effects for single-element airf2il, M : 0,30; Rc _ 6.0 x 10 -.
Flow direction _1 --8 °
_,_ _l!r! 7'!_''_' _' rT_I
_ili It ;i_j_!l' !'__r! ! I,tt;ill
' Ii
=i0 ° _ =12 ° Figure 21.- Oil flow pattern of single-element airfoil illustrating tunnel sidewall effects in 6 x 28-inch TT.
/_ ' _- Lift at mid span (_) and © _. near wall (C>) with blowing _,_, 3.0 _-- Lift at mid span without blowing Lift coefficient c_.
Blowing near nose and flap 2°5 _- Lift near wall without blowing l I J 2.0 0 5° 10 ° 15° Incidence a Typical tunnel sidewall effects for multi-element Figure 22.- airfoil showing requirement for sidewall boundary-layer control.
Design Moo= . 82 bf = 0 o c[ =. 55 .A_= 23° -1.5- ._llmien - Schlichting Cp 0 1.5 Figure 23.- Design chordwise pres- sure distributions and sonic lines on the LFC airfoil section.
NI =0.823 oo Rc=20x106 , c[ =,56 Rc=lOx106, c[ =.64 _0..0 0 O@
-. 8 ,_.___o__
_ _
4_--g__
%
C P • 4 I W •8( = 20 x 106 Design Rc=.55 IN =.82 c I I I I I 1.2 . I ] i ] I 0 0 .2 .4 .6 .8 1.0 .2 .4 .6 .8 1.0 x/c x/c Figure 24.- Comparison of measured and design pressure distri- bution for swept LFC airfoils.
Moo= O.82, Rc = I0 x 106 x 10.4 -10 -20 -30 -I0 CQ -20 0 __.l'_-/Design -30 i -50-40 Experiment J _) (_ -60 -70 Lower surface -80 I I I I I 0 .2 .4 .6 .7 .8 x/c Figure 25.- Measured and design suction distributions for swept LFC airfoi I.
M 1.0 .8 (X!rltr .60 .78 .6 ,4 I I I [ I I I 12 14 16 18 2D 22 24 x 106 R C Figure 26.- Variation of measured transition location with Reynolds number for swept LFC airfoil.
• 004 - cd . 002 - Suction drag ....... i I i I .4 .6 .8 1.0 Moo Figure 27.- Measured variation of drag with Mach number for swept LFC wing.
Rc = I0 x 10 6.
1 28