NASA aerodynamics program

Table of Contents Chapter Two Test Techniques and Instrumentation .................................................. 1 1 Density Field Measurements in a Nonequilibrium Expanding Flow Direct Numerical Simulation of Transition and Turbulence in a Spatially iii

Theoretical Determination of Noise Reduction With Increasing Blade XV-WATB Flight Investigations: Advanced Technology Blades Loads.

Development of High-Angle-of- Attack Nose-Down Pitch Control Outdoor Static Tests of the Full-scale Ejector-Lift/Vectored-Thmst .

iv

Experimental Study of Hypersonic Shock-Wave/Turbulent-Boundary Calculation of Forebody Flow Field for a Candidate Aero-Space Mane Configuration Effect of Fuselage Forebody Fineness Ratio on HSCT V Comparison of Heating Rate Calculationswith Experimental Aerodynamic/Aemtherodynamic Characteristicsof Effect of Atmospheric and Aerodynamic Uncertainties on Manned-Mars Aerobrake Feasibility .

vi

List of Figures Figure 1-2 . Separation Control for High-Lift Airfoils; Three Element Landing Figure 2.2 . Laser-Induced Fluorescence Instrumentation in the Ames 3.5-Foot Figure 3.1 . Evolution of Linear and Nonlinear Disturbances in Mach 1.6 Boundary Efficient Supersonic Wind Tunnel Drive System for Figure 3.7 .

Figure 3-11 . (A) Pressure Dilatation During 1D Rapid Compression Figure 4.1 . Direct Numerical Simulation of Transition and Turbulence in a Figure 4.2 .

Figure 4-3 . Unstructured and Adaptive Multigrid for the Three-Dimensional S3D - An Interactive Surface Grid Generation Tool ........................................................ 56 Figure 4-4 .

.

vii Figure 4.5 .

Figure 4.6 .

Performance of NAS Pseudo CFD Applications .............................................................. 74 Figure 5.2 .

Figure 6.1 .

Figure 6.2 .

Figure 6.3 .

Figure 6-4 . Directivity Characteristics of t h e XV-15 in Hover ............................................................

Figure 6.6 . Measud Contours of the 65 Day-Night Noise Level (DNL) for the XV-15 Tiltrotor Aircraft With 3 Blades, Corrected for 4 Blades, for Takeoff and ApproachConditions ................................................................

Figure 6.7 . XV- WATB night Investigations: Advanced Technology Figure 6.9 . Comparison of Measured and predicted Acoustic Signal for a Model Figure 6.10 . Comparisonof Predicted Rotor Blade Loading for 10' Azimuthal Figure 6.11 . Predicted Effects of Fuselage Scattering of an Figure 6.12 . Comparison of Mid-FrequencyNoise Contours for the BO-105 Figure 6.13 .

Figure 6.16 . University of Maryland/Ames Research Center Bearingless Figure 6.17 .

viii

Figure 7.3 . Typical Nose-Down Control Capability Characteristics for Relaxed Stability Combat Figure 7.5 .

Figure 7.7 . Outdoor Static Tests of the Full-scale Ejector-Li Wectorcd-Thrust STOVL E-7A Figure 7.8 . ... 136 Validation of Out-of-Ground Effect Prediction Capability for Powered Lift Aircraft Figure 8.3 .

Figure 8-4 . ComputationaVExperimental Parametric Study of Figure 8.5 . Technique for Hypersonic Powered Tests of Figure 8.8 . PEMACH . Computed Pressure Distributions and Figure 8.9 .

Figure 9.1 . Contours of Effective Perceived Noise Levels (EPNdB) for Standard Figure 9.3 . Aeroacoustic Loads on Thrust Vectoring Vanes of a Modcl F- 18 Figure 9-4 .

ix

Figure 10-2 . Computed Smamlines for Various Fence Heights Figure 10-3 . Integration of the VMS and ANOPP for High-speed Civil Figure 10-6 . (1) Sonic Boom Simulator . (L-90-5755) (2) Subjective Loudness of Sonic Booms Figure 10-10 . Supersonic Laminar Flow Control Program ................................................................... 2 .

Figure 11.2 . Energetics of Gas-Surface Interactions in Transitional Figure 11-5 . Experimental and Numerical Analyses of Small Rockets Figure 11-7 . Rarefied Flow Regime Hypersonic Waverider Figure 11-11 . Computed Meridional Pressure Contours from Mach 5.04 Flow of Premixed Hydrogen-Air Over a Blunt Body Using Figure 12- 1 . Additional Post-Aerocapture DV Requirements Resulting .

.

Figure 12.7 . Aeroassist Flight Experiment Ground-BasedTesting ..................................................... 2~ X

Chapter 1

Subsonic Aerodynamics

The objectiveof the SubsonicAerodynamics Program is to provide the technology to improve the performance,efficiency,and economicsof hture transport aircraft. To accomplish this objective, the program is currently focusing on the areas of hybrid laminar flow control technologyand high lift.

A hybrid laminar flow control system has been installed on the wings of a Boeing 757 transport aircraft to investigate the effectiveness, maintainability, and cost of applying this technology to subsonic transports.

High-lift studies include both wind tunnel tests and computational fluid dynamics (CFD) calculations. Two-dimensionalwind tunnel tests that focusedonthree element airfoils have been conducted. Three-dimensional tests at high Reynolds Number and large scale are planned. CFD calculations are guiding the activity for both the two- and three-dimensional tests.

Program Manager: Victor R. Corsiglia OAST/RF Washington, D.C. 20546 (202) 453-2261 , M = 0 . 8 CL= 0 . 5 H = 39,000 ft.

14 r - - Flute No.

c - - -

0.lSCq ' OMCq ' 0.45 Cq

X/C n 0.15 Dean Design D w l n

HLFC WAKE RAKE RESULTS

Total Pressure in Wake Figure 1-1. Hybrid Laminar Flow Control Flight Research ,

1-1 Hybrid Laminar Flow Control Flight Research

Objective. T o conduct flight research ex- Significance. The experiments of the past periments with the Hybrid Laminar Flow year have demonstrated that the suction re- Control(HLFC)systeminstalledintheBoeing quired to obtain large amounts of laminar 757 flight research aircraft. flow over the wing chord (as much as 65% chord) is much less than had been previously Approach. The program is jointly funded by believed, and amounts to one-third of that the NASA, the US. Air Force, and Boeing indicated by the early design calculations Commercial Airplane Co. An HLFC system (shown in Figure A). In addition, pressure measurements made in the wake behind the provides suctionboundarylayer controlin the trailing edge of the wing in the center of the leading edge (to the front spar) to control the highly three-dimensional laminar boundary 17-foot test span (shown in Figure B) con- layer. Downstream of the h n t spar, the firmed a major reduction in wing drag over a s d c e pressure distribution is used to stabi- wide variety of test conditions and showed lize the boundary layer flow and maintain that large amounts of laminar flow over the laminar flow to the wing shock or a f t pressure aircraft's wing box area had been achieved as r i s e . The suction surface in the leading edge indicated by the early hot-film data.

is a microperforated titanium skin with over Status/Plans. Flight testing is continuing.

19 million tiny, laser-drilled, closely spaced holes. Detailed experiments will be made to esti- mate the tolerance of extensive laminar flow Accomplishments. Last year the flight test to relaxation of the stringent laminar flow experiments yielded momentous results that design requirements.

could advance significantly the application of this technology into the commercial transport Dal V. Maddalon Flight Applications Division fleet. The research showed that the suction Langley Research Center system flow capacity, power requirements, (804) 864-1909 weight, and complexity had been overesti- mated by more than 200%.

M = 0.20, R = 9 Million - - - - - -

- -

Vortex Generators -

- d -<

, . ~ , l l l l , LTPT: Douglas 3-element airfoil I I 1 I 0.0 -5 0 5 10 15 20 25 a, O Figure 1-2. Separation Control for High-Lift Airfoils; Three Element.Landing Configuration; I Counter Rotating Low-Profile Vortex Generators

1-2 Vortex Generators for High-LiftAirfoils

Objective. To evaluate the effectiveness of Status/Plans. More vortex generator tests very small vortex generators for controlling are planned for 1992, as well as a vortex boundary layer separationoverthree-element generator jet test on a NASA high-lift model.

high-lift airfoils.

John C. Lin Approach. Small, vane-type vortex genera- Fluid Mechanics Division tors were mounted on a Douglas three-ele- Langley Research Center ment, two-dimensional model in landing con- (804) 864-5556 figurations as part of a cooperative test pro- gram between NASA and Douglas Aircraft Company conductedin the Langley Low-Tur- bulence Pressure Tunnel (LTPT). The inves- tigation was conducted at near-flight condi- tions with chord Reynolds numbers of 5 mil- lion and 9 million and a Mach number of 0 . 2 .

Accomplishments. Separation control with vortex generators was investigated during thesecondandthirdquarterofFY1991 onthe Douglas high-lift model. Measurements in- clude lift, drag, surface pressure, wake pro- files, and fluctuating shear stress. Low-pro- file vortex generators mounted at 25% of the flap chord, and with a deviceheight of only 0 . 2 inches at full scale, were effectivein alleviat- ing flap separation. The resulting wake of the three-element airfoil was significantly nar- rowed, providing measured drag reductions of up to 50%, (see figure). In addition, lift coefficients at approach angles-of-attack(i.e., 8" to 1 2 ' ) were increased by up to 10%.

. The results of this investiga-

tion have indicated (1) Separation allevia- tion on the flap can significantly improve the performanceof the entire high-lift system. (2) A practical vortex-generator system that not only can produce effective separation allevia- tion but also produce low drag when deployed, no drag when retracted in cruise (stow inside the flap well), and simplicity and ease of maintenance. (3) Low-profile, flap mounted vortex generators can be integrated into a modern transport to maximize performance and eliminate buffeting.

I SELF-ADAPTIVE AIRFOIL DRAG POLARS FOR SOLID AND THROUGH POROSITY POROUS NACA 0012 AIRFOILS Ma t 0.80 permeability .10 r .05 .020 Relative Drag thickness -.05 I I I I I -.lo[ 0 .2 .4 .6 .8 1.0 Chord Lift Figum 14. Porosity for Transonic Airf& With a View Toward Multipoint Design L

1-3 Porous Wansonic Airfoils for Multipoint Design

Objective. T o determine the feasibility of structed using a computational design tool.

using porosity to achieve multipoint design While the subcritical equivalent airfoil basi- for transonic airfoils.

cally retains the teardrop shape of the NACA 0012 profile, the supercriticalcompanionpiece Approach. A computational pilot study was reveals a distinct flattening of the upper sur- conducted with an in-house developed time- face combined with hump-shaped closure to- implicit upwind method for the nonconserva- ward the trailing edge.

tive Euler equations using floating shock fit- ting for accurate representations of shocks. Making the surfaces of an already optimized The computational approach allowed for a supercritical airfoil permeable also broadens faster andless-expensive assessment ofidu- its operational speed and incidence range.

encing parameters, such as profile geometry, arrangement of porous surface patches, and Significance. Euler analyses indicated a degreeofpomity,thananexperimentalstudy.

potential of using passive venting techniques for restructuring the flow past airfoils such Accomplishents. Solutions were com- that they become self-adaptive to very dis- puted for steady transonic flow over NACA similar flow conditions. This property could 0012 and supercritical airfoils with solid as open an alternate route for achieving well as porous surfaces. The porous surfaces multipoint design, that is, the design of air- were applied to both upper and lower profile foils which satisfies several, oftentimes con- sdaces, and they extended over the nominal flicting constraints.

chord. Either connected or separated cavities were assumed to lie beneath the upper and Status/Plans. The proposed concept will be lower sdaces. The porosity distribution is experimentally verified in the 8-Foot Tran- described by a modified sine wave with sev- sonic Pressure Tunnel at Langley Research eral amplitudes. The sinusoidal porosity dis- Center.

tribution was chosen to facilitatethe numeri- c a l approximation of flow through permeable Peter M. Hartwich (ViGYAN Inc.)

s - s . Applied Aerodynamics Division Langley Research Center Applied to a NACA 0012 airfoil, porosity gen- (804) 864-2881 erally increases lift for a given angle-of-at- tack, in some instances by up to 65%. As indicated by the accompanying figure, the wave drag taken at constant lift for supercritical flow past a porous NACA 0012 airfoil with separated cavities is up to one order of magnitude lower than for its solid counterpart.

The accompanying figure also illustrates the new quality of a porous NACA 0012 airfoil being self-adaptive to dissimilar flow condi- tions. Using the calculated porous surface pressure distributions as target pressures, eqdivalent solid airfoil shapes were con- dtACK AND WHITE PHOTOGRAPh

ste

Tuft Flow Visualization of the Flap System istributions on 1st Flap Element Flap deflectlon = 40 deg %= 2.46 Airspeed = 101 kts (Indlcated)

RN, c = 9.96 x 10'

Angle of attack = 10.21' c1 lower surface CP 0.0 0.2 0.4 0.6 0.8 1.0 xlc Figure 14. Flight Results on a Subsonic Transport Flap System

1-4 Subsonic Wansport High-Lift Flight Research - Phase I

Objective. To obtain in-flight aerodynamic contribute to the improved understanding of flow measurements on subsonic transport high-lift flow physics and will provide a good high-lift systems for correlation with wind- test case for correlation of wind tunnel data and validation of computationalfluid dynamic tunnel data and validation of CFD results to improve high-lift design methodology. results.

Approach. As part of a multiphase research Status/Plans. Additional flight tests (Phase activity, Phase I flight experimentswere con- IA) are planned for winter 1991/1992 to ob- ducted on a B737-100 (NASA 515) research tain full-chord pressure distributions along aircraft to measure pressure distributions on the wing and slat elements as well as the flap the high-lift flap system of a transport con- elements. Analysis offlight data will be made figuration. Phase I activities included mea- using two and three-dimensional computa- suring pressure distributions at select tional methods to help plan a follow-on Phase I1 flow physics flight experiment. Phase I1 spanwiselocations of the aircraft flap system.

Pressure data were obtained in flight using activities will incorporate the use of replace- belts of thin plastic tubing wrapped around ment 737 slat and flap spare parts to house the upper and lower surfaces of each of the instrumentation for detailed surface pres- trailing-edge flap elements at two spanwise sure distributions, boundary layer, and wake locations. In addition, surfacetufts were used flow measurements.

to document flow-separation characteristics, and Preston tubes were used to measure sur- Long Yip, Paul Vijgen Flight Research Branch face shear-stress characteristics at selected locations. Langley Research Center (804) 864-3866 Accomplishments. Phase I flight tests on the NASA 515 research aircraft were com- pleted in May 1991. Steady-state test points (about 250 test conditions) were obtained for a range of altitudes up to 20,000 feet and airspeeds as low as the stick shaker speed.

Chord Reynolds numbers ranged from 10 to 20 million and Mach numbers ranged from 0 . 1 6 to 0.40. Pressure data was obtained on each flap element of the triple-slotted flap system for deflections from 15" to 40". Flow- separation characteristics on the flap system were documented with surface tufts, and sur- face shear-stress measurements were deter- mined from Preston tube readings.

. An aerodynamicdatabase in-

cluding pressure distributions and documen- trition of flow separation was established on a complex high-lif't flap system in flight over a range of Reynolds and Mach numbers and angle-of-attack conditions. These data will

Chapter 2

Test Techniques and Instrumentation

Technology is being provided for criticalexperimental research required to improve the measure- ment of the fundamental flow properties of fluids and the overall aerodynamic performance of aircraft componentsand configurations. The developmentofinstrumentation and measurement techniques for real-time, flow diagnosis is being performed with emphasis on nonintrusive methods. These developments occur across the range of conditions from cryogenic to high temperature and &om low subsonic to hypersonic speeds.

Oneof themore excitinginstrumentation developmentsunderway is the use ofpressure sensitive luminescent paint on aerodynamic models. This will provide nonintrusive, inexpensive sensing of pressure in real time for wind tunnel and flight testing.

Other research areas being addressed are (1) heavy gas wind tunnel testing to operate at high Reyno~ds numbers; (2) solid-statecamerdoptical image correction techniques to correlate flight test dataofaircraft encounterswith wing tip vortices generatedby large aircraft, (3) nonintrusive Rayleigh-Ramanmultipoint measurements of gas density in hypersonic flow; (4) liquid crystal skin fiction sensors capable of high-frequencyresponse in high-speedflows; and ( 5 ) nonintrusive infrared thermography of global heat transfer rates or wind tunnel models in hypersonic flow.

Program Manager: Gary Hicks OASTIRF Washington, DC 20546 (202) 453-2830 , -1. Luminescent Paint Sensor Development , ' 2 . . 12

2-1 Luminescent Paint Sensor Development

Objective. To develop luminescent "paints" tion of the wing tip shock o f f the fuselage.

for the measurement of surface pressure and More subtle features such as the formation of temperature fields in aerodynamic testing.

streamwise vortices over the wing surface Our ultimate god is to develop this methodol- from the cross-flow and the subsequent sepa- ogy into a practical tool such that it can ration can also be seen.

replace conventional methods.

Significance. The oblique wing test repre- Approach. This work is a cooperative re- first use of the pressure paint under sents the search program between the Ames Fluid Me- supersonic flow conditions. The flow condi- chanics Laboratory and the University of tions for this test were outside the previous Washington Chemistry Department. Candi- operationalexperiencewith the pressurepaint date luminescent moleculdcoating formula- and were considered marginal at best. How- tions were identified, and trial coatings were ever, the paint did provide qualitative infor- made. The suitability of these trial coatings mation. From this experience we now have was assessed through static and wind tunnel more information on the practical operating tests. range of the pressure paint and the experi- mental methodology associated with its use Accomplishments. In May 1991 the super- in a large-scale supersonic facility.

sonic oblique wing test was being conducted in the Ames 9 by 7-foot Unitary Plan Wind Status/Plans, Further development of the Tunnel. The pressure sensitive luminescent pressure sensitive paint to improve its char- paint test was piggybacked onto the larger acteristics continues. Near-term work is f e experiment, whose objective was to evaluate cusing on the development of a temperature the performance and surface flow character- sensitiveluminescentpaint and a dual sensor is tics of a genericobliquewing model for Mach paint that will allow simultaneous measure- 1 . 6 to 2 . 0 . Figure A is a photo of the model in ment of pressure and temperature. We are the test section showing the model's lower also pursuing the development of lumines- surface. The lower surface of the wing was cent sensor coatings for unsteady aerody- coated with the pressure sensitive paint. namic applications.

Because of the extreme test conditions, a low test section static pressure and a high total B.G, McLachlan, J. Bell, temperature, it was not possible to determine J. Schreiner the absolute pressure level using the paint. Fluid Dynamics Research Branch Relative pressure change, however, could be Advanced Aerodynamic Concepts Branch seen. Ames Research Center (415) 604-4142 Figure B is a representative example of the data acquired. Shown is a map of the lower surface pressure field over the forward swept portion of the oblique wing. Conditions are .

notedin the figure. "he unique field measure- ment capability of the paint method is evi- dent: the paint capturing the passage of the fuselage bow shock over the outboard portion of the wing and the passage over the wing inboard portion of a shock arising from reflec- Figure 2-2. Laser-Induced Fluorescence Instrumentation in the Ames 3.5-Foot Hypersonic h Wind Tunnel

2-2 Temperature and Density Measurements in Air Using Laser-

Induced Fluorescence

Objective. To provide nonintrusive mea- Significance. Nonintrusive measurements surements of flow-fieldtemperature, density, of flow-fieldproperties and their fluctuations and their fluctuations owing to turbulence in hypersonic flows will provide significant with application capability in a l l air flows new information for turbulent model de- including hypersork wind tunnel flows. velopment and the validation of 3-D numeri- cal simulation codes that is not obtainable by Approach. Laser-induced fluorescence of other means.

oxygen is used in combination with Raleigh scattering h m the same laser beam to pro- Status/Ph& Theinstrumentation has been vide instantaneous and simultaneous mea- installedin the Ames 3.5-foot hypersonicwind surements of temperature and density from tunnel and preliminary measurements have each laser pulse.

been demonstrated. Further refinement of the instrumentation is in progress.

Accomplishments. Bench-topexperiments have demonstrated that measurements can Robert L. McKenzie be made at the fkee-stream conditions of the (415) 604-4749 Ames 3.5-foot hypersonic wind tunnel at Douglas Fletcher Mach 10 with uncertainties of less than 2%.

(415) 604-5244 Initial measurements have been made in the Experimental Fluid Dynamics Branch 3.5-ft Wind Tunnel at Mach 7. Ames Research Center , Figure 2-3. Application of Thermographic Phosphor Technique b

2-3 Phosphor Thermography Technique

Objective. To extend the capability of the ment technique. Globalheating distributions relative-intensity, two-color thermographic can be obtained minutes after a wind tunnel phosphor technique to provide global, quanti- run, allowing researchers to optimize (i.e., tative heat transfer data on models in hyper- tailor) investigation run by run. Compared sonic wind tunnels. with the thin-filmgage technique(recognized as the most accurate method for measure- Approach. Model surface temperature time ment of heat transfer rate), phosphor ther- history was accurately measured, pixel by mography provides invaluable global infor- mation, and the time and cost to construct pixel, prior to and during injection of the model through the nozzle boundary layer into ceramic models are an order of magnitude the test core and following flow establishment less than the time and cost for thin-filmmod- of fireestream flow about the model. One- els. This phosphor technique has revolution- dimensional conduction heat transfer equa- ized aerothermodynamic testing at LaRC and tions were used for each pixel on the model has generated considerable interest in the nationaMnternationa1 hypersonic-testing within the field-of-view of the camera (maxi- mum of 262,000 pixels for each time) to deter- community.

mine the quantitative value of heat transfer rate. These values of heating were compared Statdlans. A next-generation phosphor with benchmark values obtained with thin- system, which will provide tracking during f i l m resistance gages for the same model ge- the model injection process and accurate spa- ometry at the same flow conditions and atti- tial definition including regions of high sur- tude. face curvature, has been specified and is be- ing acquired. Advanced methods of applying Accomplishments. Global heat transfer phosphors are being examined and improved rates were measured on a ceramic "modified data reduction techniques are under develop- orbiter" model at Mach 10 in air using the ment. Potential applicationto impulse facili- relative-intensity, thermographic phosphor ties havingrun times less than 50 msec w i l l be investigated.

technique. Surface temperature time histo- nes were determined during the model injec- tion process, and newly developed software N. Ronald Merski was used to reduce data to values of heat Experimental Hypersonics Branch transfer rate. These results were compared Langley Research Center with heating distributionsmeasured with thin (804) 864-7539 film gages on the same configuration at the same flow conditions (referred to as HALIS orbiter CFD code calibration study). Data from these two techniques along the model centerline and several spanwise cuts showed good to excellent agreement.

Sigdicance. Phosphor thermography has evolved from a thermal mapping technique to a viable quantitative heat transfer measure- \ surements in a Nonequilibrium Expanding ,

2-4 Density Field Measurements in a Nonequilibriurn Expanding

Flow Using Holographic Interferometry

Status/Plans. In addition to these holo- Objective. To characterize the flow field of a grams, synthetic holograms using the com- two-dimensional nozzle using holographic in- puted flow-field data will be generated. The terferometry as a first step towards the inves- experimentally obtained holograms then w i l l tigation of nonequilibrium expanding flows.

be compared with the synthetic ones. Such an exercise will enable us to test our ability to Approach. The IO-cm drive section of the compute the expanding flows. Also, the den- Ames's electric arc-driven shock tube facility sity flow-field data from the holograms d l be was converted into a reflected shock tunnel a 2-D nozzle insert in the used in the next stage of the experimental facilityby installing program on the inveetigation of tube. A 3.2 W s e c shock in 100Torr nitrogen nonequilibrium expanding flows.

reflected off the nozzle insert, providing a reservoir ofcompressed test gas withP5-1500 psi and T5=7000 K. A NdYAG laser doubled Surendra P. Sharma, Scott A. Meyer, to 532 nm was used to produce holograms of Walter D. Gillespie Aerothermodynamics Branch the nonequilibrium expanding flow.

Ames Research Center (415) 604-3432 Accomplishments. A diagnostics system capable of producing a snapshot (6-11s expo- sure time) single-plate, double-exposure ho- logram has been developed. Holograms ofthe expanding flow, including startup flow, at intervals of 9,17,20,40, and 60 ps h m the time of shock arrival have been recorded.

Static pressures at 10points in the wall of the nozzle have also been recorded.

Significance. The relaxation time scale in expanding flows can be 1-2 orders of magni- tude lower than the time behind a normal shock. To develop a better understanding of this nonequilibrium in expanding flows, the vibrational populations duringthe relaxation will be measured. However, to accomplish this, the bulk density in the flow field must be known. The holograms provide the bulk den- sity data.

~~~~~

-

Figure 2-5. Infrared Map of Wakes Behind Vortex Generators in Hypersonic Flow L

2-5 Measurement Techniques for Hypersonic Flows

Objective, T o evaluate and develop experi- Status/Plans. Infrared tests will be contin- mental methods for use in scheduled Generic ued, and a parametric study ofvortexgenera- Hypersonics and NASP-related programs. tor geometry w i l l be made to determine the most effective configuration for hypersonic Approach. Two experimentaltechniquesfor flow. Measurementsofmassfluxfluctuations hypersonic flows in helium were examined in a hypersonic boundary layer will be made.

(1) An exploratory test was made to deter- Efforts to improve data reduction methods mine ifinfrared measurements could be used will continue.

to map the two-dimensionaltemperature field on the surface behind a series of vortex gen- erators. Prior to the test, it was not known if Ralph Watson the infi.ared imaging system would be sensi- Experimental Flow Physics Branch tive enough to accurately determine the loca- Langley Research Center tion ofvortex-inducedsurface heatingbehind (804) 864-5723 .

vortex generators. (2) Measurement of the constant-temperature hot-wire anemometer characteristics was made, and examinationof hot-wire data reduction techniques in hyper- sonic flow was begun by a summer research associate, Professor Eric Spina of Syracuse University in collaboration with Catherine McGinley of Experimental Flow Physics Branch, Fluid Mechanics Division.

Accomplishments. Despite the very low heat transfer rates on the surface of the base model in the High Reynolds Number Helium Tunnel, the wakes behind a series of vortex generators were visible, allowing the assess- mentofwake strengthas ahctionofgenera- tor angle and orientation. A frequency re- sponseof the hot-wire system was found to be adequate for measuring turbulence quanti- ties at Mach 11.

Infrared measurements w i l l enable a relatively simple determination of the geometry, size, and spacingof vortex gen- erators needed to trip a hypersonicboundary layer with minimum drag penalty. With a constant-temperature anemometer system, measuring mass-flow fluctuations in hyper- sonic boundary layers with adequate fre- quency response will be possible. These tests showed that a constant current system is not ne&ssary.

Chapter 3

Transition and Wbulence Physics

The objective of the Transition and Turbulence Physics Program is to develop a fundamental understanding of flow structures relating to transition and turbulence and to incorporate these flow structures into sophisticated flow models for use with computational methods. Extensive e&& are being developed within the program to solve the full Navier-Stokes equations, which include transitioning and turbulent flows. However, the realities of today's supercomputer limitations in terms of memory and speed make it impractical to consider full Navier-Stokes solutions for a l l but the simplest problems. Practically speaking, research in transition and turbulence modelingwith various levels of approximation is providingthe bridge for the gap until the computer power is available. The understanding of the physical flow structures of transitioning and turbulence flows is being examined from two directions, computational and experimental. Extensive advancements in test instrumentation being developed within other discipline programs is enhancing transitional and turbulence experimental research. The development of quiet supersonicwind tunnels will provide an experimental tool for investigating supersonic viscous flows. Exciting advances are occurring in compressible boundary layer transition with the use of parabolized stability equations derived from the complete Navier- Stokes equations. The program supports the Center for Turbulence Research which assembles world renowned experts in the field to generate new concepts regarding turbulence. A similar effort is being established for fluid mechanics research emphasizing transition.

Program Manager: Gary Hicks OAST/RF Washington, DC 20546 (202) 453-2830 , C m

*

W C Y n -0 Y a3 0 0 0 0 0 0 0 JDal(S llDM In 0- 7 I1 3.

t - I

s p a3

F X

*

“3 h

d Y II LL cc) mi

-

I I I Lo Lo 0 0

N T - .-

0 0 0 0 0 0 0 0 0 0 I 1 , Figure 3-1. Evolution of Linear and Nonlinear Disturbances in Mach 1.6 Boundary Layer Flow at the Given Nondimensional Frequency F(R= dRe,)

3-1 Compressible Boundary Layer Wansition Using PSE

wall shear. The results indicate that rapid Objective. To understand the mechanisms involvedincompressibleboundary layer tran- growth of the secondary disturbance is trig- sition, specifically to provide a capability for gered when the primary amplitude is suffi- boundary layer transition prediction in both ciently high.

“quiet” and “disturbed” environments.

Significance. The PSE method can be used Approach. For convective instabilities such to study the nonparallel and nonlinear evolu- as T-S waves in boundary layers, the govern- tions of disturbances in compressible bound- ing PDEs are only weakly elliptic along the ary layers starting from the linear stage up to dominant flow direction. The original PDEs transition. Coupled with receptivity phenom- ena this approach offers a computationally are reduced to a set of parabolized stability equations (PSE), which is parabolic along the viable means for studying and predicting the dominant flow direction, allowing solution by complex and intricate phenomenon of com- single-sweepmarching. Both nonparallel and pressible boundary layer transition.

nonlinear effects of growing boundary layers are studied up to the transition stage using Status/Plans. An axisymmetric version of this new PSE approach. the PSE will be developed and used for tran- sitionstudies in conical flows. The PSE theory Accomplishments. The PSE code for three- will be extended to 3-D boundary layers. Work is also under way on formulating the leading dimensionallinear and nonlineardisturbances edge receptivity problem using the PSE ap- was developed for flat-plate geometry. Linear proach.

PSE calculations were performed for Mach numbers 0 to 4.5. The effect ofboundary layer growth was shown to be important for both C . L. Chang the first- and second-mode disturbances in Fluid Mechanics Division supersonic boundary layers. Nonlinear PSE Langley Research Center calculations show very good agreement with (804) 864-5563 earlier temporal direct numerical simulations (DNS) at Mach 4.5 for the nonlinear evolution of a second-modedisturbance. The computa- tional time required is an order of magnitude less than that for DNS.

Part (a) of the figure compares the growth rates of a linear disturbance as computed by the parallel PSE (solid line), the nonparallel PSE (dashed line), and the multiscale DNS methods. The results indicatethat nonparallel effects are more important for oblique (y = 50°) waves than for 2-D (y = Oo) waves. A likely route to transition may consist of two oblique primary waves interacting with each other. Part (b)of the figure shows the oblique breakdown procedure for three different ini- tial amplitudes at the given frequency. Tran- sition is clearly located at the abrupt rise of Figure 3-2. Large-Eddy Simulation of Compressible Isotropic Turbulence

3-2 CompressibleLarge-Eddy Simulation of Isotropic Turbulence

Objective. To assess the subgrid-scale (SGS) not include actual large-eddy simulations.

model proposed for compressible large-eddy The results showthat this first-cut SGS model simulation (LES) by Speziale, Erlebacher, for compressible turbulence performs better Z a g , and Hussaini. than anticipated.

Approach. The assessment was performed Statdlans. The results of these first com- by conducting actual large-eddy simulations pressible LES calculations have sparked con- ofcompressible,isotropic turbulence andcom- siderable interest within the Langley group.

paringthe results with direct numerical simu- Several improvements to the compressible lation (DNS) data. part of the SGS model have already been suggested. The improved model will be ap- Accomplishments. An extensive new set of plied to more challenging flows, such as com- DNS data on 963 grids was generated. These pressible turbulence containing eddy data covered a range of fluctuating Mach shocklets, homogeneous turbulence in uni- numbers (M) and fraction of compressible form shear flow,and wall-boundedshearflows.

energy ( x ) , the key parameters in compress-

ible turbulence, according to a recent theory Thomas A. Zang developed at ICASE. The proposed SGS model Fluid Mechanics Division was utilized in LES for all of these cases. The Langley Research Center figure compares the time evolution of the (804) 864-2307 kinetic energy of decaying, isotropic turbu- lence from three different runs .. a fine-grid (963) DNS, a coarse-grid (323) DNS, and a coarse-grid(323) LES. As shown in the figure, coarse-grid LES is quite accurate on the total kinetic energy (E), but is somewhat less accu- rate for the compressible component of the kinetic energy (Ec). Moreover, the coarse- grid LES does substantially better than a coarse-grid DNS. Surprisingly, the model predictions for the compressible component sometime improve as the flow becomes more compressible. Some of the problems can be ameliorated by refining the constants in the SGS model. However, subtle changes in the model itself appear necessary to obtain more consistent predictions in the compressible regime.

S i g n i f i c a n c e . These were the first system- atic, compressible LES calculations that have tests of the SGS been performed. Previous modelwere based solely on DNS data and did Vorticity thickness Shear layer U1 M1 P1 T1 M2 p2 *2 "2 1.2 Matched total temperatures Ttl 3 1500 ' R M1 JM2 P i 3 P2 32116.8Pf 1.0

- A 2.0,2

n A m a

A 5 . 3 ~ a 4.9,l .8

- 14,2 1 e 4,.5; 8,l; l6,l.S

n . A

' 8

a i3,.5

5 0 6 - M0d.l

A 0 A a Standard k-e .4- A SEHK o M a n A A ' A 0- e 2

- A Bogdanolt compilation

m Elliot and Samimy I f I I Figure 3-3. E f f e c t of Compressibility Corrections on Vorticity Thickness Growth Rate L

3-3 Compressible Turbulence Modeling for High-Speed

Shear Layers

Objective. T o assess the performance of StatudPlans. T h i s study is ongoing. Recent several compressibility corrections to turbu- results were reported in June 1991 at the lence models applied to high-speed shear lay- AIAA 22nd Fluid Dynamics, Plasma Dynam- ers.

ics and Lasers Conference. Further studies will include testing of additional compress- Approach. "he compressibleNavier-Stokes ibility corrections and detailed comparisons equations were numerically solved using a with shear layer experiments.

two-equationturbulencemodelwithandwith- out compressibility corrections. Numerical John R. Viegas results were compared with available experi- Fluid Dynamics Division mental data to determine the most appropri- Ames Research Center ate model.

(415) 604-5950 Accomplishments. Each of the compress- ibility corrections studied was developed to increase the dissipation in the kinetic energy of the turbulence in a shear layer. By lower- ing the eddy viscosity through a reduction in the turbulence energy, these models reduced the spread rate of free-shear layers for a wide variety of flow conditions as the convective Mach number increased.

S i g n i f i c a n c e . Thevorticitythickness growth rate comparisons between computation and experiment indicate that the applicationcom- pressibility corrections can significantly im- prove predictions of high-speed shear layer mixing.

To Vacuum Nozzle Bleed Figure 3-4. Mach 18 Quiet Helium Tunnel ,

3-4 Development of a Mach 18 Quiet Helium 'limnel

Objective. To modify the Mach 20/40 Open projected for October 1991. Initial runs will Jet Helium Tunnel to operate as a Mach 18 use an existing,unpolished nozzle throat sec- Quiet Helium Tunnel. tion and will focus on ensuring that the sys- tem helium flow is free of particles that may Approach. Only the minimum essential damage the highly polished nozzle finish. In modifications required to convert the existing the event that particles are found tobe present, conventional facility to a quiet flow tunnel plans to install a filter upstream of the set- were made. The settlingchamber was outfit- tling chamber are in preparation.

ted with freestream turbulence and noise control devices. These consisted of a new inlet Stephen P. Wilkinson diffuser, two high-density woven wire porous Fluid Mechanics Division plates, and two high open-area screens. The Langley Research Center existing settling chamber was modified to (804) 864-5733 accept a slotted nozzle and associated bound- ary layer bleed system piping. The down- stream portion ofthe existingMach 20 conical nozzle was retained to eliminate the possibil- ity of nozzle wall transition due to the Gortler instability mode. The throat section, how- ever, was replacedwith a highly polished unit of the with an annular sonic oflice upstream throat for removal of the settling chamber boundary layer.

Accomplishments. Engineering design of the tunnel modifications was completedin FY 1989. Manufacture and acquisition of the newhardwarewascompletedduringFY 1990.

Installation of the new components and asso- ciated piping has occupied most of FY 1991.

S i g n i f i c a n c e . Because of the high radiated noise environment of conventionalhigh-speed wind tunnels, transition experiments with known noise sensitivity or experiments where the noise sensitivity has not been established must be conducted under quiet flow condi- tions. Virtually all hypersonic transitiontests fall in the latter category. If the modifications prove to be successful, this facility will offer the first hypersonic, quiet transition testing mpability above Mach 6.

Status/plans. At the present time, certifica- tion of the new high-pressure piping is 90% complete. Start-up and shake-down runs are o D p c ( o ~ l r m ( Y ~ o n Figure 3-5. Direct Simulation of Compressible Homogeneous Shear Flow ,

3-5 Direct Simulation of Compressible Homogeneous Shear Flow

Objective. T o increase the physical under- ratio Q(QM2t) approaches an equilibrium standing of the effects of compressibility on value of approximately 0.5 for large a.

flow turbulence so that these effects can be modeled in flow computations. S i g n i f i c a n c e . Simulations of homogeneous shear flow show that the growth of turbulent Approach. Homogeneous compressible tur- kineticenergydecreaseswith increasingMach

bulence wasconsideredwhereinalinearmean number - a phenomenon that is similar to the

velocityfield sustains the random turbulence experimentally observed reduction of turbu- field. A spectral collocation method, along lent intensities in the supersonic shear layer.

withathirdsrderRunge-Kutta time advance- The numerically generated databases have ment, was used to perform direct numerical exhibited great potential for developingmod- simulations (DNS), which are highly resolved els for terms (e.g., the compressible dissipa- in space and time. The resulting database tion) important in high-speed flows.

was analyzed for statistical and structural features of the flow. Status/Plans. Further analysis of the data- base will be carried out to improve the under- Accomplishments. Figure A shows the evo- standing of compressibility effects and aug- lution of turbulent kinetic energy, K, as a ment our present turbulence modeling capa- bilities.

functionof time, f nondimensionalizedby the shear rate, S, for several values of Mt, the Mach number of the turbulent velocity fluc- Sutanu Sarkar tuations. The figure shows that the level of ICASE turbulent kinetic energy decreases with in- Langley Research Center creasingvalues ofinitial turbulent Mach num- (804) 864-2 194 ber, Mta. Thus, compressibility decreases the growth rate of the turbulence intensities.

The exact equation for the evolutionof turbu- lent kinetic energy in homogeneous flows is

d(pK) / dt = pP - PES - p~ + p'd', where the

production, pP, and solenoidal dissipation, ES, are supplemented by two compressibility terms: compressibledissipation, p ~ , and pres- sure-dilatation, p'd'. Figure B shows that these Compressibility terms alter the balance between production and dissipation by as much as 25% for the case with Mt,o = 0.4, suggesting that incompressible turbulence models are inappropriate. Physically, the augmented dissipation that is due to com- pressibility and the transfer of energy from the velocity field to the pressure field via the pressure-dilatation term contribute to the reduced growth of K. A previous theoretical study had shown that the compressible dissi- pation obeys the scaling = M2t. Figure C shows that for a variety of DNS cases, the J J 40 80 80 120 140 20 180 100 0

Uot/M - 42

FIGURE 1 . Time development of resolved-scale turbulence kinetic energy from LES of

isotropic turbulence. o , filtered data of Comte-Bellot and Corrsin (323); ---- -, LES (323); A , filtered data of Comte-Bellot and Corrsin (643); ---.---- 8 LES (643).

2.0 1.6 *)

- 1.2

X A

2 0.8

V 0.4 I I I

0.01 ' I I _ _ _

0 50 100 150 280 250 300.

t

FIGURE 2. Time development of the plane-averaged w a l l shear stress <rw> in Re =

8000 transitional channel flow. A DNS (Zang et al. 1990); - present results; .... -... LES (Piomelli & Zang 1990b).

F'igure 3-6. Dynamic Subgrid Scale Modeling and the New LES Program ,

3-6 Dynamic Subgrid Scale Modeling and the New LES Program

Its inherent versatility should facilitate its Objective. Subgrid scale models used in use in applied CFD.

large eddy simulationsofturbulentflowshave had several drawbacks: the inability to cor- StatudPlans. The dynamic model will be rectly represent, with a single universal con- applied to complex flow configurations. These stant, different turbulent fields in rotating or include flow over a cylinder, backward facing sheared flows, the requiring of ad hoc damp- step, separated flat plate boundary layer, ing functions near solid walls, and the requir- compressible flow transition on a flat plate, ing of ad hoc intermittency functions to pre- and flow over a concave wall.

dict the transition from laminar to turbulent flow. A new eddy viscosity model has been Parviz Moin developed that alleviates many of these draw- Center for Turbulence Research backs. At present, our objective is to evaluate Ames Research Center and Stanford Uni- this model in a variety of incompressible, versity compressible, transitional, and fully devel- (451) 723-9713 or 604-5127 oped turbulent flows.

Approach. In large eddy simulations (LES) the large-scale field is directly computed and the effects of small scales are modeled. In the dynamicmode approach,the coefficient of the subgrid scale eddy viscosity is a function of space and time, and hence varies in the differ- ent flows and flow regimes. The coefficient is computed dynamically during the computa- tion rather than input a priori. This model was used in large eddy simulations of transi- tional and turbulent channel flow with heat transfer and in compressible homogeneous turbulence.

Accomplishments. The concept of dynamic modeling was used to derive expressions of the subgrid scaleeddyviscosity and turbulent Prandtl number. The results were in good agreement with the experimental data and direct numerical simulations, and were bet- ter than those of LES using conventional models that include ad hoc intermittence and dampingfunctions. The model has performed remarkably well without any adjustments in very different flow situations.

&gdficance. The idea of dynamic modeling is an altogether new approach in turbulence modeling. It has led to physically sound models for momentum and energy transport.

M Figure 3-7. Efficient Supersonic Wind Tunnel Drive System for Transition Research at I Mach 2.5

3-7 Efficient Supersonic Wind Tunnel Drive System for Transi-

tion Research at Mach 2.5

Objective. To validate design principles for ary for M = 2.0. The injectorspull conditioned a new generation of low-cost Laminar Flow air through the test section. The ratio of mass Supersonic Wind Tunnels (LFSWTs) using flow through the primary injector over that of “Quiet” technology. The LFSWT is essential the test section is an unprecedented 9.17:l.

for boundary layer receptivity studies at high- speed civil transport (HSCT) Mach numbers.

Significance. This low-budget project has demonstrated that we can operate a super- Approach. A V8 scaleProof of Concept (PoC) sonicwindtunnelwithanonspecializedindraft model of the LFSW” was built to research compressor. We now have the technology to unknowns for the full-scale design. The pri- design a drive systemfor the W-scale LFSWT.

mary purpose of PoC was to establish that the The LFSWT will enhance supersonic nozzle Fluid Mechanics Laboratory (FML) indraft and transition research progress withinNASA compressorcould support the desired LFSWT while impacting HSCT.

test envelope of M = 2.5 @ Re = 1.3 million per foot. Thelow compressionratio (-1.8:l) ofthe StatudPlans. The LFSWT drive system is compressor precluded the use of any conven- being designed, and the LFSWT design prin- tional tunnel drive system. For the past 19 ciples are being studied for the settlingcham- months, there has been an experimental re- ber and supersonic nozzle, which are critical search effort, supported by CFD studies, to to establishing quiet flow.

develop a unique drive system for the LFSWT.

In parallel, innovative design techniques for J. Laub, S. Wolf, L. King, D. Reda the “Quiet” settling chamber and supersonic Fluid Dynamics Research Branch nozzle are being investigated. Ames Research Center (415) 604-4136 Accomplishments. In February 1991, we achieved M = 2.5 in the PoC test section over Re range. To achieve the lowest the desired Re of 1 million per foot, a stagnation pressure (Po) of 5 psia, is required. APo of 5 is less than the exit pressure (Pe) of 8 psia, which makes the compression ratio across the test section 0.6259, which is a major and unique accom- plishment. By comparison, a conventional supersonic tunnel requires a compression ra- tio greater than 2:l to run and an overpressure to start. The PoC has an advantage in that the starting and running test conditions are the same. Operation at such a low compression ratio is achieved through the use of a super- sonic diffuser (optimized by CFD support) at the end of the test section and dual ambient injector system. The primary injector is de- signed for an M = 2.4 exit flow and the second- xt, at surface z: 1 *t2m (from heat transfer) .Lu- x2.39 X t r

= 1.12m

50 75 100 125 150 175 200 0 25 Xk3" Figure 3-8. Precursor Effect in Hypersonic Transition ,

3-8 Numerical Simulation of Laminar Breakdown in Supersonic

Transition

Objective. Toexplaintheoretically, the “rope- Significance. A comparison of U T , SIT, like structures”and the “precursortransition and DNS results for the spreading angle of effect” that have been observed in supersonic the disturbances indicates that the spreading transition experiments since the 1960s. The rate cannot be attributed to linear instability alone (as has long been believed), nor even to rope-like structures refer to the braided pat- terns that are visible near the boundary layer secondaryinstability,but is, indeed, a strongly edge in schlierenphotographs. The precursor nonlinear effect.

transition effect refers to the origination of transitional disturbances near the boundary Status/Plans. The numerical databases al- layer edge a considerable distance upstream ready generated will be analyzed in detail, of the first detection of transition by surface with particular emphasis given to their impli- cations for transition modeling.

measurements.

Approach. The approach was to conduct C.D. Pruett high-resolution direct numerical simulations Fluid Mechanics Division (DNS) of laminar breakdown in supersonic Langley Research Center boundary layers on cylinders and cones, uti- (804) 864-6788 lizing linear stability theory (LST) and sec- ondary instability theory (SIT) to select cases that represent the most likely paths to transi- tion.

Accomplishments. Two high-resolution temporal DNS ofthe stronglynonlinear, lami- nar breakdown stage of supersonic transition to turbulence were conducted: for Mach 4 . 5 flow past a cylinder and for Mach 6.8 flow past a cone. The top frame of the accompanying figure shows a schlieren photograph that is typical of experimental results for transition on a cone in hypersonic flow. The essential features are displayed in the cartoon in the middle of the figure. The spatially recon- structed turbulenceintensities extractedfrom the DNS of the Mach 6 . 8 cone case (bottom frame) display these key features. In particu- lar, the “precursor transition effect” has been captured, with even the spreading angles of the disturbance regions in good agreement with the experimental range: between 0 . 5 ” and 1”. Moreover, numerical schlieren flow- field visualizations fhm the DNS display “rope-like”structures are remarkably similar to those that have been observed experimen- tally.

, eceptivity of Low-Speed Boundary Layers

3-9 Receptivity of Low-SpeedBoundary Layers

Objective, T o develop a research program to analytical streamlines induced by the “rough- understand the essential issues underlying ness element” and the down wind extent of the receptivity process and to use this knowl- this disturbance.

edge in developingnew approaches to transi- tion prediction. Transition can be considered Significance. T h i s integratedcomputational/ a three-stage process consisting of receptiv- experimental project is designed to under- ity, instability, and breakdown to turbulence. stand the essential features ofboundarylayer While most previous work has been concerned receptivity. By collecting a detailed database with the issue of flow stability, receptivity is concerning the receptivity process, enough now considered a critical first-order param- knowledge will be gained to incorporatethese eter. effects in transition prediction schemes. The final result will be amplitude- and scale-de- Approach. A combined experimentallcorn- pendent criteria for transition.

putational effort is under way to develop a receptivity database. Followingprevious work StatudPlans. Current work is focused on on receptivity to suction surfaces (see FY 1990 low-speed, incompressible boundary layers.

Research Accomplishments), two new experi- Further receptivity studies on suction holes menta are now under way. One experiment are also planned. Once the basic physical concerns the transition of a laminar, low- processes are understood, emphasis will shift speed boundary layer in an adverse pressure to high-speed flows. A special interest is in gradient. The disturbance created by a puff receptivity of supersonic boundary layer (1.5 periodically introduced through a small hole I M 5. 2 . 5 ) to suction surfaces as are envi- in the test swface is tracked with an auto- sioned for the next-generation supersonic mated data acquisition system using phase- transport.

lock-averaging software. Pattern formation S. Davis, J. Watmuff, M. Tadjfa, in the boundary layer - expressed as velocity vectors and vorticity contours - is examined D. Reda using video animations. The second experi- Fluid Dynamics Research Branch ment concerns the receptivity to a roughness Ames Research Center element and follows a theoretical treatment (415) 604-4197 using triple deck theory. This experimentwill study the nature of roughness-inducedrecep- tivity resulting f r o m freestream acoustic dis- turbances.

Accomplishments. The response of a low- speed boundary layer to a periodic jet at one phase interval is shown in the left panel of the accompanying figure. The minimum local boundary layer thickness is approximately 6 mm. Animated flow sequences clearly show the generation of vortical flow structures and tkeir downstream convection. The second experiment is currently being designed based on results of the computational visualization.

The,right panel in the figure shows the

Energy Spectra in Kolrnogorov Units

10'1 0, 10 ..

-1- 10 ; !

-2 , , , , , , I , n I I r ' ,

1 o-2 10-l 1 oo

k . eta

Figure 3-10. Simulation of Homogeneous Turbulence on the Intel i860 Hypercube ,

8-10 Simulation of Homogeneous Wbulence on the

Intel is60 Hypercube .

Objective. To develop programming and a very high-performance level; in the case of compiler techniques to exploit highly parallel 128 40h4Hz i860’s, at a speed exceeding that computers for turbulence research. of 10 Y-MP processors. Within the next 2- years, machines of the iPSC/860 type w i l l Approach. The three-phase approach was have performance in the 50-Gigaflop range, first toconstruct a highly optimizingcompiler far exceedingthat ofcurrent supercomputers.

for the i860 microprocessor, then to develop and implement efficient inter-processor com- Status/Plans. We will continue to refine the munication procedures, and finally to apply i860 Vectoral compiler, since this processor and its successors will be used in important these to a program and compile an existing homogeneous turbulence algorithm. new multiprocessors, port other turbulence codes to the i860 machines to be used as true Accomplishments. First, the vectoral com- productionsupercomputers, and port the com- piler was ported to the i860 microprocessor piler and turbulence codes to other new mul- tiprocessor computers, such as the Stanford and the hypercube multiprocessor environ- DASH machine.

ment. A number of new compiler techniques were developed to exploit the high level of internal concurrency available in the i860. Alan W a y Techniques were also developed to alleviate Computational Fluid Dynamics Branch the problem of inadequate raw memory band- NASA Ames Research Center width. The homogeneous turbulence code of (415) 604-6066 Rogallo was then ported from the Cray Y-MP to the 128 processor NAS iPSC/860 Gamma Prototype. A highly efficient implementation of interprocessor communication was added tothecode. Awall-clockspeedof1.64 Gigdo s was obtained for the largest mesh size (256 B ).

Excellent agreement with experimental re- sults was obtained (see accompanying figure) at Reynolds numbers obtainableonly at great expense on Cray computers. The code has also been ported to the 528 processor Caltech Delta Machine, which has a mesh rather than a hypercube architecture.

Significance. It is recognized in the indus- try that highly parallel, distributed-memory computers are the future of supercomputing, and it is therefore crucial to begin learning how to implement our most computationally intensive applicationsonthese machines. This project has shown that a typical turbulence simulation code can run on such a machine at , I I I I I I I I I A -2 -4 E* -6 -8 -10 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Aot 1.4 I I 1 1 I I I I ei I I I I I I I I DNS (Lee) New model 1.2 w / o press-dilat.

............. ...... ............................

q 1

%

0.8

t I I I I I I I I . I I

0.6 0 2.5 5 7.5 10 12.5 15 17.5 20 22.5 25 x*k, Figure 3-11. (A) Pressure Dilatation During 1D Rapid Compression (B) Response of Turbulent Energy to the Normal Shock 4 4 3-ll Compressible Wbulenee:

Modeling Rapid Compression

Objective. To develop new, and improve the Significance. In the flowsof practical inter- old, models for compressible turbulent flows. est, the condition of directional rapid com- A specific goal reported here is investigation pression arises in flow codigurations where of the rapid compression effect on turbulence.

turbulence passes through a shock. A typical Thisefhtispresentinflowswith turbulence/ example of this kind is a compression corner shock interactions. flow. The discussed mechanisms of the rapid compression effect on the pressure dilatation Approach. In real flows, the individual has so far not been considered in turbulence me&anisms and processes of turbulence dy- modeling. Its significance is apparent in a namics cannot be studied in isolation. There- real flow configuration represented by a uni- fore, the basis for the investigation of the form turbulent flow passing through a normal rapid compression effect on turbulence has shock wave and shown in Figure B. Here, the been the direct numerical simulations (DNS) DNS data (Lee 1991) of the turbulent kinetic of compression of homogeneous turbulence energy (q2)response to the shock passage are (Coleman and Mansour, 1991). The study led compared with a turbulence model results, to a development of new turbulence closure with and without the rapidcontribution to the models that represent the physics of rapid pressure-dilatation term; the abscissa xkl compression. represents the distance normal to the shock.

Accomplishments. The study of the DNS of Status/Plans. The new model is being tested compressed turbulence yielded the following in the compression corner flow configuration; important findings: it was established that rapid compression theory has been devel- when nearly incompressibleturbulence (with oped; and modeling of turbulence subjectedto 1.111.8. Mach numbers Mt<<l) is rapidly com- rapid expansion is also being studied.

pressed in one direction (lD), unexpectedly high levels of negative pressure-dilatation Otto Zernan correlation are generated. The pressure-dila- Center for Turbulence Research tation term (nd) appears in the turbulence Stanford UNASA Ames Research Center kinetic energy and its magnitude during the (415) 723-9596; (415) 604-4726 1D compression can become by O(10) larger than the total dissipation (et); hence, nd can represent a significant loss of turbulent ki- netic energy (to pressure fluctuations), The striking aspect of this rapid compression mechanism is that it is most effective when Mt<<l, and that it is inefficient when the compression is more isotropic, i.e., acting in all three directions. All these aspects have been included in the new model for pressure dilatation. The model-DNS comparison of the ratio ll& for two values of Mt = 0.02 and 0.44 are shown in Figure A. Here, the abscissa &t represents the compression time with 1/( 1- &t) being the total compression. The differ- ence,in kinetic energy q2-levels (soley due to the pressure dilatation effect) is also pre- dicted by the model.

Figure 3-12. An Experiment to Guide Turbulence Modeling for Separated Flows ,

3-12 Separating Boundary Layer Experiment for

Turbulence Modeling

Objective. To evaluate turbulence models Statua/mane. The experimental database used in the prediction of separating flows in will be distributed to other researchers i n the such applications as high lift Moils. turbulence modeling community for use in model development.

Approach. A computational study was un- dertaken to numerically calculate two cases David M. Driver, Florian Menter of adverse pressure gradient: one with flow Experimental Fluid Dynamics Branch separationand the other attached. Solutions Ames Research Center to the incompressible Navier-Stokes equa- (415) 6045396 tions were obtained with INS3D and various different turbulence models. Comparisons between numerical results and previously measured velocities and turbulent Reynolds stresses were pedormed to determine which is the most appropriate model to use in ad- verse pressure gradients Accomplishments. Various turbulence m o d e l s such as the Baldwin-Lomax model, the Johnson-Kingmodel,the Wilcox(k-omega) model, and the Baldwin-Barth model were incorporated into INS3D. Solutions were obtained with each of the models for both the separating case and the attached case of ad- verse pressure gradient. These solutions re- vealed that only those models that accurately predicted the Reynolds shear stress would accurately predict the pressure distribution.

The numerical results as well as the ex- perimental measurements of velocities and Reynolds stresses were reported at the June 1991 AIAA meeting.

S i g n i f i c a n c e . Accuracy of flow calculations involving strong adverse pressure gradient is inextricably linked to the accuracy of the turbulence model that is employed. Compu- tational Fluid Dynamicsis being significantly advanced by the advent of the Johnson-King turbulence model, which was developed at NASA Ames Research Center and has been M e r validated in this study. k i n g Com- mercial Aircraft Company is actively and successfdy using the Johnson-Kingmodel in the design of future aircraft.

, ,

Chapter 4

Computational Methods and Validation

The objective of the Computational Methods and Validation Program is to develop and apply advanced analysis andcomputational methods for solvingcomplex,fluid dynamics problems and to perform detailed, benchmark experiments using redundant facilities and instrumentation to produce high-quality archival data sets to which computational fluid dynamics (CFD) solutions can be compared. Areas of interest include modeling turbulence and transition and computing complex flows (steady and unsteady, inviscid and viscous) over two-and three-dimensional geometries rangingin speed from zero to hypersonic and including such effects as mass injection and withdrawal. Additional objectives are to (1) demonstrate proof-of-concept computations for pioneering applications, (2) disseminate validated computer codes to the aerospace c o d u n i t y and provide maintenance and consultation on their use, (3) develop innovative techniques for scientific visualization of flow-field solutions, and (4) provide an experimental database that is taken in the form and detail consistent with CFD modeling requirements and that has documented accuracy and limitations of the experimental data.

The Computational Methods Program is focused on the present and future technology needs of the aerospace community. These needs include (1) developing faster and more efficient numerical algorithms to facilitate solutions of the full Navier-Stokes equations by large-eddy, simulationdsmall-scaleturbulence modeling; (2) developing advanced geometric modeling and grid generation techniques for complex, three-dimensional configurations; (3) improving under- standing of the effects of grid characteristics on solutions accuracy, convergence, and stability; (4) enhancing computational capabilities through development and use of advanced computer architectures and expert systems concepts; and (5) developingimproved methods for numerical simulation of aerothermodynamic flow phenomena associated with hypersonic cruise and maneuver vehicles, including real-gas chemistry.

Experiments are designed for comparison with numerical CFD results in order to (1) understand flow physics, (2) develop physical models for CFD codes, (3) calibrate CFD codes, and (4) validate CFD codes. The experiments range in speed from subsonic to hypersonic and include a variety of configurationsincluding generic, fightedattack, subsonic transport, rotorcraft, ASTOVL, and propulsion systems. Work continues in developinghigh-quality data bases for several classes of flows, including (1) high-and low-aspect ratio wings in subsonic and transonic flows; (2) simple 3-D turbulent flows, including time histories; (3) flow fields about aircraft components; (4) propulsive lift flow interactions in ground effect; and (5) unsteady flow interaction in rotor flow fields.

Program Manager: Pamela F. Richardson OAST/RF Washington, DC 20546 (202) 453-9857 x I O 4 I I I X Computation (Grid 'A') Computation (Grid O B ' ) (B ~xperim@nt (Suder et. a!.)

e ~xp@rimen~ (Sohn et. ai.)

-

\c, Tu~bulent correlation

. 1 Cf 4 0 2.5 5.0 7.5 10.0 32.5 x 105 flex e 4-1. Direct Numerical Simulation of Transition and Turbulence in a Spatially , Evolving Boundary Layer

4 1 Direct Numerical Simulation of Wansition and Turbulence in

a Spatially Evolving Boundary Layer

Objective. To develop a high-order-accu- and high-order-accurateschemes).The finite- rate, f d i f t k r e n c e method to computecom- differencemethod developed in this study can pressible, transitional, and turbulent flow be extended in a straightforward manner to and to demonstrate the capability of the curvilinear grids and thus enable direct simu- method by computing transition to turbu- lations of transitional and turbulent flow over lence on a flat plate. A study of the physics of general geometries.

transition and the creation of a data base for turbulence and transition modeling are an Status/Plans. The method has been ex- integral part o f this effort. tended to curvilinear grids. A new code to computecompressible,turbulentltransitional Approach. The unsteady, compressible for- flow over a turbine airfoil is being developed.

mulation of the Navier-Stokes equations are solved using a high-order-accurate, upwind- Man Mohan Rai biased, finite-differencemethod. This method Fluid Dynamics Division is an extension of an earlier high-order-accu- Ames Research Center rate method developed to compute incom- (415) 604-4499 pressible, turbulent flows. The new code that has been developed uses a zonal methodology toefficientlyprocess the number ofgridpoints used for the computation.

Accomplishments. The high-order-accu- rate method was successfullyused to compute to turbulence on a flat plate in a transition high freestream disturbance environment.

The numericalresults agreequalitatively with the availableexperimental data. Flow visual- ization of the computed flow indicated that the transition region was characterized by detached shear layers and pairs of counter- rotating streamwise vortices. The results indicate that the essential features of the transition process have been captured in the computation. Figure A shows instantaneous spanwisevorticitycontours just above the flat plate and Figure B shows the computed skin friction along the plate compared with experi- mental data.

Significance. This computation is the first “transition to turbulence”simulation. It indi- cates that computingtransition to turbulence ia certain spatially evolving boundary layers to a reasonable degree of accuracy is possible on currently available supercomputers (pro- vidqd there is judicious use of zonal methods

INRIA II CASE 3.4

Mach Contours

Detail of Separation Zone

0.200 Xsep 0.1 50 0 250 500 750 1000 CPU Time, seconds Figure 4-2. Multigrid Algorithm for Hypersonic Viscous Flows

4-2 Multigrid Algorithm for Hypersonic Viscous Flows

Objective. To develop a multigrid algorithm Significance. The multigrid method, previ- ously shown to be highly successful for sub- for the efficient solution of viscous flows at hypersonic Mach numbers. sonic and low supersonic flows, has been ex- tended to viscous hypersonic flow. An order- Approach. A Full Approximation Scheme of-magnitude increase in efficiency obtained (FAS) multigrid was developed based on an with the scheme allowed grid refinement stud- implicit upwind-biasedalgorithmfor the com- ies to be completed for all test cases of the pressible Navier-Stokes equations. The algo- INRIA code-validation workshop.

rithm uses upwind differencing for the con- vective and pressure terms and central Status/Plans. Further applications of the algorithm are continuing, including blunt differencing for the viscous terms. The im- plicit equations were modified so that the bodies and three-dimensionalconfigurations.

primitive variables, instead of the conserved James L. Thomas, David H. Rudy variables, were updated during the iteration.

Fluid Mechanics Division The time advancement and correction stages Langley Research Center of the FAS algorithm were modified to main- tain positivity of the thermodynamic vari- (804) 864-2146 ables, density and pressure. The FAS algo- rithm was modified to include multiple itera- tions on coarser meshes in order to speed up the upstream transfer ofinformation through the subsonic part of the boundary layer.

Accomplishments. The convergence rates for a number of hypersonic strong-interaction viscous flows were improved with the multi- grid scheme. The test cases were those of the second Institut National de Recherche en Informatique et en Automatique (INRIA) Workshop on Hypersonic Flows for Reentry Problems. An application to case 3 . 4 , corre- sponding to a laminar strong-interactionflow over a compressionramp at a Mach number of 11.68, is shown in the accompanying figure.

The Mach contours show the leading-edge shock and separation-induced shock at the compression; the enlarged view shows the separated flow in the region of the compres- sion corner. The variation of separation ex- tent with computer time is shown for the multigrid and single grid schemes. An order- of-magnitude improvement in efficiency oc- curs with the multigrid scheme.

d for the Three- nstructured an ,

4-3 Unstructured and Adaptive Multigrid for the Three-

Dimensional Euler Equations

Objective. To develop an accurate and effi- StatudPlans. Future work will center on cient method for computingsteady-state com- the inclusion of more adequate surface model- pressible flow about complex three-dimen- ing and grid generation techniques, as well as sional configurations.

the extension of the present work to viscous turbulent flowcasesin three dimensions. T h i s Approach. The steady-state 3-DEuler equa- will require the implementation of a multiple tions are solved on an unstructured tetrahe- field-equation turbulence model.

dral mesh by a Galerkin finite-element scheme. The flow variables are stored at the Dimitri J. Mavriplis vertices of the mesh, and an edge-baseddata- Institute for Computer Applications in structure is employed to minimize memory Science and Engineering (ICASE) requirements. An unstructured multigrid Langley Research Center technique is used to accelerate convergence to (804) 864-2213 steady-state. This procedure operates on a sequenceofnon-nestedcoarse and fine meshes, and the patterns for interpolation between the various meshes of the sequence are deter- mined in a preprocessing step, using an effi- cient search algorithm. This strategy is com- bined with an adaptive meshing approach, where new, finer meshes are automatically generated as the solution evolves.

Accomplishments. The present methodol- ogy is capable of providing accurate and effi- cient solutions for steady-state inviscid 3-D flows without incurring large memory over- heads. The figure illustrates the computation of transonic flow over an ONERA M6 wing. A multigrid sequence of four meshes has been employed, with the last two being adaptive meshes. The final solution was obtained in 100 multigrid cycles on the finest grid, which required 35 minutes of CPU time on a single Cray Y-Mp processor, and 22 million words of memory.

Significance. The ability to accurately and efficiently predict compressible flows over complex 3-D geometries is of particular im- portance to the aircraft industry. The use of adaptivemeshingenablesunprecedented reso- lution of highly localized phenomena, while the multigrid strategy maintains the overall efficiency of the scheme.

,

Figure 4-4. S3D - An Interactive Surface Grid Generation T o o l

,

4-4 S3D - An Interactive Surface Grid Generation Tool

Objective. To develop a surface grid genera- edge distributions to surroundingpatches. In tor with geometry modeling and surface the area-of database management, S3D per- gridding capabilities in an integrated, inter- mits a variety of data formats including that active environment, with the goal of vastly of PLOT3D for ease of data transfer.

reducing turnaround time.

Significance. The efficiency and ease with Approach. The scope of surface grid genera- which S3D can be used to handle commonly tion requirements were defined, and areas encountered tasks in surface grid generation that have not been adequately addressed or will help CFD analysts drastically reduce the treated by existing software were identified.

time required in gridding complex surface The initial development effort of S3D has geometries.

been directed at filling or enhancing those areas. A prototype S3D with primitive inter- StatudPlans. A workshopon S3Dis planned face was developed and applied on a variety of for this October. Future plans include adding complex geometries. These applicationswere options to handle some of the common data helpful in identifying problem areas and per- exchange formats under Interim Graphics mitted a greater focus on such issues as user- Exchange Standard (IGES) and capabilities friendliness and ease of operation. The S3D to design and analyze geometries based on code has since been expanded and enhanced non-uniform rational B-spline (NURBS).

with a user interface incorporatingthe latest in workstation technology. The code is now Raymond Luh ready for hands-on application by CFD ana- Fluid Dynamics Branch lysts.

Ames Research Center (415) 604-4494 Accomplishments. S3D can start from a geometry definition based either on a discretized curve set or a rectangular point set. Prominent geometric features such as discontinuities and high curvatures are eas- ily preserved and deficiencies in geometry data easily removed in transforming from geometry definition to surface definition. At the heart of S3D is the analysis of curves and surfaces by robust and widely applicablepiece- wise cubic and bi-cubic interpolation tech- niques made possible by highly reliable curve and surface fitting schemes. Point redis- tributions are accomplished, through smart and user-fiiendly interfaces, by the use of hyperbolic fimction-based, two-sided stretch- ing functions which have been shown to re- duce the spacing-induced truncation error.

Some of the more advanced features of S3D include surface-surface intersections, opti- mized surface domain decomposition and redomposition, and automated propagation of reparation of Grid Generation Input Data I

4-5 Accelerated Preparation of Grid Generation Input Data

Objective. To greatly reduce the user time The next step is to combine the three modules required to generate three-dimensionalmul- to produce a truly integrated interactive 3-D tiple-block computational volume grids.

volume grid generator. But this integration task is more significant than one might ex- Approach. The approach is to use the pect. The graphical user interface must run graphical capabilities of a powerful scientific on a workstation,but the elliptic grid genera- workstation to accelerate the process of col- tor must run on a supercomputer. Thus, the lecting and formatting the input data re- integrated system must run over a network- quired by a grid generation program. This connecting workstation and supercomputer new software tool leads the user through the running concurrently. This formidable pro- process of specifying input data by asking gram design and coding task is now well questions in sequence, error-trapping the re- under way.

sponses, and then formatting those data for the grid generator. Users having more expe- The integrated system will further accelerate rience can view all the input data in a random- the process by enabling the user to view the access fashion, enter or m o m it, and then iterative grid generation process as it is tak- format it with automatic error-trapping. A ing place. This will give the user a better suite of useful utilities completes the software understanding of the grid generationprocess, system.

and the ability to improve a grid generation effort that is working or to abort one that is Accomplishments. Preliminary tests indi- not.

cate that a reduction by a factor of ten has been achieved in the number of man-hours Reese L. Sorenson required to generate 3-D volume grids.

Fluid Dynamics Division Ames Research Center Significance. The generation of a suitable (415) 604-4471 grid for computational analysis of a realistic aerodynamic configuration can take as long as 6 months, hquently dwarfing all other aspects of the computational fluid dynamics (CFD) process combined. This problem be- comes even worse when the physical system to be modeled undergoes change requiring modification to the grid, such as deflecting control surfaces, rotating nozzles, or aeroelasticeffeds. Thus, any tool that greatly reduces the user time required to generate grids is a very significant advance.

Status/Plans. The software system to effect the above, called 3DPREP, is complete. An improvedprogramtoview 3-D computational grids, called 3DECANT, is complete. The elliptic volume grid generator, called 3DGRAPE, is also complete.

, Figure 4-6. Numerical Simulation of the YAV-8B Harrier VSRA in Ground Effect ,

4-6 Numerical Simulation of the YAV-SB Harrier VSRA in

Ground Effect Objective. To develop a validated powered- and potentially money saving new technique lift flow-field analysis capability useful in the for performance prediction of powered-lift design of advanced powered-lift aircraft. designs in the ground environment.

Approach. The flow field surrounding the Status/Plans. As a High-Performance Com- YAV-8B harrierin low-velocityjet-borne flight puting and Communications Program has been simulated through the numerical (HPCCP)ComputationalAero Sciences (CAS) solution of the Reynolds-Averaged Navier- Grand Challenge, the powered-lift project is Stokes equations. This simulation also uses developing numerical analysis tools for com- an engine model to specify inlet and nozzle plete powered-lift aircraft on massively par- mass flows and nozzle temperatures based on allelcomputers. Simulationmethods for aero- throttle inputs and inlet temperature and dynamics, propulsion, and controls will be pressure. Over 2.8 million grid points in 18 combined. Additional flight tests are sched- grids are used to resolve the flow field using a uled for the collection of infrared imaging diagonalized approximately factored algo- validation data. Following validation, it is rithm and the chimera overset grid method. expected that an accurate and neccessary design and analysis tool for powered-lift air- Accomplishments. Simulation conditions craft will result.

of 30 knots forward speed at 30 feet above ground level with exhaust nozzles rotated 81" William R. Van Dalsem, below horizontal have been modeled. This Fluid Dynamics Division computation was performed on a Cray Y-MP Ames Research Center 8/128 using 70 single processor CPU hours (415) 604-4469 and running at a rate of 130 MFLOPS(million floating point operations per second). Im- proved, spatially varying time step methods for powered-lift flow fields have allowed a factor of six increase in the global time step and a corresponding decrease in the cost of computation. Software has also been devel- oped for the analysis of infrared visualiza- tions which will be used in the validation of the aircraft simulation capability.

S i g n i f i c a n c e . Groundeffectflow fields about powered-lift aircraft, such as the Harrier, are highly complex. Analysis is generally resis- tant to small-scale wind tunnel techniques due to a wide variety of scaling effects. This leads to expensive full-scale powered wind tunnel tests to quantify the effects of ground proximity on vehicle performance. Some of the hazards of operating in the ground envi- ronment include hot-gas ingestion, foreign object damage, and the "suckdown" effect.

Th& simulation demonstrates a promising (&Im = 3.0, Rei = 6.3~10~)

(b) Drag coefficient

(a) Lift coefficient

CI Experimental data - Present Results

r Oqo5 r

0.20 0.04 0.15 0.03 n 0.10 0.02 0.05 0.01 -0.05 0*0° ~ 0.00

0 4 a 0 4 a

a I p ha( d eg .)

alpha(deg.)

Figure 4-7. High-speed Civil Transport Navier-Stokes Configuration .

4 7 High-speed Civil Transport Navier-StokesComputations

Objective. T o develop an efficient numeri- Status/Plans. The method described here cal procedure for computingviscous flow over will be generalized to accommodate block- a high-speed civil transport configuration. structured grids so that flow over complex aircraft can be calculated.

Approach. Amultistage Runge-Kutta time- stepping scheme with multigrid acceleration Veer N. Vatsa technique wm employed for computing the Fluid Mechanics Division steady-statesolutionsof the thin-layerNavier- Langley Research Center Stokes equations. The accuracy of the nu- (804) 864-2236 merical solutions was enhanced by the use of a matrix dissipation model.

Accomplishments. A multigrid-based fi- nite volume numerical scheme developed for computing high Reynolds number viscous flows over aircraft components at transonic speeds has been extended to accommodate supersonic flows. The resulting numerical d e , known as TLNS3D, has been applied to compute viscous flow over a generic high- speed civil transport model at the design cruise Mach number M = 3. The angle-of- attack varied from 0" to 8" in these computa- tions to correspond to an in-houseexperimen- t a l investigation, and each test condition re- quired approximately 2.5 hours on a Cray-2 computer. The computed lift and drag coeffi- cients are compared with the experimental data in the adjoining figure. The computed solutions are in excellent agreement with the data over the entire angle-of-attack range, even for the drag coefficient, which is domi- nated by the viscous component at lower lift values.

S i g n i f i c a n c e . Navier-Stokes solutions for the flow over configurations of practical inter- est can be obtained routinely using the multi- grid-based TLNS3D code. Because of the improved efficiency of this code, parametric studies to examine the performance of differ- ent con&prations are feasible without re- sortingto wind tunnel testingin early phases of vehicle development.

MACH 15 HELIUM NOZZLE, MACH NUMBER CONTOURS LEAST-SQUARES/PARABOLIZED NAVIER-STOKES DESIGN METHOD-OF-C~ARACTERISTICS/~OU~DARY LAYER DESJGN IFigUre 4-8. CFD-Based Hypersonic Wind Tunnel Nozzle Design I 4-8 CFD-Based Aerodynamics Design of Hypersonic Wind Tunnel Nozzles Objective. There was animmediate need for teristics with boundary layer (MOCBL)cor- a new design procedure for high Mach num- rection methods. The CFD-based nozzle de- ber hypersonic wind-tunnelnozzles, for which sign reduced the fluctuations of static pres- the classical designprocedures (usingmethod- sure to less than & 2% for the uniform core of-characteristics with boundary layer correc- region. This is a significant improvement tion) have failed to produce the high-quality over the classically designed nozzle, which flow fields necessary for fundamental flow had fluctuations of static pressure greater experiments and computational fluid dynam- than ~ 3 0 % for the core region.

ics (CFD) validation studies. The classical design procedure for hypersonic nozzles has Significance. The new procedure for design- been shown to break down at Mach numbers ing hypersonic wind-tunnel nozzle contours greater than 8 and when the boundary layer demonstrates that a CFD computer code can covers a large percentage of the test section be efficiently coupled to an optimization algo- diameter. rithm. The effect of changes in the nozzle wall parameters is evaluated by computing the Approach. To develop a procedure that uni- nozzle flow field using the PNS equations.

fies the best of classical design practices, This new procedure provides a method for designing hypersonic nozzles of high Mach CFD, and optimization procedures, an effi- cient CFD code, which solves the parabolized numbers in which classical procedures break Navier-Stokes (PNS) equations using an ex- down.

plicit upwind algorithm, was coupled to a nonlinear least-squares (LS) optimization StatuslPlans. Work is under way to apply procedure. A LS problem was formulated to the new procedure to design other wind-tun- minimize the difference between the com- ne1 nozzles and to extend it for use in the puted flow field and the objective function, design of hypersonic inlets.

and consisted of the centerline Mach number distribution and the exit Mach number and John J. Korte crossflow velocity profiles. The aerodynamic Fluid Mechanics Division lines of the nozzle were defined using cubic Langley Research Center splines, the slopes of which are optimized (804) 864-6920 with the design procedure. The thick bound- ary layer typical of hypersonic nozzles is ac- counted for by incorporating the solution of the PNS equations in the design procedure.

Accomplishments. Initial programmingand validation have been completed on a com- puter code that exercises a LWNS optimiza- tion procedure for designing hypersonic wind tunnel nozzles. The program couples an opti- mum design procedure based on the LS mini- mization problem to an explicit upwind PNS CFD code. In the accompanying figure, a Mach 15 helium nozzle designed using the new CFD-basedprocedure is compared with a nozzle designed with the method-of-charac- e Turbulent Flow Over a Backward Facing Step

4-9 Turbulent Flow Over a Backward Facing Step

Objective. To develop a detailed database Significance. The data being acquired in of the turbulent flow over a backward facing this investigation are unique in that they step €or the validation of direct simulations, document the Ml Reynolds stress tensor.

large-eddy simulations, and turbulence mod- Upon completionof the experiment, this data- els. base will represent the most complete statis- tical database for the validation of CFD meth- Approach. State-of-the-art instrumenta- OdS.

t i o n was used to acquire the mean velocities and turbulence quantities in the turbulent Status/Plms. The acquisition of the flow- field data is currently under way and w i l l flow. The primary instrumentation was a three-component laser velocimeter, which al- continue throughout the next fiscal year.

lowed the measurement of a l l three velocity components and the full Reynolds stress ten- Scott 0. Kjelgaard sor. Test conditions and geometries were Fluid Mechanics Division defined through interaction with the compu- Langley Research Center tational fluid dynamics (CFD) developers. (804) 864-2160 Accomplishments. During F Y 1990, the initial investigation of the flow over the step was performed. The test conditions for that experiment were chosen to be suitable for direct simulations. The detail and quality of the data generated interest in the turbulence modeling community, which requested a higher Reynolds number case. A new facility has been constructedthat allows a wide range of Reynolds numbers and expansion ratios.

Currently, the flow at a Reynolds number of 47,000 based on step height with an expan- sion ratio of 1 : 2 is being investigated.

DOPPLER SIGNAL TIME HISTORY ENERGY SPECTRA m-krr U-component FREQUENCY TIME 3 T REYNOLDS STRESS PROFILE

t

O S '~

0 _____I .UUL.

10 Y+ 100 Figure 4-10. Near-Wall Laser Velocimeter

4-10 NearcWall Laser Velocimeter Measurements

Objective. To increase near-wall Reynolds StatudPlans. Further develop the multifke- stress database for three-dimensionalturbu- quency signal processing system to obtain lent boundary layer flows with significant velocity components -vw and -uv Reynolds turning and separation. shear stresses. Fiber-optics are being devel- oped to facilitate 3-D measurements.

Approach. A new Laser Doppler Velocime- ter (LDV) instrument was developed capable Dennis A. Johnson of obtaining near-wall measurements in a Fluid Dynamics Division three-dimensional boundary layer. Measure- Ames Research Center ments were performed in a 3-D turbulent (415) 604-5399 boundary layer flow apparatus, located at Stanford University for which previous mea- surements of the Reynolds stresses were ob- tained with a hot-wire anemometer.

Accomplishments. Turbulence measure- ments were acquired in a 3-D boundary layer at a distance of 0.02 mm (0.001 in.) from the surface. Conventional LDV approaches are limited to a distance of about 1 mm from the surface. The technique has undergone fur- ther improvements, such as the development of a single channel multifrequency signal pro- cessingmethod for measuring the cross corre- lations. 3-D Navier-Stokes calculations (on a coarse grid) have been performed for the 3-D wedge flow experiment.

Significance. The accurate prediction of skin fiction and heat transfer is critically dependent on the modeling of the Reynolds stresses within the viscous sublayer. The present LDV development will allow the mea- surement of these stresses within the sublayers even under highly 3-D conditions.

Such data will aid in the development and validation of improved turbulence closure models.

,

Chapter 5

Numerical Aerodynamics Simulation (NAS)

The Numerical Aerodynamic Simulation (NAS) program provides readily accessible supercomputing capability to the United States’ top aeronautical researchers in Government, industry, and academia. The NAS program also includes research and technology development to ensure application of emerging technologies to computational fluid dynamic and other computational sciences. Specifically, the current research involves the enhancement of user interfaces and software and hardware technology for parallel computer architectures.

The objectives of NAS are (1) to maintain a pathfinding role in providing leading-edge to NASA, DoD, and other Government agencies, industry, and supercomputing capabilities universities as a critical element for continued leadership in computational aeronautics and related fields; (2)to stimulate the development of state-of-the-art,large-scale, computer systems and advanced computational tools for pioneering research and development; and (3) to provide a strong research tool for QAST.

To maintain the lead in large-scale computing capability, NAS is implementing a strategy of installing, at the earliest possible opportunity, the most powerful high-speed processor (HSP) available. NAS maintains at least two HSPs, one of which is fully operational and represents more mature technology, and the other, which is a higher performance prototype or early production model. The current HSPs are a Cray-2 installed in January 1988 and a Cray Y-MP installed in November 1988. The Cray Y-MP is the first computer to sustain a computation rate of a billion floating point operations per second (GFLOP). Current plans are to replace the Cray- 2 with a new processor (HSP-3) in 1992.

NAS was the first supercomputing facility to install a standard operating system (UNIX) and communication software on all systems. UNIX offers the flexibility of both batch and interactive a common user interface on all user-visible subsystems. NAS is computing and provides currently upgrading its remote network nationwide from NASnet, a switch-based, high-perfor- mance communication network, to Aeronet, a router-based system, allowing researchers at remote locations to have similar interactive capability as local users at the NAS facility.

The vision for the NAS program is to provide by the year 2000 an operational computing system capable of simulating an entire aerospacevehicle system within a computingtime range f h m one to several hours. It is estimated that a computing time rate of one trillion floating point operations per second (TFLQPS) is required to accomplish this goal.

Program Manager: Pamela F. Richardson OAST/RF Washington, DC 20546 (202) 453-9857 F Figure 5-1. Parallel Computers in the NAS Program

5-1 Parallel Computers in the NAS Program

Objective. To evaluate the performance available is 1 gigabyte. The iPSC/860 is a capabilities of parallel computers running MIMD (multiple instruction, multiple data) computational fluid dynamics (CFD) codes. architecture. Interprocessor communication is through message passing. The intercon- Approach. A base level of performance sta- nect network is a hypercube. The iPSC/860 tistics was established and algorithms that has a concurrent file system of 10 gigabytes increase the performance of parallel comput- supporting a peak transfer rate of 10 mega- ers were developed. bytes/second. At present, all communication is done through a small (PC 80386) front-end Accomplishments. The NAS Progrm has computer, which also handles partitioning of a Connec- the nodes and loading of jobs for execution.

acquired two parallel computers, tion Machine-2 (CM-2) from Thinking Ma- chines, Inc. (TMI) and an iPSC/860 system Significance. The highest performance at &om Intel. The CM-2 has 32,000 l-bit serial NAS for CFD-related applications on parallel processors and 1000 &bit Weitek floating computers is 1.6 GFLOPS using a 32-bit iso- point processors with a SIMD(single instruc- tropic turbulence code. This code used an tion, multiple data) architecture. The peak assembly language FFT (Fast Fourier Trans- speed is 14 GFLOPS (giga [billion] floating form) and a special compiler, Vectoral. The point operations per second). Scalar code is problem size was 2563 nodes. A similar code executed on either Sun or VAX front-end runs at about 160 MFLOPS on a single Cray Y-MP processor. This code is achieving a 10- computers, which broadcast instructions to the CM-2 processors. Interprocessor commu- fold performance-to-priceincrease on the Intel nication occurs three ways: general router, compared to the Cray Y-MP.

compiled, and NEWS (North, East, West, South, that is, nearest neighbor). The CM-2 StatudPlans. NAS will continue to explore has 4 gigabytes of memory, and a 25 gigabyte the potentials of both parallel systems.

Datavault for on-line storage as well as a high-performance parallel interface (HiPPI) T. A. Lasinski to UltraNet. The CM-2 was upgraded in NAS Systems Division January 1991 to take advantage ofTMI's new Ames Research Center slice-wise architecture, which presents a pro- (415) 604-4405 gramming model of the CM-2 based on 1000 floatingpoint processors,instead of the 32,000 bit serial processors. Such a model is more relevant to the CFD work done at NAS. TMI has recently provided a time-sharing operat- ing system and network queing system.

The iPSC/860 system was installed in Janu- ary 1990 as part of Intel's gamma prototype project with DARPA. This system has 128 nodes, each with an Intel 2360 64-bit micro- processor and 8 megabytes of memory. It has a 40 megahertz RISC chip rated at 60 MFLOPS (M=mega [million]). Peak perfor- manee is thus 7.5 GFLOPS. Total memory

Performance of NAS Pseudo CFD Applications (7/24/91)

Algorithm Computer Number of TimeIIter MFLOPS Efficiency 643 grid Processors (YW) (actuaVpeW LU BBN 62 12.13 36 1/17 CM2 32K 2.84 150 1/86 iPSC 128 2.34 182 1/42 YMP 1 2.10 203 YMP 8 0.35 1218 <I12 SP BBN 62 3.71 95 1fJ CM2 32K 11.63 25 11521 113 1 iPSC 64 2.42 122 YMP 1 1.49 198 >1/2 YMP 8 0.26 1335 1 1.83 BT BBN 62 77 1/8 11352 CM2 32K 24.25 37 1/19

iPSC 64 4.54 199

YMP 1 4.22 214 YMP 8 0.64 1411 >I12 LU: Lower Upper Triangular Solve SP: Scalar Pentadiagonal B T Block Tridiagonal Figure 5-2. Per€ormanceof NAS Pseudo CFD Applications

5-2 NAS Parallel Benchmarks

Objective. To develop anew methodology for Status/Plans. Implementationon other par- the performance evaluation of highly parallel allel computers w i l l continuein the next year.

computers running computational fluid dy- The inclusion of additional kerneldapplica- namics (CFD) applications. tions into future releases of the benchmarks is under investigation.

Approach. The principal distinguishingfea- ture of the new benchmarks is their “pencil Horst D. Simon Computer Sciences Corporation and paper“ specification-all details of these benchmarks are specifiedonlyalgorithmically. NAS Systems Division In this way many of the difficultiesassociated Ames Research Center with conventionalbenchmarkingapproaches (415) 604-4322 on highly parallel systems are avoided. The benchmarks consist of a set of kernels, the “parallel kernels,” and a simulated applica- tion benchmark. Together they mimic the computation and data movement character- istics of large-scale CFD applications.

Accomplishments. A set of benchmark ker- nels and pseudo applications has been se- lected and completely specified. Sample se- rialimplementationsfor the benchmarkshave been developed. Parallel versions of the bench- marks have been implemented on the 8-pro- cessor Cray Y-Mp, the 128-processor Intel iPSC/860, and the 32,000-processor Connec- t i o n M a c h CM-2(fromThinkingMachines).

The benchmarks have been distributed to a large number of computer vendors and aca- demic researchers interested in performance evaluation.

Significance. TheNAS ParallelBenchmarks represent the first comprehensive effort to assess the performance ofhighlyparallel com- puters. They present a new methodology that allows the suitability of parallel supercom- puters for aerosciences applications to be evaluated. In addition, the benchmarks are an invaluable tool for selectingcomputers for the High Performance Computing and Com- munication Program (HPCCP) testbeds and will ensure the selection of computers suit- able for the HPCCP “Grand Challenge” prob- lems.

, NAL PAGE WfTE PHOTOGR

. Virtual Wind Tunnel

,

5-3 Virtual Wind Tunnel

Objective. T o explore the effectiveness of Status/Plans. Future plans include distrib- virtual environment technologyfor flow visu- uting the software so that computation and alization.

data management is done on asupercomputer and upgrading to a higher resolution head- Approach. Virtual environment technology tracking display system.

provides a new approach to user interfaces in computer software. This approach involved Creon Levit integrating a variety of input and display NAS Systems Division devices to give the user the illusion of being Ames Research Center immersed in an interactive computer-gener- (415) 604-4403 ated environment. The computer-generated scene was displayed in stereo creating the illusion of depth, and was rendered from a point of view that tracks the user‘s head movements. The user had an input device, typically an instrumented glove, through whichobjectsappeared tobe directly manipu- latedin the computer-generatedenvironment.

Accomplishments. A virtual environment was implemented for exploring numerically generated three-dimensional unsteady flow fields. A variety of interesting techniques were used for visualization and for navigation through the flow. A boom-mounted six de- gree-of-freedom head position-sensitive ste- reo CRT system was used for viewing. A VPL DatagloveTM Model 11, which incorporates a Polhemus 3 SpaceTM tracker, was used to sense the user‘s hand position, orientation, and finger joint angles. The user‘s gestural movements were interpreted as commands for injecting various tracers (e.g., “smoke”) into the virtual flow field. An eight-processor Silicon Graphics computer system was used for computation and rendering.

Significance. We have shown that a high- performance graphicsworkstation can be used tovisualize three-dimensional flow fields in a virtual environment at acceptableframerates; virtual environments are an effective tool for the rapid exploration of three-dimensional flow fields, both steady and unsteady; and the system may be applicable to the visualization of other vector fields.

arrier Topological Vortex Cores ,

iel ati W

To create simple visualizations We plan to extend the soft- e the important properties of vec- e visualization of vortices and tor fields. to integrate topological from the neighborhoodof critical p w i l l inves- module, called Topology, tigate implementingTopologyin o the Flow Analysis Soft- ization environments, particularly the ) graphics software pack- (FLOW Real-Time Analysis) software, also display a portion of the topol- developed at Ames, which enables a user to ogy of vector fields. FAST is a multipmcess interactively generate particle traces and sur- software environment that consists of a col- faces, and EXPLORER, develo lection of separate modules that can be run Graphics, Inc.

simultaneously, sharing data and allowing the user to create, load, analyze, visualize, A I Globus animate, and record computational fluid dy- Computer Sciences Corporation namics (CFD) data. FAST was created at NAS Systems Division Ames Research Center, and betaversions ofit Ames Research Center (415) 604-4404 are currently being tested at various U.S.

sites.

nts. Topology, the new FAST graphics software module, finds, classifies, and displays critical points, i.e., where a vec- tor field vanishes. The software integrates curves through the field from initial positions near critical points on the invariant mani- foldsofthe linearized field near the points. By carefdly choosing the curves to integrate, the user can find the vortex cores of all vortices containing critical points and/or examine the skin friction surface topology to determine the location of separation lines and other features of interest. The software has very flexible mechanisms for exploring other topo- logical features as well.

. Users of Topology can quickly

find and display some of the important fea- tures of vector fields. In particular, separa- tion lines and most vortex cores may be found quickly, easily, and, in many cases, automati- cally. Since the software functions properly on the large, multi-zoned, blanked (masked) grids typical of state-of-the-art CFD, scien- tists can gain additional insight into the re- sults of their computations.

Shared Space

Receiver/Sender * Conversation --+ Receiver/Sender

Scientist + Conversation Scientist

Su pemornpute r Graphics Monitors Figure 5-5. Distributed and Cooperative Visualization of Unsteady Fluid Flow ,, cooperative environment is created via a net- work connection to the supercomputerand is izing computational fluid uninhibited by geographical proximity.

The initial implementation of nterview is a proof-of-the-con- cept capability providing interactive visual- a supercomputer, ization of unsteady fluid flows by using dis- application. The supercom- ted processing. Interview‘s prototype ties of high performance, large IWIS (“what you see is what I see”) disk storage, and fast disk interface provides a base for further research to contain, access, and into group communication via a visualization medium.

ata sets endemic to un- The high-speed graph- tation presents high- s. A prototype of Tempus FugitJ Interview has been completed. Additional ddition, the human visualization techniques and cooperative in- face environment provided by the work- teraction methods will be added as opera- ling the analytical tional experience is gained.

tools of the visualization system.

Michael J. Gerald-Yamasaki th cooperative processing, visualization is NAS Systems Division as a communication medium. The para- Ames Research Center is illustrated by the analogy of one (415) 604-4412 scientist outlining ideas on a chalkboard for t. Equal accessibility to the th scientists provides a plat- form for the exchange of ideas. The images producedon graphics workstations can be the platform. In combination with distributed processing, scientists may use this platform at geographically distant locations.

ompli s . TempusFugitAnterview is a CFD visualization application. Tempus Fugit (“time flies”) interactively creates im- ages animated over time from large data sets representing unsteady fluid flows. The com- panion program, Interview, presents simul- taneously, identical and animated images to another workstation in a shared environment WithTempus Fugit. Interview attaches to the supercomputer process created by Tempus Fugit and transforms the process to a shared envirpnment. Identical images are presented to both workstations simultaneously. Thus, a ,

Chapter 6

Rotorcraft

The objective of the Rotorcraft F’rogram is to provide the technologies for helicopters and other rotorborne aircraft to achieve quiet, low-vibration operation with increasedperformance, agility, maneuverability,and stability, all with acceptable handling qualities. Much of the work is done in conjunction with the U.S. Army and the FAA and in cooperative programs with industry.

Improved analysis can now handle many local aerodynamic phenomena, but the transonic, unsteady, complex wake interaction flow requires another generation of codes. Therefore,much of the mtorcraft program still is empirically based, with validation of analysis as a goal. This approach requires new test techniques, upgraded test facilities, carefully instrumented models, sophisticated simulations, and increased use of the latest generation of computers. New ideas also are part of the program, since the complexity of rotorcraft makes it a fertile area for innovative approaches.

The heart o€theaerodynamicportion of the program is airloads research. Small scale, pressure- tapped blade data have become available with the full-scaleresults due in 1992. Complementary efforts are under way in component, interference, and wake testing and prediction. These databases are used in acoustic and vibration research, which constitutes more than half of the program resources. The airload models also will be used for rotor state control and higher harmonic control to suppress noise or vibration or to enhance maneuverability.

Flight dynamics research in the Aerodynamics Division is based on unsteady rotor dynamics, with emphasis on simulation and eventually flight tests using a variable stability UH-60. New flight test capability is planned, with research challenges in higher-frequency control for rotor state control to achieve high agility.

Higher-speed rotorcraft activities are now focused on commercial tiltrotor technology. Certifi- cationissues are being addressedon the simulator and in noise testing and prediction. Also under way are improvements in performance, interior noise, vibration, and stability. The importance of these research opportunitiesfor civil applicationshave been reported under contract to Boeing Commercial Airplane Company.

Program Manager: Mr. George Unger OAST/RF’ Washington, DC 20546 (202) 453-5420

w

Large-Scale Tiltrotor Performance ,

6-1 Large-ScaleBltrotor Performance Program

Objective. To acquire rotor performance StatudPlans. The Ames Prop Test Rig will and loads data with and without the wing be rebuilt for both 40 by 80 cruise perfor- installed to determine the influence of the mance testing and 80 by 120 transition test- wing on rotor operation, provide rotor perfor- ing for performance and acoustics. The test- mance data up to and beyond the aircraft ing will use existing government-owned 25 R operating envelope, and evaluate acoustics diameter XV-15 rotor hardware.

for terminal area operations.

Jeffrey Light Approach. A large-scale test program was Rotorcraft Aeromechanics Branch developed for the 40- by 80 foot wind tunnel. Ames Research Center The Prop Test Rig (PTR) was used to evaluate (415) 604-4881 the forward flight performance for three dif- ferent large-scale rotor blade sets including the V-22, XV-15 Metal, and XV-15 ATB (Ad- vanced Technology Blades). The required measurements include acoustic as well as dynamic and performance data. Measure- ments made with and without the wing would be used to quantifj. effects of the wing on the rotor performance and acoustics.

Accomplishments. Testing was initiated, and isolated rotor data up to 240 knots was acquired on the 0,658-scaleV-22 rotor prior to incident. The PTR recovery plan was initi- ated to rebuild/strengthen hardware to pro- vide full test envelopesin both 40 by 80 and 80 by 120 test configurations.

S i g n i f i c a n c e . The data obtained during testing will be used to provide experimental validation of the V-22 tiltrotor aircraft cruise performance. Future acoustic data will pro- vide criticalinformationinto the feasibility of tiltrotor terminal air operations.

, w

n

I

Figure 6-2. Tiltrotor Aeroelastic Stability Control ,

6-2 Tiltrotor Aeroelastic Stability Control

Objective. To investigate the use of active cantilevered-wing tiltrotor configuration the controls to delay tiltrotor whirl-flutter type wing-chord instability couldbe increased from instability. 280 knots to above 300 knots by feeding back vertical and horizontal accelerations mea- Approach. The comprehensive rotorcraft sured near the wing tip to the longitudinal analysis code CAJMFUD/JA was used to ob- cyclic pitch.

t a i n a set of linear differential equations, which describe the motion of a tiltrotor air- Significance. The results of the investiga- craft at variow speeds. Thehub motions from tion have shown that (1) active controls,using wingmOay vibration are a standard input to the rotor cyclic pitch can be used to delay the CAMRAD/JA. The hub-motion is calculated occurrence of whirl flutter instability; (2) an usingaseparate structuralanalysis program. activecyclicpitch controllevel of 0.010°-0.0120, The CAMRAD/JAoutputconsists of the open- which equates to a 10-15 pound active control loop system matrices which describe the air- force applied at the hub, was required to craft motion in the state-variable domain. stabilize the tiltrotor; and (3) contamination The matrices formed an input to a separate of the sensor output signal with a 25% noise program, which performed the closed-loop, level did not adversely affect the closed-loop active control calculations. Additional input stability.

consists of the sensor model and the active controls feedback gain factors. The program StatudPlans. The whirl-flutter alleviation performed an eigen-value analysis to deter- investigation will be extended to include the mine the flutter stability for both the open- anti-symmetric flight mode and for the com- and closed-loop systems. Time response cal- plete aircraft in which the symmetric and d a t i o n can be performed to estimate the anti-symmetric flight modes are coupled magnitude of the required active control in- through the active control system. Optimiza- put for closed-loop stability. A separate util- tion of the feedback system for a given sensor ity was used tocombine aircraft trim informa- model will also be studied.

tion from CAMRAD/JA with wing vibration data h m the structural analysis program to Johannes M. van Aken define a sensor model, which is based upon Rotorcraft Aeromechanics Branch Ames Research Center physically measurable signals.

(415) 604-6668 Accomplishments. The use of active con- trols to alleviate whirl-flutter was investi- gated for a cantilevered, high-wing tiltrotor model and an advanced joined-wing tiltrotor aircraft. Sensor models, defining the feed- back of either pure state-variables or physi- cally measurable wing accelerations to the rotor cyclic pitch were evaluated. The analy- sis showed that for both tiltrotor aircraft con- figurations, the feedbackof accelerations, rep- re-senting the wing torsion and chordwise motions, to the longitudinal cyclic pitch could increase the flutter velocity for the symmetric flight mode. The figure shows that for the 0.7-SCALE V-22 SEMI-SPAN 67 DEG WING FLAP ANGLE

2*04 1.5

-2.5 1 I I I I I m I d 0 . 0 0.2 0.4 0.6 0.8 1.0

x/c

Figure 6-3. Wing Surface Pressure Distributions Within Wake , 6-3 'Tiltrotor Download Reduction The 2-D velocityhorticity formulation results Objective. To understand the physics of tiltrotor download and develop download re- are the first-ever calculations to accurately ductionconcepts to minimize the adverse per- predict the distributed airfoil pressures on formance impact of the tiltrotor's wake on the both the upper and lower wing surfaces. The aircraft's wingin hcvering, vertical, and tran- staggered grid method provides for the direct sition flight. solution of the incompressible Navier-Stokes equations by means of a fully coupled implicit Approach. An experimental and analytical technique. This unique method was demon- program has been pursued to quantify and strated to accuratelypredict the mean lift and demonstrate download reduction concepts. A drag and the periodic loading for several 2-D model 7-foot diameter V-22 rotor system was bodies including an NACA 0012 airfoil with used in hover at the Outdoor Aerodynamic and without a deflected flap at -90" incidence.

Research Facility (OARF). Full potential and Navier Stokes models of tiltrotor hovering StatudPlans. Plans are toreport the results flow fields have been successfully developed of the experimental download program and to to characterize and analyze the physics of the conduct a second test evaluation of other unsteady, vortical flow fields. download reduction concepts. The prediction methodology will be expanded to include 3-D steady and unsteady analysis, turbulent flow, Accomplishments. A successful hover test and rotor/fuselage interaction.

was conducted at the OARF in FY 1991.

Results included a demonstration of down- Paul M. Stremel load reduction with upper surface wing blow- ing. Shadowgraphs of the wakehinginterac- Rotorcraft Aeromechanics Branch tion were also acquired. Several papers were Ames Research Center (415) 604-4563 presented on three tiltrotor wake analysis codes. Two steady codes, a full-spanpotential flow code andasemi-spanNavier-Stokes code, were developed. Wing pressures within the rotor's wake as well as outside the wake along the wing were predicted and compared. A new two-dimensional vortical wake interac- tion code using a velocity/vorticity formula- tion on a staggered grid was reported.

Significance. Results from the experimen- tal program are directly applicable to aircraft configurations and result in a 10 to 20% re- duction in download on current generation tiltrotor designs. The shadowgraph wake flows quantify the unsteady flow field, and the wing pressure database will be used to validate new 3-D analysis codes. The compu- tational methods are the first successful analy- ses of tiltrotor flow fields in hover, and are avqilable to industry for wing design on the next generation of tiltrotor aircraft.

Figure 64. Directivity Characteristics of the XV-15 in Hover

6-4 Tiltrotor Hover Acoustics

Status/Plans. The acoustic data from the Objective. To perform full-scale flight tests for validation of ROTONET and to determine hover test program is being prepared as two methods of minimizing rotorcraft noise. articles for publication. A H S Technical Spe- cialist Meeting Rotorcraft Acoustics and Fluid Dynamics. Follow-on test programs are Approach. Acooperative program with Ames planned for level flight and terminal area Research Center has been defined to evaluate descenvascent conditions: a quick program the acoustic impact in hover of the Advanced scheduled for August 1991 and a more com- Technology Blade system on the XV-15 prehensive program scheduled for early 1993.

tiltrotor aircraft. A database for validation of ROTONET is an objective of this program.

Accompfishments. A hover test evaluation Danny R. Hoad Applied Acoustics Branch of the XV-15 tiltrotor aircraft was conducted at Ames Research Center. The hover test Langley Research Center (804) 864-5060 program was conducted in December 1990.

Significance. The results of the hover test Brent Wellman, Martin Maisel Flight Experiments Branch program have shown that: (1) A significant reduction of in-plane noise was achieved by Ames Research Center as expected. (2) The (415) 604-6573/6372 reduced rotational speed, longitudinal directivity measurements iden- tified a strong impulsive character propagat- ing aft of the vehicle. This character has been identified before and related to the “fountain effect.” Triangulation is currently in progress to identify the location of the impulsive source (see figure).

6-5. Civil Tiltrotor Technology

6-5 Civil Tiltrotor Technology

Objective. To determine the technical, eco- come the technology barriers and to develop nomic, and environmental factors that are enhancing technology to encourage civil ap- critical to the success of commercial, passen- plications.

ger-carryingtiltrotor aircrafl in the national transportation sy&em.

S i g n i f i c a n c e . The United States holds a commandingleadintiltrotor technology. This Approach. A follow-on study was pedormed studyidentified the opportunities, challenges, with Boeing Commercial Airplane Group and actions required to develop a civil tiltrotor teamed with Bell Helicopter Textron and aircraft that is both market responsive and Boeing Helicopter to identify highest payoff effectively integrated in the national trans- technologies, economic evaluations, market portation system utilizing its unique capa- size and “cost/price/ market” loop closure, bility. The development of the recommended civil requirements, and the potential changes actions will reduce the r i s k to industry.

needed to the V-22 for civil application. Op- erational analyses were conducted for steep Status/Plans. The Phase I1 Study is com- approachestovertiports, and actionsrequired plete and a final report was released in Febru- to achieve a national civil tiltrotor transpor- ary. NASA is working with FAA to help tation system were defined. determine Pre-TerminalOperatingProcedure Criteria for safe, steep approaches to urban Accomplishments. The study found antici- vertiports. Barrier technologies and other pated demand for a fully designed commercial high payoff technologies are being assessed passenger carrying tiltrotor is forecast to be by NASA for future R & T consideration and 2625 vehicles by the year 2000, growing to the Advanced TiltrotorTransport Technology 4925 in the year 2010. The market value of new initiative.

2625 aircraft is $32 to $42 billion, with half this being a domesticmarket and the remain- John Zuk der apotential worldwide export market. The Civil Technology Office study found that the tiltrotor has promise as Ames Research Center a means to relieve airport congestion signifi- (415) 604-6568 cantly. A projected airline schedule of airline service in the northeast corridor indicated that as replacement commuter aircraft a tiltrotor network could open up 1000 airport slots per day (1/3 of 1989 daily slots).

The study found that the cost/price/market loop can be closed. A fully designed commer- cial passenger-carryingtiltrotor was found to be differentfromthe V-22 and must overcome three barrier technology challenges: (1) low external noise, (2) human factors-based pilot controls for commercial flight, and (3) termi- nal area low-speed control and navigation, allowing safe, precise, steep approaches. The study recommends increasedresearchtoover- I -0 e 9 -0

cv

I : -0

c)

l-

a

U

-0

n

n

a

-0 F I -0

cv

I -0 e 9 0 0 ' 0 0 0 0 v) 0 v) l- F I Figure 6-6. Measured Contours of the 65 Day-Night Noise Level (DNL) for the XV-15 Tiltrotor Aircraft With 3 Blades, Corrected for 4 Blades, for Takeoff and Approach Conditions

6-6 Theoretical Determination of Noise Reduction W i t h

Increasing Blade Number for the - 1 5 lfltrotor Aircraft

Objective. T o predict the change in the 65 Significance. This study has demonstrated day-nightnoise level (DNL)contour area for a a capability to use the ROTONET system tiltrotortype aircraft as a result of increasing noise prediction code to predict trends and the number of rotor blades per rotor. The magnitudes of resulting noise reduction fkom flight conditions of particular interest were Changing rotor system design parameters.

takeoff and approach.

S t a t d P h Future plans will examine Approach. Contour corrections were ob- the predicted noise results to explain the tained theoreticallyby comparing ROTONET differences in relative magnitudes ofthe noise acoustic predictions made for the baseline corrections and to apply the approach to other XV-15 tiltrotor system (three blades per ro- combinationsof rotor head design and operat- tor)withpredictionsmadefortheXV-15 modi- ing parameters.

fied to four blades per rotor. The resultant correction factor was applied to modifya mea- Robert A. Golub Acoustics Division sured XV-15 ground contour.

Langley Research Center The predictions were calculated using (804) 864-5281 ROTONET Phase I and included propagated tone and broadband noise. Effective per- ceived noise level (EPNL) values were calcu- lated at five rows under the ground track 63" and 45" port of the ground track; on the ground track; and 63" and 45" starboard of the ground track. All predictions were made for the XV-15 in a level flyover at 90 knots air- speed, 85" nacelle tilt, 95 knots groundspeed, 250 feet altitude, and gross weight of 13,000 pounds. Wing off-loading of rotor thrust was accounted for. Noise values from the port and starboard rotors in isolation were summed to give the combined rotor noise signature. The difference between the predicted baseline EPNL and extra blade EPNL was averaged over the five observer locations. This average was assumed to be a noise reduction attrib- uted to a unit increase in the number ofblades per rotor.

Accomplishments. The resultant noise re- duction due to one additional blade per rotor was 6.2 EPNdB. This correction, when ap- plied to the measured XV-15 ground contours (seefigures),corresponds to a shrinking of the 65 DNL contour surface area by a factor of 0.23 for both takeoff and approach conditions.

LOW SPEED (35-40 KTS) Shimmed Gimbal Shimmed Glmbai & Untwlrted TIPS Configurations Figure 1. XV-16 collective actuator loads trends with configuration

modifications - comparison of predicted and measured loads at low

speed.

0 Baseline 0 Shimmed Gimbal A Shimmed Gimbal with Untwisted Tips

- CAMRADMA

m 800 L

s

400 Endurance $ L Limif-317 LB \ #)o e3 30 50 80 70 Velocity (knots) Figure 2. XV-18 collective actuator loads correlation using a modified version of CAMRADlJA -Helicopter Mode, MSL, 36 to 80 I C T S Figure 6-7. XV-lYATB Flight Investigations: Advanced Technology Blades Loads, , Stability, and Performance

6-7 xv-1 Flight Investigations: Advanced Technology

Blades Loads, Stability, and Performance

Objective. T o investigate advanced technol- animportant mnfigurationperturbation(hm ogy applied to tiltrotor blades for its impact on the original XV-15 metal blades) to develop the performance, dynamics, loads, and and validate predictive codes.

aeroelastic stability of the advanced Technol- ogy Blades. These data provide a configura- Status/plans. A scheduled 200-hour major tion perturbation for the development and transmission inspection and overhaul will validation of analytical prediction codes. follow the completion of the FY 1991 XV-151 ATB flight tests. A vibration survey will be Approach. Advanced aerodynamics and conducted immediately following the flight mixed matrix composite materials are com- activity to document the airframe resonant bined in the design of the Advanced Technol- frequencies in support of the dynamics and ogy Blades (ATB) for XV-15 Tiltrotor Re- load analyses. Control system and blade search Aircraft. The highly twisted blades modifcations to increase stability and reduce are configured with a compound planform, loads will be performed during the inspection thin tips, and an increased solidity to improve period. A full spectrum of tiltrotor flight experiments in support of the advanced hover lifk capability without degrading per- formance at high speeds in the airplane mode. tiltrotor transport technology activity will commence in late FY 1992 after the inspec- Flight tests of the ATB will support the devel- opment of a broad range of analytical method- tionloverhaul effort is completed.

ologies.

Brent Wellman, Martin Maisel Accomplishments. Initial flight testing re- Flight Experiments Branch vealed unanticipated high oscillatoryloads in Ames Research Center the XV-15 rotor control system. Subsequent (415) 604-6573/6372 analyses identified potential control system and blade modifications to reduce loads and improve structural dynamic stability. Tovali- date the analyses, variations of tip twist, blade chordwise, e.g., control stiffness, and blade sweep have been examined in flight.

The most significant factor in the reduction of loads tested to date is the increase of the cyclic swashplate stiffness, accomplishedby a tem- porary modification,which displaced the natu- ral fkequency of the system sdiciently far &om the rotor 3-per-rev excitation to enable evaluation of the ATB’s performanceand loads in all operational modes. Also, improvements to the Stability and Control Augmentation System (SCAS) have been installed for flight evaluations.

S i g n i f i c a n c e . The measured XV-l5/ATB control loads Bight investigationrevealed limi- tatiops and deficiencies in current analytical and design methods. The flight data provides Lidt=317 LB 50 60 70 80 90 30 4 0 Velocity (knots) F’igum A W-l!YATB colktive actuator loads correlation Using

a modified version of CAMRAD/JA - Helicopter Mode, 35 to 80 KTS

Low Speed (35 K’IS), M%R Flight Data -.-.-.-.-.-- uniform M o w _.-.-._.-.-. Prescribedwake --- ^ - - Free Wake .___..-.-..._.

----____ -.

-.

Azimuth (de@ Figure B XV-1WATB Normal bending moment correlation Using a m a d version of CANRAWJA showing VBIpiEltion i n correlation for different

wake models - Helicopter Mode, MSL, 35 KTS

Figure 6-8. Analysis of Tiltrotor Phenomena Utilizing CANRAD/JA ,

6-8 Analysis of Tiltrotor Phenomena Utilizing C

Objective. To evaluate the accuracies of the Extensive correlation results have been ob- state-of-the-art comprehensive rotorcraft tained utilizing CAMRAD to predict airloads analysis codes, such as CAMRAD/JA (Com- for the AH-1G and correlating them with AH- prehensive Analytical Model of Rotorcraft 16 airload measurements.

DynamicdJohnson Aero- Aerodynamics and nautics), in the prediction of rotor system A major effort was also undertaken to up- performance, loads, and stability and to iden- grade the UH-60 CAMRAD model through tify and address sources of inaccuracies to improvedinput and input source d m e n t a - enhance the analysis and design of advanced tion as requested by UH-60 PEER Review.

rotor systems.

S i g n i f i c a n c e . Improvements in correlation Approach. Extensive correlation effortsiden- will strengthen the validity of the optimiza- tified the limitations of the current state-of- tion work and lead to a realistic rotor blade the-art in comprehensive rotorcraft analysis design for advanced helicopters and high- codes. Improvements in the comprehensive speed tiltrotor configurations. Improvements codes will be judged by respective improve- in performance and aeroelasticstability must ments in correlation. Advanced high-speed be achieved to enhance the feasibility and tiltrotor blade optimization efforts also have various measures of aircraft efficiency for been initiated using a rigorous two-point op- rotorcraft at high speeds.

timization of the XV- 15/ATB(AdvancedTech- nology Blades) rotor system. The validity of Status/Plans. Further aeroelastic stability the results of the optimization effort will be and airloads correlation efforts are planned strengthened by improvementsin correlation for the XV-l5/ATBs, the AH-1G Cobra, and of the current state-of-the art comprehensive the UH-60 Phase I1 program to ensure that codes. inaccuracies are identified and comprehen- sive rotorcraft analysis codes are improved Accomplishments. Extensive correlation and reflected in the advanced high-speed efforts have been performed and reported tiltrotor blade optimization effort.

using predictions generated from a modified version of CAMRAD/JA and XV- 15/ATBrotor J. J. Totah, J. F. Madden I11 correlated with control system loads flight Rotorcraft Technology Branch data, as shown in Figure A. This effort iden- Ames Research Center tified the importance of accounting for control (415) 604-4126 system flexibility. The modified version of CAMRADJAreflectedthe flexibility,and this resulted in improved correlation. This modi- fied version of CAMRADIJA is currently be- ing used to generate results for the advanced high-speedtiltrotor blade optimizationeffort.

The predictedperformance and stability gains resulting from this effort will reflect the im- proved accuraciesobtained with the modified version of CAMRADIJA.

, Predicted acoustics Measured acoustics 40 I 20 20 Acoustic pressure, 10 Pa 0 0 -1 0 -1 0 -20 t . - . . ' . . . - ' . - . . . . . ' -20

0.0 ' 0.5 1 .o

0.0 0.5 1 .o

Timelrotor revolution Timelrotor revolution Figure 6-9. Comparison of Measured and Predicted Acoustic Signal for a Model Helicopter I at a Strong Blade-Vortex Interaction Noise Condition

6-9 Blade-Vortex Interaction Noise Prediction Validation

Objective. To understand rotorcraft noise blade variability in the measured data. The source mechanisms and to develop and vali- major features of the measured data are fairly well reproduced in both shape and magni- date noise prediction codes, specifically to tude.

validate a Langley-developed noise predic- tion code for blade-vortex interaction (BVI) Significance. The rotor noise prediction noise usingmeasured rotor blade surface pres- code WOPWOP can predict BVI noise given sures.

accurate high resolution blade loads data as Approach. Far-field noise and blade pres- input. This is significant because noise pre- sure data obtained from a model rotor test dictions made with predicted blade loadshave to validate the rotor noise not been in good agreement with measured were employed noise data. The recently obtained acoustic prediction code, WOPWOP, for BVI noise.

and blade pressure data will add immensely The test was a cooperative effort between to the understanding and development of pre- Sikorsky Aircraft, the United Technologies dictions for both rotor blade loads and acous- Research Center (UTRC),NASA Langley and tics.

Ames Research Centers, and the U.S. Army Aeroflightdynamics Directorate. The model Status/Plans. Noise predictions for avariety rotor was a 116 scale, 4-bladed, swept-tip of operatingconditions will be made and com- design. The rotor blades were instrumenkd pared with measured data. In addition, noise on both the upper and lower surfaceswith 176 predictions from predicted blade loads will be recess-mounted pressure transducers. This with measured results.

large number of transducers, combined with a made and compared high data digitizationrate, resulted in a high The predicted and measured blade loads will resolution aerodynamic database. The blade be compared and deficienciesin the predicted pressure resolution was sufficient to capture blade loads will be identified. This will allow short time events, such as BVI. Noise predic- for the development of an improved blade loads prediction and thus a better noise pre- tions were made using the measured blade pressures as input into the noise prediction diction capability.

code WOPWOP. The predicted noise time Casey L. Burley histories were then compared with the mea- Acoustics Division sured time histories.

Langley Research Center (804) 864-3659 homplishments. Measuredand predicted acoustic time histories have been compared for a number of rotor operating conditions and measurement locations. A comparison be- tween predicted and measured results for a BVI case at a moderate flight speed of 85 knots is presented in the figure. The top figure illustrates the microphone position under the flight path, and the lower figures show the measured and predicted acoustic results. The results are presented for one- rotor period and clearly show the impulsive BVI noise f h m each of the four blades. The coniparison is good, allowing for blade-to-

otor Disk Loading

l o

1 oo

resolution

resolution

e 6-10. Comparison of Predicted Rotor Blade Loading for 10" Azimuthal , Resolution and 1" Resolution

6-10 High Resolution Rotor Blade Loads for Blade-Vbrtex

Interaction Noise Prediction

Objective. At present, the missing element applied successfully in support of an experi- in providing Langley with the first fully ana- mental effort to reduce BVI noise through the lytical predictioncapabilityfor helicoptermain use of higher harmonic control (HHC)ofblade rotor noise is the specification of detailed pitch. After incorporating HHC capability blade loading. A study was undertaken to into the code, the effect of HHC on BVI noise provide this detail, through a unique recoding levels was accurately predicted.

and post-process modeling of the rotor perfor- mance code Comprehensive AnalyticalModel Significance. The effort has produced a of Rotorcraf't Aerodynamics and Dynamics more sophisticated and versatile prediction (CAMRAD). capability for the harmonic loading and im- pulsive BVI noise. The long experience of Approach. The accurate prediction ofimpul- government and industry researchers in the sive blade-vortex interaction (BVI) noise, as use of CAMRAD for aerodynamic and well as the prediction of low frequency loading aeroelastic calculations makes a high resolu- noise when BVI occurs, requires accurate tion version particularly attractive.

blade loading definition at azimuthal steps of 1" or less, as well as small radial blade incre- StatudPlans. Efforts are under way to corn- ments. The well-known and often used per- plete the loads definition by installing a near formance code C A M W has sophisticated as well as an advanced unsteady wake model, aerodynamicsmodel. Detailed noise directiv- fkee-wake and aeroelastic blade modeling, but it has a 15" azimuthal resolution limita- ity predictions are to be obtained by using the tion. The approach taken to obtain a high high-resolution loading calculations in resolution version of CAMRAD was to first Langley's WOPWOP rotor noise code, and increase the code resolution version of these will be compared with existing acoustic CAMRAD to the intrinsic limit (found to be data sets. The code will continue to be applied 10" azimuth). Post-processing codes were in support of the noise reduction research on higher harmonic control.

developed to determine blade positions and tip vortex strengths and trajectories at very small incremental steps of 1". At each incre- Thomas F. Brooks mental step, the induced velocities and blade Acoustics Division sectional loadings were calculated. Langley Research Center (804) 864-3634 ents. The CAMRAD code modifkations and the post-processing code for the free wake and first order unsteady aerodynamics have been completed. An ap- plication of the code is shown in the figure, where the rotor disk loading, the lift encoun- tered by the blades as they rotate, is plotted.

Compared with the 10" resolution results, the 1" high-resolution results offer an order of magnitude increase in the level of detail and milerstandingofthe blade-wakeinteractions.

The loading reveals the tip vortex positions and the locations where they pass through the rotor ,fisk. The high-resolution code has been Observation a , r 1.5 ~ w

.I -

0.5 Q Q) -0.5 > .I CI I I I I I I II m -1.5

E

a , r 1.5 w .I I 0.5 m Q) -0.5 ..I w

- m -1.5

.I CI II,,,,

- Q -1.5

a , 0.0 0.2 0.4 0.6 0.8 1.0 K

Re I at ive ti me

Figure 6-11. Predicted Effects of Fuselage Scattering of an Impulsive Rotor Noise Signal ,

6-11 Ray-AcousticsApproach to Fuselage Scattering of

Rotor Noise

Objective. To predict the acoustic field o f there is no fuselage, and a single impulse is installed rotors and propellers. The specific seen. For the case of a sphere, the initial objective of the present research is to develop impulse is received cleanly, and is followed by a numerical method for the aircraft fuselage a smaller, broader pulse due to scattering scattering of noise from a rotor. from the sphere. Where the sphere is re- placed by an ellipsoid, a larger amplitude Approach. A numerical implementation of scattered pulse is seen. These results are ray acoustics theory in conjunction with a highly dependent on observer position in ad- paraxial ray approximation is used to model dition to the shape of the scattering body.

the scatteringof an impulsive noise sourceby a fuselage. This is a two-point ray tracing Significance. Acoustic scatteringfrom wing problem, where both the source position and and fuselage surfaces for full aircraft configu- observer location are known in space. This is rations is always a concern in experimental modeled as a boundary value problem that aeroacoustics. In addition, model supports, requires an iterative solution, and is very microphone struts, and other wind tunnel expensive e0 solve computationally. In order hardware can cause reflection and scattering to improve numerical efficiency, paraxial ray problems. This technique can be used to help approximation (PRA) is used. This is devel- understand measured acoustic results, and oped from a Gaussian beam method (GBM), especiallyto explain differencesbetweenmea- where Gaussian beams are high-frequency sured results and theoretical predictions, as asymptotic time-harmonic solutions to the predictions usually do not include scattering wave equation, and are concentratedclose to surfaces.

rays. A GBM solution requires no iteration, but does involve a superposition of a large Status/Plans. A “user-friendly”version of number of beams to compute the noise at a the numerical implementation of this theory given observer location. The PRA evaluates is being developed. This w i l l be a useful the time-travel ray path and amplitudes both diagnostic tool to the experimental acousti- along the ray and in its vicinity. So a single cian, who will be able to define the specific ray can be used to obtain the solution at the geometrical setup of a particular test.

observer,providing it passes sufficientlyclose Michael A. Marcolini to the observer position.

Acoustics Division Accomplishments. An example of the fuse- Langley Research Center lage scatteringresult is given in the figure for (804) 864-3629 rotor impulsive noise source modeled as a point load rotating in a plane above the scat- tering body. The geometry of the model is shownin the figure. The point loadingrotates at angular velocity R, on a disk of radius a, at a height H above the body. A spherical body of radius b is depicted in the figure. Numeri- cal results are also shown in the figure for three different scattering bodies at an ob- server angle 8 of 45*, b/a=0.5, Wa=0.5, and a soyce Mach number (R dspeed of sound) of 0.7: The result for a fkee-field case, where

Measurement plane

No HHC With HHC

Upstream

/.

-Toyvns,treqm I

I I I I

12 =I 0 1 2

-2 =I 0 1 2

Cross flow direction, m

Figure 6-12. Comparison of Mid-Frequency Noise Contours for the BO-105 Rotor s in Normal Flight and With Higher Harmonic Control

6-12 Rotor Impulsive Noise Reduction Using Higher

Harmonic Control Objective. Impulsive blade-vortex interac- tive levels and directivity can be modified by tion (BVI) noise is one of the most objection- changing the area of the rotor disk in which able types of helicopter noise. A previous the HHC is employed.

rotor noise experiment in the Langley Tran- sonic Dynamics Tunnel (TDT) proved that Significance. This database is the first higher harmonic control (HHC) of blade pitch available on BVI noise radiation and directiv- could reduce overall sound power. This ex- ity using HHC. The results support the key periment was undertaken to obtain the effect conclusion of the previous Langley test that of HHC on the directivity of BVI noise. HHC can be used to reduce BVI noise for descendingflight conditions where BVI noise Approach. The test, conducted in the ane- is most important. These results are doubly choic German-Dutch Wind Tunnel (DNW), significant because of the two different rotor was a cooperative effort between NASA Lan- types used in the investigations. The present gley, the German aerospace research estab- rotor is hingeless with dynamically scaled lishment DLR, and the European helicopter blades, while the rotor tested in the TDT was companies MBB and Aerospatiale. A large fully articulated with much stiffer blades.

database was acquired for a 40% scale model of the MBB BO-105 main rotor. Noise and Statufllans. Analysis is continuing to es- vibration measurements were made for a tablish more fully the BVI noise reduction range of operating conditions where higher benefits versus operational and HHC param- harmonic blade pitch was superimposed on eter variations. In addition,the vibration and the normal collective and cyclic trim pitch. low frequency noise data will be used to help HHC control modes of 3 per rev (3P), 4P, and establish the practicality of particular con- 5P at various amplitudes and phases were trols. The results will be compared with tested, as well as special mixed 3P/4P/5P acoustic and vibratory predictions based on modes. The noise was measured over a large an improved high-resolution loading version measurement plane underneath the model of the CAMRAD rotor performance program rotor using a traversing in-flow microphone currently under development at Langley.

array.

Thomas F. Brooks Accomplishments. It was found that the Acoustics Division BVI noise directivity could be significantly Langley %search Center modified by using HHC. The figure presents (804) 864-3634 BVI noise contour plots over the measure- ment plane for a particular descent flight condition, advanceratioof0.15 and an equiva- lent descent angle of 6", which is considered typical for full-scale helicopters. On the left, a baseline case is shown where HHC is not used. The levels given are mid-frequency values and represent the BVI portion of the total noise. The plot to the right is the noise contour for a particular HHC pitch schedule which causes a significant 6 dB reduction in the advancing side lobe levels, although the retseating side increases by 2 dB. The rela- Figure 6-13. Joint ArmyBoeing Research Program ,

6-13 CooperativeArmy/NASA/BoeingPressure-Instrumented

Rotor Program

Objective. To (1) investigate the basic for- sure blade loading without expensive and mation of the rotor wake directly behind the troublesome pressure transducers in the ro- advancingblade,(2)validate acirculationbox tor blades.

technique for measuring rotor loads, and (3) capture a blade-vortexinteraction occurrence Status/Plans. The pressure data are being in the rotor wake. digitized and reduced at k i n g Helicopters.

The laser velocimeter data will be processed at Langley. Ajoint publication is anticipated Approach. A wind tunnel test was con- ducted in the 14- by 22-foot subsonic tunnel next year. Correlation with analysis will be using the Boeing Single Rotor Helicopter done both at Boeing with their proprietary B- (SRH) drive system, a pressure-instrumented 65 code and at Langley using CAMRAD-JA model of the Boeing 360 rotor and a two- with FPR coupling.

componentlaser velocimeter. Measurements of the dynamic blade surface pressures and Susan L. Althoff the flow-fieldvelocities were obtained for sev- Subsonic Aerodynamics Branch eral flight conditions. Langley Research Center (804) 864-5059 Accomplishments. Measurements of the blade surface pressures using the 66 trans- ducers installed in the four rotor blades and velocity measurements using the laser were made for nine flight conditions ranging from high-speed forward flight to descent condi- tions. Velocity data were obtained for over 1600 flow-field measurement locations. In all cases, average pressure and velocity data as well as time-dependent data were acquired.

Significance. These data will be used to correlatesurfacepressure measurements with wake velocity measurements to determine the nature of the forcing function for rotor blade airloads. These data also provide a link between the acousticmeasurements obtained in the German Dutch Wind Tunnel (DNW) and the flow-field phenomena which gener- ates rotor noise. The data identifying the formation of the rotor tip vortex and wake will be used to improve analytical predictions of rotor wakes. The velocity data around the blade in a box will be used to validate the circulation measurement technique, and, if successful, will lead to the capability to mea- , Full- and Small-scale, p = 0.20, CT = 0.055, aB = 0" Dlfferential pressure coefficlent I a \ (AcP)R

Hub Q

-4 t I

I I I I 1 I 1 I I I

.5 '

0 .2 .4 .6 .a 1 .o

Normalized distance along body (X/L) Figure 6-14. Effect of Rotor on Upper Surface Pressures

6-14 RotorE’uselageAerodynamic Interactions Program

Objective. To quantify rotodfuselage aero- stat . Plans are to complete the

dynamic and acoustic interactions through documentation of the full-scale wind tunnel wind tunnel testing and computation. The test results.

program results will assist in the design of rotorcraft systems by taking advantage of ThomasR.Normanand Rotorcraft Aeromechanics Branch mutuallybeneficial aerodynamicinteractions.

Ames Research Center Approach. To achieve the program objec- (415) 604-6653 tive, a fid-scale test of a Bell 412 rotor system with a simple axisymmetricfuselagewas con- ducted in the 40- by 80-foot wind tunnel at Ames. Small-scale tests with similar fuselage shapes have been conducted in the past.

Accomplishments. Data from the 40- by 80- Foot Wind Tunnel test has been reduced and analyzed. A paper based on the test results was presented at the 1991American Helicop- ter Society Forum.

Significance. Results from the wind tunnel test revealed that the magnitude of the aero- dynamic interactions on the simple fuselage were small, The importanceof rotor shaft and hub interactions was also shown. In addition, comparisons between full-scale and small- scale results demonstrated the importance of model configuration and installation differ- ences. For example, the figure shows a com- parison of the rotor effect on the upper surface pressures for full- and small-scale data. In the figure, (ACp)? represents the difference in pressure coeffiaent that is due to the rotor.

The large differences between the full-scale and V6-scale results are attributed mainly to the lack of a rotor mast in the l/6-scale test.

The reason for the offset between the full- scale and 0.15-scale data is unclear, although potential causes include the unscaled hub and significantly different model mounting system used in the 0.15-scale test.

, RWF code output (LaRC-Berry) Figure 6-15. Rotor-Wake-Fuselage Code Development , 6-15 Rotor+FITake-FuselageCode Development Objective. To predict the aerodynamicenvi- code is under modification for production op- ronment of a realistic helicopter configura- eration on the Langley SupercomputerNet- tion which accounts for the interaction of the work System (SNS). The code is being used to rotor wake and the helicopter fuselage.

assess h l a g e effects on rotor inflow.

Approach. This program has been an in- Significance. The computer code RWF is a house development effort. Existing code ele- valuable tool for investigating of realistic fu- menta for computing the wake &om a single selage effects on helicopter configurations.

rotor blade and the influence of a nonlifting source panel have been integrated into a code Status/plans. After validation of the SNS that computes the interacting flow field of production version of the RWF code, the code multiple rotor blades with cyclic pitch, their will be made available to interested U.S. us- shed wakes, and a nonlifting fuselage. The ers. A low level of continuing development is resulting computer code Rotor-Wake-Fuse- planned to respond to requirements of the lage (RWF) has been under preliminary de- target user community.

velopment since 1986.

John D. Berry Accomplishments. During F Y 1991, pre- Subsonic Aerodynamics Branch liminary copies of the implementing code Langley Research Center (RWF) have been provided to two requesting (804) 864-5090 U.S. helicopter companies. A version of the ,

------I

w w U L L t ,

a

I u) 8 I I I I n * C ) m u r

t

a

X u) Figure 6-16. University of MarylandAmes Research Center Bearingless Rotor , Aeroelastic Analysis

6-16 Five-BladedBearingless Rotor Program

Objective. To document the performance, rotor wind tunnel test data for shaft fixed and loads, and stability of five-bladedbearingless shaft free (see figure) have shown good agree- rotors, and to develop and validate csmpre- ment.

hensive analyses to accurately predict and design new rotor configurations. StatudPlans. Wind tunnel test planning is continuing for both the MDAR and SBMR.

Approach. Testing of the McDonnell Dou- Development and validation of UMARC con- glas Helicopter Company (MDHC)Advanced tinues as data becomes available.

Bearingless Rotor (MDAR) and the Sikorsky Five-Bladed Bearingless Main Rotor (SBMR) Steve Jacklin was conducted in the 40- by 80-foot wind Rotorcraft Aeromechanics Branch tunnel through the flight envelope up to 2Q0 Ames Research Center knots. A comprehensive database on rotor (415) 604-4567 performance,blade, flexbeam and torque tube loads, and aeroelastic stability was acquired.

The results of these test programs will assist in the development and validatiosl of the Uni- versity of Maryland Advahced Rotorcraft Code (UMARC).

Accomplishments. A joint cooperative pro- gram was established for the MDAR tests.

The first test was scheduled for the fourth quarter of CY 1991 in the 40 x 80 on the MDHC test stand. The second phase of the program with a pressure instrumented rotor has been initiated. A joint cooperative pro- gram was established with Sikorsky for test- ing of the SBMR in’the 40 x 80 on the Rotor Test Apparatus in the second quarter of CY 1992. The UMARC Code has been clocu- mented and correlation is in progress for the SBMR stability data. The MDHC HARP rotor has been modeled and the MDAR modeling is awaiting data release from MDHC.

Significance. The correlation efforts with UMARCand experimentalstability data have been encouraging. Correlation with the SBMR hover stability data has been good. Correla- tion with a four-bladedmodel scalebearingless BVI loading calculations A K l G Cobra (V=82 knots, r/R-0.970) 0.2 flight test,averaged - - - calculation,core=O.OlSR - _ _ _ _ _ calculation,core=0.030R 0.16 ._.-__...._.

calculation,core=0.040R

2 0.12

h v 0.08 U

-

0.04 .- c , Q) m O -0.04 0 90 180 270 360 azimuth, deg dLACK A tor Airloads Correlation Using CFDLifting-Line

6-17 Rotor Airloads Correlation Using CFDLifting-LineMethods

Objective. To assess the predicting capabil- S i g n i f i c a n c e . The accurate prediction of ity of current state-of-the-art rotorcraft aero- rotor airloads is one of the biggest challenges dynamic codes on modern rotors by compar- in the field of theoretical aerodynamics and ing results with existing flight test data.

one which remains to be solved. Current rotorcraft aerodynamic codes make use of Approach. The comprehensive rotorcraft, many approximations and assumptions that code, CAMRAD/JA, was used to model the ultimately must be replaced by more accurate aerodynamic and dynamics of the complete methods. To determine the validity of such rotorcraft configuration. To obtain the de- methods and improve upon them, correlation tailed blade airloads, the Full Potential Rotor work with experimentaldata is needed. Such Code (FPR)was coupled with CAMR,AD/JAin improved codes will help in the design of an iterative loop. The final solution was the better rotor systems.

predicted airloads on the rotor system which were then compared with Plight measure- StatudPlans. CFD methods are currently ments at different operating conditions. being used to model the UH-60 Blackhawk rotor system from hover to high advance ra- Accomplishments. A detailed airloads cor- tios. Preliminary results have been compared relation effort on an AH-1G Cobra helicopter with model wind tunnel data at different has been completed. Computationalmethods flight conditions. These methods will help in were used to calculate the blade airloads and the prediction of rotor airloads during Phase give insight into the nature of blade vortex I1 of the UH-60 Airloads Program.

interaction loads at low advance ratios. Fig- ure A shows the correlation between calcu- Francisco J. Hernandez lated measured blade section lift during blade Rotorcraft Technology Branch vortex interaction encounters. As shown in Ames Research Center figure, the section lift was calculated for (415) 604-1322 the several vortex core sizes. The effect was minimal for this case, implying the vortex did not pass near the blade. At high advance ratios, the same approach was used to predict the transonic compressibilityeffects that are present on the advancing side of the rotor disk. The use of computationalfluid dynam- ics (CFD) allowed both a quantitative and qualitative analysis of the flow phenomenain a very efficient manner. Figure B shows predicted Mach contours at a high advance ratio where transonic flow is present. Also, computational fluid dynamics methods have been used to predict the rotor airloads on the UH-60 Blackhawk and to compare them with egstingexperimental data from the German- Dutch Wind Tunnel (DNW). This CFD model is being developed in preparation for Phase I1 of the UH-60 Airloads Program, currently under way at Ames.

e 6-18. Roto Flow Field Investigation

6-18 Rotor Flow Field Investigation

Objective. To provide a comprehensive da- lage interaction code validation. The flow tabase documenting the velocity flow field visualization work provides the beginning of generated by a thrusting helicopter rotor in a visual database of the rotor vortex wake forward flight. This database is intended to structure.

provide a means to validate the accuracy of the velocity and rotor load predictions of rotor Status/Plans: The data obtained in this computational methods.

program will be used in-house for correlation with rotor performance codes and rotor-fuse- Approach. A wind tunnel test was con- lage interaction codes. The flow visualization ducted in the 14- by 22-foot subsonic tunnel video data will be digitized and enhanced for using the 2-Meter Rotor Test System and a further analysis by image processing soft- two-component laser velocimeter to measure ware.

the velocity induced by the rotor above the plane of the rotor blade tips. Measurements Susan L. Althoff, Joe W. Elliott were also made in two planes through the Subsonic Aerodynamics Branch separated flow region behind the hub. A Langley Research Center second laser was used to generate a light (804) 864-5059 sheet to provide flow visualization of the vor- tex structure in the rotor wake.

Accomplishments. Measurements of the rotor-inducedvelocity were made at 180 loca- tions i n planes above the rotor blade tips at a height of 0.50 chord for an advance ratio (ratio of freestream velocity to rotor tip speed) of 0.23. Measurementswith and without blades were made in the region of separated flow behind the rotor hub at 15 longitudinal sta- tions, with 15 points at each station. In all cases, average velocity data as well as time- dependent data were acquired. Flow visual- ization studies were conducted for the rotor wake region above the rotor disk and down- stream of the rotor for advance ratios of 0.15, 0 . 2 3 , and 0.30.

Significance. The rotor inflow measure- ments, which were obtained during this test program, demonstrate the first production use of the newly developed Frequency Do- main Processors (FDPs) for the laser veloci- meter data acquisition system. The hub sepa- ration measurements provide the basis to define the extent of the separated flow behind the rotor hub for computational rotor-fuse- , ,

Chapter 7

FightedAttack Aircraft

The objective of the FightedAttack Aircraft Program is to provide t h e required technologies for fighter/attack aircraft to achieve efficient, sustained supersonic cruise and maneuver perfor- mance; efficient store carriage and deployment at supersonic speeds; increased agility at subsonic speeds and acceptablehandling qualities at extreme angles-of-attack;and short takeoff and vertical landing (STOVL) operation.

Improved prediction methods and/or experimental techniques are now available for high-lift aerodynamics, propulsion integration, weapon carriage, supersonic store cavity and separation aeroacoustics, integrated flight controls, and systems design for fightedattack aircraft. Wind tunnel and piloted simulation studies have demonstratedthe potential effectiveness of multiads thrust vectoring for propulsive control at extreme angles-of-attack. In addition, powered lift systems consistent with the operation of advanced STOVL aircraft, have been identified, and configurationstudies have been conducted to assess the impact of integrating these systemswith supersonic airframe designs.

The FightedAttack Aircraft Program is focused on (1) CFD modeling and validation of cavity effects and near-field trajectory simulation for weapons launch during maneuver and modeling of 3-D flow fields for vehicles at high angles-of-attackwith separated flow; (2) development of innovative propulsive and aerodynamic control systems concepts to provide increased vehicle control at angles-of-attacknear and beyond maximum lift; (3) improved understanding of control requirements for control throughout the relevant flight envelope; (4) large-scale,ground-based testing of a STOVL fighter concept using ejector liftlvectored thrust to define critical transition aerodynamics; (5) studies of STOVL ground effects including inlet reingestion and ground erosion; and (6) integration studies to quantify the impact of emerging technologies and define the critical research areas for supersonic STOVL.

Program Manager: Benjamin Neumann OASTiRF Washington, DC 20546 (202) 453-2795 ORlGlNAL PAGE BLACK AND WHITE PHOTOGRAPh F/A-18 Aircraft Angle of Attack, degrees Comparison of small-scale and full-scale tail buffet frequency versus angle of attack. Experimental for fluctuating pressure measurements at a single point on the vertical tail while the results are CF'D point is for the unsteady root bending moment on the tail.

Figure 7-1. Tail Buffet Frequency vs Angle-of-Attack %

7-1 Full-scale %st of F/A-18 at HighAngles-of-Attack

Objective. To examine the aerodynamics of tem can provide more yawing moment than a high angle-of-attack flight of fighter aircraft. full rudder deflection does at low angles-of- Particular emphasis is placed on understand- attack.

ing the ][LEX-vortex (LeadingEdge Extension) induced t a i l buffet and on developing a pneu- Significance. The pressure distributions matic forebody flow control concept for lateral measured on the full-scalemodel will provide control at high angle-of-attack. The test has a rich source of CFD validation data, which also provided important data for computa- will help toimprove the accuracyof 3-DNavier- tional fluid dynamics (CFD) validation and Stokes computations. Current CFD efforts are ongoing to compute the unsteady burst- resolution of scaling issues, which have be- come apparent in comparingsmall-scalewind- vortex induced pressures on the F-18 tail.

The wind tunnel measurements will also be tunnel data with flight results.

used in the development and validation of Approach. A retired Navy F/A-18A has been CFD for unsteady flows.

converted into a wind-tunnel model for test- ing in the 80- by 120-foot wind tunnel. The It has been demonstrated that pneumatic model is heavily instrumented to measure forebody flow control can generate control yawing moments of sufficient magnitude to time-averagedforebodypressure distributions provide directional control at extreme angles- and the unsteady pressure distribution on one of the vertical tails. In addition, the nose of-attack where rudder control is no longer of the aircraft has been modified with the effective. These full-scale data are critical in addition of jet-type and slot-type blowing for order to define the planned flight experiment of pneumatic forebody flow control system on flow control at very high angle-of-attack.

the High Alpha %search Vehicle, HARV.

Accomplishments. Preliminary data from the wind-tunnel tests matched very well with StatusiPlans. A second wind-tunnel test flight data for overall forces and moments as will be conducted in 1992 to optimize the well as for the radome and forebody pressure forebody flow control system in preparation for flight testing, extend the surface pressure distributions. The unsteady vertical tail pres- sure data will be used to address wind tunnel database to include the wings for use in CFD scaling issues, to identify flow structures and validation and correlation with flight, and to refine the measurement of unsteady pres- dominant frequenciesin the burst vortex, and sures on the vertical tails of the aircraft.

to validate CFD analyses. Preliminary analy- sis of these data indicates that the frequency content of the fin pressure measured in the James C. Ross full-scale wind-tunnel test is close to that Fixed-WingAerodynamics Branch observed in several small-scale experiments.

Ames Research Center Similar pressure measurements are not yet (415) 604-6722 available from the flight experiment.

The pneumatic forebody flow control has been shown to be effective for practical blowing slot lengths and blowing rates. At a given angle- of-attack, the yawing moment is linear over a wide range of blowing mass flow coefficient.

At a 50" angle-of-attack, the pneumatic sys-

. Tail Buffet Research

7-2 Tail B d e t Research

Objective. To study the effects of vortical aircraft's configuration can have a signiscant flow fields on tail buffeting of modern high- impact onits tail buffet characteristics. Thus, performance aircraft, and explore methods of consideration of tail bdeting should be ad- reducing buffet levels while maintaining the dressed early in the design stage of an air- favorable effects of vortex flow on the high craft, when configuration changes are still angle-of-attack aerodynamic characteristics. possible. Two issues to consider are: (1) re- gions of stalled or separated wake flow can be Approach. A cooperative program was es- a mqjor contributor to tail buffet; and (2 ) tablished between NASA, the Navy, and aerodynamic/stab~ityabilitypadcanIbe due more McDonnell Aircraft Company to conduct low- to interactions of the vortex flow field with speed wind tunnel tests on a 0.16-scale F/A-18 other aircraft surfaces than to changes in the model with dynamically scaled and instm- vortex flow field itself.

mentedvertical tails. This research provided data on the unsteady aerodynamic loadings Also, successfulcorrelationwith previous data and surface pressures of the tails. Changes to using scaling parameters suggests that the the configurationaimed at modifying the vor- same approach may be used to develop wind tical flow field were tested to investigate the tunnel-to-flightscaling methods.

effect on both buffeting response levels on the tails and the overall aerodynamic character- StatudPlans. Follow-on investigations of istics of the configuration. Additional tests various innovative concepts to alleviate tail were conducted to obtain detailed aerody- buffeting are planned. Correlation of results namic information on the impact of the modi- with research programs of other organiza- fications. on t a i l tions is planned through a workshop buffet research in August, 1991, for Aero- Accomplishments. Testing in the Langley nautics Technology Technical Panel HTP-5 30- by 60-foot tunnel provided a large data- members of The Technical Coordination Pro- base of dynamic response characteristics of gram. Results from current and future wind the vertical tails and steady-state total force tunnel tests w i l l be correlated with airplane and moment data for several configurations. flight test data for development and valida- Correlation of measured tail buffet levels with tion of scaling methods.

previous, limited industry testing of the baseline F/A-18 configuration was very good. Gautam €3. Shah The study showed that repositioning and/or Flight Dynamics Branch weakening of the primary vortex system can Langley Research Center reduce t a i l buffet levels but can also have a (804) 864-1163 large impact on the overall lifi and stability characteristics. In addition, substantial amounts of tail buffeting were seen due to hselage and wing wake flow, in the absence of avortex system. Follow-on aerodynamic-only testing indicated that many of the stability effects were caused by changes in interactions between the vortex system and tail surfaces.

Significance. The results of this test indi- cate' that relatively minor changes in an S

z

L .-

-- a , c

E

.-

t

=t - 0 9- L h, [L 0 t : Figure 7-3. Typical Nose-Down Control Capability Characteristics for Relaxed Stability Combat Aircraft

7-3 Development of High-Angle-of-AttackNose-Down Pitch Con-

trol Requirements for Relaxed Static Stability CombatAircraft

Objective. To develop high-angle-of-attack cept requires design guidelines for defining nose-down control requirements that can be high-angle-of-attack nose-down control re- used as design guidelines early in the devel- quirements early in the development process.

opment cycle of advanced combat aircraft The output from this study is being used by incorporatingrelaxed static stability in pitch. the Navy to d e h e these requirements for design and flight testing of the F-18 Em. The Approach. A cooperative program was es- Air Force is also interested in this study, both tablished between NASA and the U.S. Navy in terms of potential application to Advanced to perform the research. A systematicpiloted Tactical Fighter and for development of hand- simulation study was conducted on the Lan- book criteria.

gley Differential Maneuvering Simulator to Status/Plans. A preliminary set of guide- develop and assess candidate guidelines. The simulation process involved specific assess- lines and flight demonstration requirements ment maneuvers and pilot rating approaches has been developed based on the results ob- developed for this application. A key element tained to date. Simulation tests w i l l be con- of the investigation was a parametric study ducted to verify that the design guidelines involving systematic variations of critical are valid for complex maneuvering and to parameters aecting nose-downcontrolcapa- prepare for full-scale flight tests. A limited flight evaluatiodvalidationprogram on atest bility, particularly the minimum level avail- able at high angles-of-attack (Cm*, see ac- F-18 airplane will be conducted by the Navy companying figure). Extensive involvement in September 1991. Detailed flight testingof of Navy researchers and pilots permitted ac- proposed guidelines using the NASA High celeration of the research process. Alpha Research Vehicle (HARV) is planned.

Marilyn E. Ogburn, John V. Foster Accomplishments. A comprehensive data- Flight Dynamics Branch base of piloted simulation results was gener- ated to enable the assessment and develop- Langley Research Center ment of potential guidelines. The results (804)864-1136 were based on analysis of flight motions and pilot comments and ratings obtained during the parametric study. These results show a strong correlation of pilot rating to the air- plane short-term pitch acceleration and pitch rate buildup in response to nose-down com- mands applied at high-angle-of-attackcondi- tions. Using these data, candidate design guidelines and flight demonstration require- ments were defined. Additional simulation tests were conducted to verify that these de- sign guidelines could be used to predict pilot opinion of the aircraft response.

Significance. The use of relaxed static sta- bility in pitch for performance enhancement is a key design approach for advanced combat aircraft. Successful application of this con- 1 27 nvestigation of the Tumbling Phenomenon in Aerodynamic Configurations

7-4 Investigation of the Tumbling Phenomenon in Aerodynamic

Configurations

Objective. To determine the effect of plan- Status/l?lans. The aerodynamic database form geometry and mass distribution on the will be completed. The instrumentation sys- tumbling characteristics of wings.

tem for the free-to-pitch apparatus is nearly Approach. Static and dynamic force tests complete. Analysis of the free-tumbling re- sults is continuing and the tumble simulation were conducted in the 30- by 60-foot tunnel is being refined.

and free-tumbling tests yere performed in the 20-foot Vertical Spin Tunnel using a se- Raymond D. Whipple ries of generic wing models. Each set consists Flight Dynamics Branch of twelve models with four different leading- Langley Research Center edge sweep angles and three different trail- (804) 864-1194 ing-edge sweep angles. Aerodynamic data through 360" angle-of-attack was used to de- velop a three-degree-of-freedomsimulation for calculating tumbling behavior.

Free-tumbling tests -were conducted over a range of center-oPgravity and inertial condi- tions. A free-to-pitch apparatus is being de- veloped and instrumented to enable continu- ous rotation testing of a dynamically scaled model for single degree-of-freedomanalysis.

Accomplishments. Free-tumbling results have shown that higher aspect ratio wings generally have a tumbldno tumble boundary at a more positive static margin than lower aspect ratio wings. This boundary can be shifted forward for the low-aspect ratio wings by changing the mass distribution. Initial results of the simulation show tumbling be- haviors similar to those observed in free- tumbling tests.

Significance. As fighter design trends drive toward more tumble-susceptible configura- tions, the information generated from this research will provide design guidelines to assist in configuration development.

Figure 7-5. F-106 Vortex Flap Flight Experiment

7-5 F-106 Vortex Flap Flight Experiment

Objective. To demonstrate the feasibility of StatudPlans. Testing of the aircraft was the vortex flap concept in a free-flight envi- completed in April 1991. Analysis and publi- ronment to characterize the wing flow field cation of the test results are continuing.

and to provide a database for sharp leading.

edge wings to be used for calibrationof design James B. Hallissy, and analysis tools. W. Elliott Schoonover, Jr., John E. Lamar Applied Aerodynamibs Division Approach. An F-106B aircraft was modified Langley Research Center by the addition of wedge-shaped, leading- (804) 864-2865 edge flaps to the existingwing. The wing flow field was characterized using surface pres- sure measurements and flow visualization data obtained through the use of flow cones, surface oil flows, and a light sheetlflow seed- ing system. A limited evaluation of the han- dling qualities and performance of the modi- fied aircraft relative to the baseline aircraft was conducted.

Accomplishments. Flight activities for the modified aircraft with the vortex flap at 30" and 40" deflections have been completed.

Aircraft performance data, detailed wing sur- facepressure measurements,and surfaceflow visualization were obtained at subsonic and transonic test conditions. Light-sheet and flow-seedingsystemswere installed to visual- ize the off-body flow and provided significant additional details of the wing flow-field char- acteristics at subsonic test conditions. The accompanying photograph shows the F-106 ajrcraft in flight with the vortex flap deflected 40" and a coating of oil defining surface flow patterns on the flap and wing upper surfaces.

Significance. Flight testing has demon- strated the performance enhancementpoten- tial of the vortex flap and provided consider- able insight concerningits flow field and load- ing distribution characteristics.

7-6. Generic Fighter Models Installed in the Unitary Plan Wind Tunnel ,

7-6 Wing Camber Effects on Flap Effectiveness at

Supersonic Speeds

Approach. A cooperative program was es- Gloria Hernandez tablished between NASA and General Dy- Applied Aerodynamics Division namics to perform the research. Cambered Lagley &search Center anduncambered wing models were built with (804) 864-5572 the same p l d o r m and flap geometry. The wing was designed to meet a broad spectrum of performance goals including efficient trimmed supersonic cruise pedormance and maneuver capability. The effects of camber on the flap performance were then studied through a comparison of force and moment measurements on both wings. The models were tested across the range from 1.6 to 2.16 i n test section I of the Unitary Plan Wind Tunnel. The wings were tested with leading- edge flap deflections of -4", -2", 0", 5 " , lo", and 15" and trailing-edge flap deflections of -30", -20" -lo", 0", a d 10".

Accomplishments. The experimental re- sults show that leading-edge flap deflections of up to 5" reduced drag at high lift on the flat wing. Small negative leading-edge flap de- flections reduced drag on the cambered wing at low lift. Trailing-edge flap effectiveness decreased with increasing negative flap de- flection. Wing camber had little effect on trailing-edge flap effectiveness.

Significance. This study shows that lead- ing-edge flap deflections, which are.custom- arily used at subsonic speeds, can be used to obtain drag reduction benefits at supersonic speeds. The drag produced by wing camber at supersonic speeds can also be reduced by negative deflectionsof leading-edgeflaps. The study also showed that wing camber has a negligible effect on the trim-producing trail- .

ing-edge flap effectiveness at supersonic speeds.

,

. Outdoor Static Tests of the Full-Scale Ejector-LiftNectored-Thrust

, STQVL E-7A Configuration

7-7 Outdoor Static Tests of the Full-scale Ejector-Lift/

Vectored-Thrust STOVL E-7A Configuration

Objective. To test the full-scale STOVL E- effect on the performace of the thrust-aug- 7A model at Ames Outdoor Aerodynamic Re- menting ejectors. Laser light sheet, smoke, search Facility ( O m ) to measure the thrust and infrared flow visualization techniques of the lifting ejectors and vectoring ventral have revealed that the hot ventral nozzle jet nozzle of the vehicle. The tests w i l l also reveal penetrates forward in a thin layer beneath the interactions of the plumes from the lifting the lower energy ejector jets.

jets beneath the vehicle as they impinge upon the ground. The E-7A research program is Significance. Designers of ejector-lift/vec- part of a joint technology development pro- tored-thrust STOVLconfigurations have been graminvolvingBoeingCanada’sDe Havilland workingunder the assumption that an ejector Division and NASA Ames Research Center. thrust augmentation of 1.6 was achievable at The E-7A configuration emerged from a Gen- full-scale. The level of ejector performance eral Dynamics design study of ejector-lift con- and forward-flight transition performance cepts.

during the Ames E-7A tests demonstratesthe viability of supersonic STOVLconceptswhich Approach. Static thrust calibrations of the employ lifting ejectors.

propulsion system in the STOVL E-7A model were performed in the test sections of the StatudPlans. The 1ateraVdirectionalperfor- Ames 40- by 80- and 80- by 120-foot wind mance of the E-7A model in hover mode is tunnels as part of the forward-flight tests of scheduled to be tested at low speeds and high the model. The static test of the model at the yaw angles in the 80- by 120-footwind tunnel 0AR.F will yield more accuratemeasurements beginning in June of 1992.

of static thrust because jet-induced recirculation flow patterns in the wind tunnel Tim Naumowicz, Brian E. Smith tests due to the presence of tunnel wall bound- National Full-scale Aerodynamics Complex aries w i l l not be a factor. The two wind tunnel Ames Research Center tests have provided low- and high-speed aero- (415) 604-6674 dynamic data. The outdoor test will complete the database on the full-scale model by pro- viding static thrust calibrations. The E-7A model has been mounted approximately 25 feet above ground at the OARF.

Accomplishments. Static tests of the E-7A model are complete. The measured thrust augmentation ratio of the lifting ejectors is approximately 1.6. The augmentation ratio is defined as the net ejector thrust divided by the thrust of the primary nozzles. This value is close to that measured in the test section of the 80- by 120-foot wind tunnel. Changes in the shape of the base of the fuselage between the ejector systems were shown to have an , ffect Prediction Capability for I Lift Aircraft

7-8 Validation of Out-of-GroundEffect Prediction Capability

Powered Lift Aircraft

for

Objective. To evaluate the capability of computedvortices are less diffuse than in the computational fluid dynamics (CFD) to pre- experiment, and the jet wake region is not dict the critical flow features for a ShortTake- adequately simulated. By retaining viscous O f f Vertical Landing (STOVL) aircraft con- terms in all three directions, quantitative figuration through comparison of computed agreement with experiment has been im- results with the measurements of a compan- proved. This modification is incorporated ion experiment designed for CFD validation. into the STOVL model simulation.

Approach. Previous simulations of critical Significance. A comprehensive and well- powered lift flow components such as a jet in documented set of measurements is provided crossflow, which isolates the lift jetlaerody- for CFD validation comparisons.

namic interaction during transitional flight, showedfavorable agreement with experiment. Statufllans. Computational grids are be- To extend the application of the F3D code, a ing developed for the STOVL model geom- simplified but representative STQVL aircraft etry. Computations (out-of-ground effects) configuration was selected for testing and will be completed and compared with experi- ment. Flow field measurements in and out of simulation. The model geometry, a Delta wing (E-7A) planform with twojets located in ground effects will be made using laser a blended hselage, minimizes the geometric velocime try.

modeling complexity while retaining the im- Karlin R. Roth portant physics of the lift jet/aerodynamic interaction. The unsteady Reynolds-aver- Fixed Wing Aerod,ynamics Branch aged Navier-Stokes equations applied on Ames Research Center patched or overlapped grids will be used for (415) 604-6678 the computations.

Accomplishments. The transitional flight characteristics of the STQVLaircraft configu- ration were measured in the NASA Ames 7- by 10-foot wind tunnel. The results include forces and moments, steady and unsteady surface pressures, andjet pressures and tem- peratures. Measurements of the flow were also made in the tunnel test section upstream and downstream of the model and at the jet exits to provide boundary conditions for the computations. Flow visualization and total pressure measurements in the jet plumes provide a descriptionof the three-dimensional jet eBux flow field.

Computational emphasis has been placed on establishing the numerical modeling require- menta for the propulsive flow field. Previous thin-layer Navier-Stokes computationsusing F3D ‘for a jet in crossflow indicated that the

4 = Oo, canards off

t 1.90

- 05

= o 5 15 20

Figure 7-9. Effect of Tail Fin Shaving ,

7-9 Comparison of Shaved and Beveled Fins for Rolling Missile

Applications

Objective. To investigate the effectiveness Status/Plans. Supersonic tests on this con- of shallow shaving along the leading and cept have been completed. Analysis of the trailing edges of wrap-around missile fins for data is continuing, and the results w i l l be producing roll on the configuration. published in a NASA report. Subsonidtran- sonic tests of this concept may be performed.

Approach, An experimental investigation was conducted to investigate the aerodynam- Jerry M. Allen ics of a tube-launched air-to-surface tactical Applied Aerodynamics Division missile configuration that employs planar Langley Research Center canards and wrap-around tail fins arranged (804) 864-5592 in opposing pairs. Fin shaping in the form of shallow shaving along the leading and trail- ing edges of the tail fins was investigated as a candidate for producing the rolling motion on this configurationthat is needed for controlla- bility. A second set of f i n s having a more conventional 4 5 O bevel along the tail fin lead- ing edges was investigated for comparison.

Force and moment tests on this concept were conducted in the low Mach number test sec- tion of the Langley Unitary Plan Wind Tunnel at Mach 1.60, 1.90, and 2.16 at angles-of- attack up to 20°and roll angles up to 90°.

Accomplishments, To be most effective,the rolling moments on this configuration should be constant over the anticipated flight condi- tions. The results of this experiment show that the shaved fins were much more effective than the beveled fins in producing these fa- vorable rolling moment characteristics. As shown in the figure, the shaved fins produced significantlylarger rollingmoments, and these momentswere essentiallyinvariant with both Mach number and angle-of-attack. The bev- eled fins, however, produced large variations with both parameters, in some cases even producing negative rolling moments.

Significance. Shallow shaving of the lead- ing and trailing edges of fins has been shown to be an effective roll-producing technique for missile configurations that require rolling motion for controllability.

,

a i

e S T Q ~ S T Q V ~ Concepts for High-Performance Aircraft I

7-10 STQWSTQVL Concepts for High-Performance Aircraft

Objective. Toprovide advancedaerodynamic flowfields show very good agreement. (3) The technology and advanced concepts applicable data from the generic VTOL test and the to STOLlSTOVL operations of current and isolatedjet tests in SBRT are under analysis.

future high-performance aircraft. Preliminary analysis of the effect of the mov- ing belt ground plane on the ground vortex Approach. Through experimental and ana- flow field, both in 14- by 22-FST and SBRT, lytical studies, using powered models in both show good agreement with results obtained the Subsonic Basic Research Tunnel (SBRT) using a moving jet in the Vortex Research and the 14-by 22-Foot Subsonic Tunnel (14- Facility (VRF). These results indicate signifi- by 22-FST),key problems were examined and cant changes in the extent of the ground promising solution concepts for providing vortex when the exhaust flow penetrated STOUSTOVL capabilities for advanced air- against the freestream (belt on or movingjet) craft were identified. The SBRT and 14- by or against the wall boundary layer (belt off).

22-FST provide a unique combination of re- sources for the study of takeoff and landing StatudPlans. The results of the ground aerodynamics in and out of ground effect. vortex investigations will be analyzed to de- termine the effects of wing location, nozzle Accomplishments. In FY 1991, two major arrangements, and pressure ratio on configu- tests were conducted in the 14-by 22-FST and ration aerodynamics in and out of ground the SBRT, which were made operational with effect. A detailed quantitative determination a moving belt ground plane and data system.

of the actual extent of the ground vortex w i l l A detailed investigation of augmenter perfor- be made and compared with the VRF moving mance, ground effects, stability and control, model results. The SBRT will be completed and a Laser Velocimetry(LV) map of the with a model support system to allow model augmenter inlet flow field was completed on testing and an LV system for flow measure- an advanced VTOL concept shown in the ments.

accompanyingfigure. A study of the effects of high nozzle pressure ratio (NPR) exhaust John W. Paulson, Jr. (804) 864-5071 flows; nozzle type, location, and number;vari- Guy T. Kemmerly (804) 864-5070 ous wing planforms; and moving belt ground Kevin J. Kjerstad (804)864-5022 plane on the aerodynamicsof a genericVTOL Subsonic Aerodynamics Branch configuration shown in the figure was also Applied Aerodynamics Division completed. The SBRT was used to assess the Langley Resaearch Center effect of a moving belt ground plane on the development of the ground vortex flow field from a single vertical jet in a cross flow.

Significance. These three experimentalin- vestigations have yielded the following sig- ' nificant results: (1) A method of venting the lower and upper surfaces to equalize pres- sure, which minimized the pitch-up tendency typical of configurations during transition, was demonstrated using the advanced VTOL configuration. (2)Comparisons ofthe compu- tational and LV measured augmenter inlet Figure 7-11. Effect of Canard on Wing-Body Aerodynamics ,

7-11 Effect of a Close-Coupled Canard on Wing-Body

Aerodynamics

Objective. T o accurately simulate the aero- Si e . Utilizingcanardsforenhanced dynamics of a canard-wing-body configura- aircraft performance requires a thorough tion and investigate the characteristicsof the understandingoftheassociatedmmplexaero- canard-wing vortex interaction and break- dynamics. Accurately predicting the nonlin- down. ear effects of the canard, including the ca- nard-wing vortexinteractionand wing vortex Approach. Using an extension of the Tran- breakdown characteristics, improves the ap- sonic Navier-Stokes (TNS) d e , Reynolds- plicability of computational f l u i d d p a m i a averaged thin-layer Navier-Stokes equations towards the design and optimization of ca- were solved for the flow about a canard-wing- nard-configured aircraft.

body configuration. Flow wasconsideredfblly turbulent and turbulence effects were mod- StatWlane. The effects of canard deflec- eled using the Baldwin-bmax algebraiceddy tion, positioning, and size on wing-body a e m viscosity model. The surface and flow-field dynamics are currently under investigation.

grids were generated using the S3D code and A study of unsteady canard motion with ap- the 3DGRAPE elliptic solver, respectively. A plications towards aircraft control and ma- grid refinement study was performed and the neuverability is planned.

total number of grid points ranges from 250,000 to over 1 . 7 million. Eugene L. Tu Applied Computational Fluids Branch Accomplishments. Computations were Ames Research Center made at a transonic Mach number of 0.90, (415) 604-4486 angles-of-attack from 0" to 1 2 " , and a Reynolds number based on mean aerodynamic wing chord of 1.52 million. Very good agreement of computed wing surface pressures and inte- gratedforces(e.g., lift, pitchingmoments, and drag) with experimental measurements has been obtained and indicates that the com- puted results are accurate. The accompany- ing figure illustrates the effect of the canard at a 1 2 ' angle-of-attack. Specifically, the eff'ect of the canard on the characteristics of wing vortex breakdown was noted. A side-by- side comparison of the canard-on and canard- off cases showed that the wing exhibits a stable vortex in the presence of the canard.

This result indicates the potential of the ca- nard to delay wing vortex breakdown and has been previously observed in numerous ex- perimental studies.

The surface pressure map also shows the significantinfluenceof the canard on the body aerodynamics.

Figure 7-12. Fundamental Research i n Vortex Interactions

Chapter 8 Hypersonic Aerodynamics The Generic Hypersonics Aerodynamics program exists to institute a fundamental, permanent base of technological superiority in hypersonic aedaerothermodynamics and to provide NASA with the long-term capability to achieve the goals of f'uture NASADoDDOE hypersonic vehicle research and development programs. It focuses on increasing the understanding of hypersonic aerdaerothermodynamics phenomena associated with slender, air-breathing hypersonic ve- hicles that use highly integrated airframe/propulsion systems and increasing the capability to experimentally and computationally simulate and analyze these phenomena. The program emphasizes both enabling and enhancing technologies related to hypersonic flight such as numerical simulation of turbulence and boundary layer transition, real-gas effects, rarefaction effects, shockmoundarylayer interactions, and validation of computationalmethods. It explores flight research opportunities for technology areas that cannot be adequately addressed by computationalmethods or ground facilities. Amajor objective of the program is to structurecost- effective partnerships with industry which will enhance the rapid and efficient transfer of research results into the design and development of aerospacevehicles. University participation involves sponsored research and a refocusing and rededication of the aeronautics/astronautics curricula to hypersonics.

Program Manager: Jim Moss OAST/RF Washington, DC 20546 (202) 453-2820 -1. Hypersonic Shock-Wave Boundary Layer Interaction I

8-1 Experimental Study of Hypersonic Shock-Wave/Turbulent-

Boundary Layer Interaction F~OWS

Objective. To provide accurate new hyper- Significance. These new data represent the sonic data for turbulence modeling and code first 3-D hypersonic shock-wavehoundary validation. layer interaction data suitable for turbulence modeling and code validation.

Approach. Design, construct, and test a series of simple geometries that create three- Status/plans. More complex flow fields w i l l dimensional(3-D)hypersonic shock-wavdtur- be investigated with two fins and/or compres- bulent-boundary layer interaction flows in sion ramps, which will begin to look like the Ames 3.5-Foot Hypersonic Wind Tunnel. hypersonic inlets.

Accomplishments. A new testbed was con- C. C. Horstman, Marvin Kussoy structed for the 3.5-foot wind tunnel which Fluid Dynamics Division enables one to test a series of 2-D and 3-D flow Ames Research Center fields of varying complexity. The first test (415) 604-5950 series was successfully completed. The test geometry consistedof a single sharp fin whose shock-waveinteracts with a turbulent bound- ary layer. Test conditionswere Mach number

8.2, Reynolds number Re - 9 x 106 and fin

angles from 5" to 15". The measurements included surface flow direction, pressure and heat transfer, as well as detailed Pitot pres- sure and yaw angle surveys of the flow fields.

Figure 8-2. Numerical Pefiormance Estimates for a Generic Hypersonic Forebody ,

8-2 Numerical Performance Estimates for a Generic Hypersonic

Forebody

Objective. To examine the effects of nose performance measures were generally more bluntness and real-gas models on the com- sensitive to real-gas effects than forebody puted flow physics andintegratedperformance performance measures such as lift, drag, and parameters of a generic hypersonic forebody pitching moment.

operating at NASP-like flight conditions.

Significance. The bluntness study provided Approach. The flow over the McDonnell confidence in the ability of the present CFD Douglas Generic Option #2 36-inch blended- codes to predict surface pressures accurately wing-body was computed for Mach 16 flow at at high Mach numbers for relatively complex an altitude of 125,000 ft. The parabolized configurations. Confident heat transfer pre- Navier-Stokessolver, UPS, was employedwith diction remains elusive because of grid sensi- each of three availablegas models: (1) perfect tivity and boundary layer transition uncer- gas, g = 1.4, (2) equilibrium air, and (3) finite- tainties. The high energy results give an rate air. Initial conditions for the space- indication of the level of importance of real- marching calculations were provided by the gas effects at NASP-like cruise conditions on time-dependent Navier-Stokes solver,TUFF, both forebody flow-fieldphysics andintegrated which possesses the same real-gas capabili- performance parameters.

ties as UPS. Bluntness effectswere studiedat wind tunnel conditions (Mach 1 1 . 4 , R ~ L = 2.904 StatusPlans: The Option 2 geometrywill be 1 0 7 ) . employed again to address issues involved in chemistry modeling such as loose vs. tight Accomplishments. Solutionswere obtained couplingand conservative vs. nonconservative for sharp- and blunt-nosed versions of the form of the species equations.

Generic Option vehicle under wind tunnel conditions, assuming a perfect gas. At zero Scott Lawrence incidence, the effect of bluntness was ob- Applied Computational Fluids Branch served to be most significant along the lower Ames Research Center centerline because of the accumulation of en- (415) 604-4050 tropy-layer fluid into a viscous core along this line. The magnitude of the bluntness effecton centerline pressure and heat transfer distri- butions predicted by the computations gener- ally agrees well with experimental data.

Real-gas effects were studied for the Mach 16 flight conditions by computing forebody flow fields using each of the three available gas models in the UPS and TUFF codes. A limited investigation of chemistry coupling and grid effects was also performed. These param- eters were studied to evaluate their effect on the lower centerline flow physics and inte- grated performance measures such as lift, drag, and inlet mass flux. Integrated inlet Figure 8-3. Forebody Pressure Contours

8-3 Calculation of Forebody Flow Field for a Candidate

Aero-Space Plane Configuration

Objective. To demonstrate the application calculation shows pockets of inflow and out- of the Langley Aerothermodynamic Upwind flow on the lower surface and a buildup of the Relaxation Algorithm (LAURA)code to com- viscous boundary layer along the lower pute the flow over the forebody of a candidate centerline. Also, anomalies in profiles of the Aero-SpacePlane configuration,and toinves- total enthalpy at the engine inlet face, which tigate flow-field anomalies predicted with were predicted in the other computational another state-of-the-artflow-fieldcode and to results, were not confirmed by the present compare the results. calculation.

Approach. The LAURAcodesolvesthe time- Status/Plans. Efforts are ongoing to resolve dependent, thin-layer Navier-Stokes equa- the anomalies in total enthalpy profiles.

tions using a point-implicit scheme to con- Knowledge gained in the present work will be verge the numerical solution. A two-block applied in a forthcoming experimentdcom- grid was used to calculate the blunted nose putational study in support of the National region and the downstream portion of the Aero-Space Plane program.

forebody in sequential steps. For comparison purposes, the flow-field grids used in this Richard A. Thompson work were identical to those used by the other Aerothermodynamics Branch computational methods. The calculation was Langley Research Center performed for flight conditions at Mach 15 (804) 864-4367 and laminar, perfect gas flow was assumed.

Accomplishments. A converged solution over the Aero-Space Plane configuration was obtained which yielded complete flow-field quantities and surface properties (i.e., heat- ing and pressure) for study. Computations were performed on the NAS Cray-2 computer and required approximately 22 hours of CPU time. Surfaceproperties and the nature of the flow field at the engine inlet (forebody end) have been examined and comparisons with the other computational results have been made.

Significance. Results show that the LAURA code is capable of flow-field calculations for this type of configuration and is able to pro- vide neededinformation about flow-field qual- ity and surface heating. For example, the Figure 8-4. ComputationaVExperimental Parametric Study of 3-D Scramjet , Inlets at Mach 10

8-4 Computational5xperimental Parametric Study of 3-D

Scraxnjet Inlets at Mach 10

Objective. To calibrate current computa- Significance. The results demonstrate the ability of this family of CFDcomputer codes to tional fluid dynamics (CFD) computer codes for hypersonic forebodyhnlet flows using ex- predict forebody pressures and heat transfer perimental data from the Generic Option #2 rates with reasonably good accuracy. The results in the inlet are encouraging, but also program.

indicate areas of needed improvement9 the Approach. Numerical simulation of three- numerical modeling of internal flows. In particular, the turbulence modeling needs to dimensional hypersonic flow past a forebodyl be improved.

inlet configuration is performed using a state- of-the-art family of upwind methodology CFD computer codes, CFL3DE and CFL3D. Nu- StatudPlans. Comparisons of the computed merical results, including surface pressure results and experimental data will be sent to and heat transfer along the forebody and the Generic Option #2 report authors for their inside the inlet, are compared with experi- comments and to learn of any update in the mental data for the same configuration and experimental data. After their response, the results will be documented in a NASP publi- freestream conditions.

cation. Also, new turbulence models, better suited to internal flows, will be incorporated Accomplishments. The forebodyhnlet cal- culation has been completed. The computed into this family of CFD computer codes.

forebody pressures and heat transfer exhibit Arthur D. Dilley, George F. Switzer, reasonably good agreement with the experi- William M. Eppard mental data as does the computed results in the inlet. In some regions of the inlet, the Langley Research Center agreement is quite good considering the (804) 864-2288 complex nature of the flow. In other regions, the agreement is poor indicating deficiencies in the numerical modeling.

. Technique for Hypersonic Powered Tests of Airbreathing Configurations

,

8-5 Technique for Hypersonic Powered Tests of Airbreathing

Configurations

Objective. To experimentally evaluate an Significance. This test program produced unproven technique of using metric model the first hypersonic powered force/moment parts to conduct simulated exhaust powered data on a National Aero-Space Plane-type tests of an airbreathing vehicle in a hyper- configuration. The results of this effort have sonic wind tunnel. shown that (1) the effect ofthe exhaust plume on the measured forebody forces/moments is Approach. The method required separate negligible; (2) the internal inlet contribution tests of two identically scaled models. The to the forebody drag is less than 10%; and (3) external surface force/moment contributions the interfkrenceeffect on the measured af'tbody from the nose to the engine cowl trailing edge foradmoments as a result of the forebody were determined from tests of the sting strut mount is negligible.

mounted,unpowered forebody model that was fully metric. The aftbody external surface Status/Pla.ns. Additional tests of the pow- force/moment contributions from the cowl ered "D model in the 20-Inch Mach 6 Wind trailing edge rearward were determinedfrom Tunnel are planned to investigate the use of tests of an entire model that had a metric air, as opposed to the tetrafluorornethime/ argon mixture, as a simulated exhaust gas.

aftbody. This entire model was supported by is part of GovernmentWork Pack- a strut mounted to the forebody and had a This effort closedinletandametric aftbody. The scramjet age 5 in support of the National Aero-Space exhaust was simulated by using a 70/30 (per- Plane Program cent mole fkaction) mixture of tetrafluoro- methane and argon. W. Witte David Applied Aerodynamics Division Accomplishments. Both models were tested Langley Research Center in the Langley 20-Inch Mach 6 Wind Tunnel (804) 864-5589 at a Reynolds number of 1 . 0 x 1 0 % . The appropriate forebody and aftbody force ac- counting procedures for this metric model parts technique were developed and demon- strated. In addition, facility operating issues related to conductingexhaustsimulationpow- ered testing with a metric aftbody model were examined. Also, the powered metric aftbody model was configured with various wing inci- dence angles (-1.5", -3.0", -4.5") and elevon deflection angles (O", +lo", and 220" ) to examine the trim characteristics with simu- lated exhaust.

8-6 Effect of Inlet Representation on Powered Hypersonic

Aftbody Flows

Objective. To assess the aftbody perfor- and the cowl. For the faired-overinlet, there mance of a powered hypersonic airbreather is only an expansion. Also, the size of the (PHA) model by computationally examining boundary layer on the flow-through inlet case the effects of a flow-throughinlet and a faired- is significantly smaller because it develops over inlet at a variety of freestream condi- only from the cowl leading edge instead of on tions.

the entire forebody. In the aftbody region, the flows are quite similar, including the location Approach. One method for obtaining pow- of the cowl trailing-edge shock, the shear ered effects on PHA concepts is to fair over the layer, and the weak compression within the inlet and then route alternate gases through shear layer. The line plot compares aftbody the combustor section to simulate the com- surface pressures and integrated force and bustion products. However, the fairing alters moment components for the two PHAconfigu- the external flow field and possibly affects the rations at this freestream condition. The aftbody local flow field and surface pressures.

result is that the difference in integrated The effect of the fairing was studied using the aftbody pressures due to different represen- General Aerodynamic Simulation Program tations is less than 1%. This comparison (GASP) to analyze two PHA configurations.

remained consistent for the entire set of The flow-through inlet model provided the freestream conditions that were simulated.

necessary inlet geometry to simulate the flow associated w i t h the cowl leading edge. The Significance. Within the scope of this in- faired-over inlet model eliminated all cowl vestigation, two-dimensional CFD results in- leading-edge effects by covering the inlet, dicate that there is no significant effect on resulting in a simple expansion over the cowl.

powered aftbody performance due to fairing over the inlet compared to a flow-through Accomplishments. Two-dimensional CFD inlet. Flow-field differences exist for the two simulations were performed on the two con- inlet representations, but they are isolated to figurations at various freestream conditions, the flow region near the cowl and appear to includin Mach numbers 3.0 to 14.0 at Re, = have minimal impact on aftbaody surface %; 2.0 x 10 / f t . and Reynolds numbers 0.5 to 7.0 pressures.

x 106/ft. at M, = 6.0. These conditionsyielded static nozzle pressure ratio (SNPR) values Status/Plans. There is a need to extend the from 12 to 256 because of the different CFD solutions to three-dimensional to ac- freestream pressures. The internal condi- count for 3-D relief effects as well as proper tions were identical for all solutions, namely expansion of the plume onto the entire span of M = 1.0 at the nozzle throat with a jet total the aftbody and any wing or control surfaces pressure of 30 psi.

that may be subject to influence by the ex- haust flow.

Comparison of Mach number contours for the two inlet representations is shown in the Lawrence D. Huebner figure fort e conditions at PLL, = 14.0 and%, Applied Aerodynamics Division

B

= 2.0 x 10 /ft. Sfj$d%x.nt differences can be Langley Research Center seen in the flow fields near the cowl for the two (804) 864-5583 configurations. The cowl leading edge on the flow-through inlet creates a shock followedby a qoderate expansion of the flow and then a recompression between this expansion region

Pitot pressure

26.

0% Argon, 50Yo Freon - 12 jet: SNPR

.\ 2-D GASP PNS Turbulent Solution vs Experiment

,

8-7 Generic Scramjet Nozzle/Aftbody-Studies

Objective. To determine optimum numeri- of surface pressures and flow-field details cal algorithms and codes for analyzing &- around the cowl trailing edge. A 3-D explicit, sive multiple-species, hypersonic 3D flow elliptic full Navier-Stokes code was modified fields, including comparisons with appropri- to accept block-structured grids and two mul- ate experimental data. The goal is to develop tiple-species, half-span nozzle calculations procedures for the analysis and design of were performed.

generic hypersonic aftbody geometries.

Significance. Proposed hypersonic vehicles Approach. Several Euler/Navier-Stokes will operate in flight regimes beyond the ac- codes, employing a variety of numerical algo- cessibility of ground-based facilities. The rithms, were exercised to analyze flow about ability to numerically analyze and design a powered generic scramjet nozzle/aftbody scramjethftbodycomponentsfor suchvehicles model. The calculated results were compared is crucial for efficient, interdisciplinary pro- with measured surface (static) and flow-field pulsiodairframe integration.

(pitot) pressures for several single- and mul- tiple-species flows.

Statufllans. A NASA Technical Paper has been written documentingthe flow-fieldpres- Accomplishments. Flow-field pitot pres- suremeasurementexperiment (to be released) sures were measured about a powered Ge- and an AIAA paper documenting the CFD neric ScramjetNozzle/Aftbody (GSNA)model results and data comparisons (surface and in the Langley 20-Inch Mach 6 Wind Tunnel.

flow field) has been published. A full 3-D The GSNA has been modeled using computa- hypersonic vehicle will be numerically mod- tional fluid dynamics (CFD) codes ranging eled with poweredscramjet exhaust flowsimu- from a 2-DEuler (inviscid)code to 3-DNavier- lation and comparisons made between the Stokes(viscous)codes. An explicitParabolized CFD results and ongoing wind tunnel mea- Navier-Stokes (PNS) code with multiple gas- surements. In addition, the GASP code will eous species capability was modified to allow be coupled with an iterative desigdoptimiza- solid wall boundary conditions internal to the tion code to permit automated design of grid to define a cowl separating a jet flow from scramjet aftbody geometries for optimum lift a freestream flow. Calculations in Euler mode and thrust constrained by pitching moment were made to determine the accuracy obtain- requirements.

able with inviscid calculations. Results from both 2-D and 3-Dcalculations were compared Kenneth E. Tatum, William J. Monta with the experimental data. The implicit, Applied Aerodynamics Division multiple-zone General AerodynamicSimula- Langley Research Center tion Program (GASP)code was similarly exer- (804) 864-5587 cised and shown to be more robust than the explicit code for high nozzle pressure ratios.

Thejet fluid was modeled as air (perfect gas), a mixture of nitrogen and oxygen, and two mixtures of argon and Freon to simulate the ratio of specific heats of actual hot scramjet exhaust gases. GASP 2-D PNS calculations, including turbulence modeling (as shown in the figure), showedimproved agreement with exNrimentoverEuler solutions,both in terms

e 8-8. PEMACH - Computed Pressure Distributions and Experimental

I Data Comparisons

8-8 Advanced Aero-PropulsionPerformance Design Tool

(PEMACH)

Objective. To improve the current capabil- Significance. This capability brings the ity to rapidly predict local pressure and ther- ability to quickly predict configuration loads mal loads on hypersonic conceptual designs suitable for structural, thermal, and stability for structural analysis and stability and con- and control analysis.

trol analysis.

StatualPlans. A preliminary version of Approach. An engineeringtechnique based PEMACH has been released and is currently on the method of characteristics (MOC) was being used for aerodynamic (lift, drag, mo- chosen as a starting point for this study, ments) analysis and design of hypersonic ve- which began last year. Although this particu- hicles. Workis continuing to efficientlymodel lar technique was originally developed for the inlet and nozzle flows, t o fully couple the applications to supersonic wings and conical boundary layer to the 3-D effects, and include shapes, it has shown very good agreement boundary layer transition prediction.

with experimental and flight data up to a Mach number of 4. The method is being Suresh H. Goradia, adapted to complete hypersonic configura- VIGYAN Associates tions and results compared with experimen- tal data, Navier-Stokes(NS)calculations, and Abel 0. Torres, Charles R. McClinton engineering methods currently in use.

Hypersonic Technology Office Langley Research Center Accomplishments. The code, PEMACH, (a (804) 864-6253 - Practical Engineering Method for Aero-pro- pulsion characteristics at &personic Mach Numbers) can now solve the flow over the fuselage and wings, including capture of en- gine mass flow, in a quasi-3-D fashion at hypersonic Mach numbers. The code includes solutioncapabilities for either sharp or blunt leading and trailing edges, and for completely arbitrary planform.

P a r t of the method development was calibra- tion to experimental results, as illustrated in the figure. The illustrated comparison used the experimental results from the National Aero-Space Plane (NASP)Generic Option #2 3-D forebody-inlet model. Excellent agree- ment with wall pressures, both on and off centerline,and shock locationsare illustrated in the figure. Comparison with CFD solutions verify accurate prediction of mass capture, apd comparisons with unpowered aerody- namic test results from the NASP Test Tech- nique Demonstrator('ITD)model verify accu- rate prediction of Cma over a large variation in angle-of-attack.

BLACK AND VlrWiTE PHOTOGRAYh Figure 8-9. Wing Glove for Pegasus Crossflow Transition Experiment ,

8-9 Wing Glove for Pegasus Crossflow Transition Experiment

Objective. T o design and fabricate a wing Significance. Based on this work, a FX-0 glove that provides a smooth non-ablating glove will be fabricated and flown on the test s d c e for the measurement of hyper- Pegasus flight, scheduledforJune 1992. This sonic 3-0 crossflow boundary layer transi- FX-0 glove is the first of three gloves planned tion.

for this hypersonic flight research program.

FX-0 is to validate the glove The goal of Approach. Ablation products from the cur- concept for measuring crossflow transition on rent thermal protection system on the Pega- a swept delta wing at hypersonic speed. Fu- sus delta wing will interfere with sensitive ture gloves will be optimized for making more aerodynamic measurements like boundary precise measurements of this phenomena.

layer transition. To minimize this interfer- ence, an aerodynamically shaped glove with a Status/Plans. Further work is required to smooth nonablating surface was installed on finalize the fairing design and glove assem- the existingwing. A ceramic glove fairingwas bly. Fabrication of the FX-0 glove and its used to provide a smooth transition between fairing is scheduled to begin January 1992.

the wing and glove surface. A l l of the flight instrumentation was contained within the Paul Kolodziej glove so that the integration with Pegasus Thermal Protection Materials Branch was simply a cable connection.

Ames Research Center (415) 604-5377 Accomplishments, A prototype wing glove was fabricated to identify and demonstrate concepts for meeting the requirements of the crossflow transition experiment. The glove consists of a copper skin attached to the ce- ramic blanket known as TABI (Tailorable Advanced Blanket Insulation). Since the aerothermodynamic heating at hypersonic speeds is substantial, a copper skin was cho- sen to quickly conduct heat away from the leading edge and therefore reduce the ther- mal gradient and stresses. TABI both sup- porta the copper skin and insulates it from the composite wing. Recent work with the rigid ceramic insulation known as TUFI (Tough Uni-Piece Fibrous Insulation) shows good progress toward fabrication of the ceramic glove fairing.

, ,

Chapter 9

Aeroacoustics Research and Technology

Aeroacoustics is concerned with the fundamental understanding, prediction, and control of noise and acoustic loads produced by the motion of f l u i d s and bodies moving through the atmosphere.

To achieve this objective, research is being conducted in three main areas: 1) Computational Aeroacoustics - establish CFD methodology as a discipline for acoustics technology development foeusing on computational models of high-speed flow noise and shockhortex interactions. 2) SupersonicJ e t s h a d s - understandlpredictduce noise and acoustic loads generated by high- aircraft and, in particular, establish innovativeconcepts for supersonicjet noise suppres- speed sion and develop a detailed acoustics loads database on &ame structures. 3) Long Range

Propagation - model atmospheric propagation effects supported by fundamental verification/

validation experiments.

Program. Manager: Benjamin Neumann OAST/RF Washington, DC 20546 (202) 453-2795

100% thrust, standard lift configuration

Sideline 11 6 . 3

Centerline 11 6 . 3

r

80% thrust, 30% lift increase

Sideline 11 2 . 3

Centerline 11 2 . 3

110 I

Figure 9-1. Contours of Effective Perceived Noise Levels (EPNdB) for Standard I Takeoff and Reduced Thrust/Increased Lift Takeoff for a Jet Aircraft

9-1 Jet Noise Predictions for Reduced Thrust Thkeoff Study

Objective. To demonstrate the use of the Signi?gcance. The use of the ANOPP predic- Aircraft Noise Prediction Program (ANOPP) tion code shows that reduced thrust takeoff to evaluate the effectsof reducbdjet thrust on procedures can be beneficial. Sideline noise community noise levels. Redukd jet thrust levels are of greatest concern during takeoff.

levels at takeoff increase the required dis- Since the FAA sideline noise measurement tance on the runway (ground roll) before lift- point location depends on the point of lift-off, OE To compensate for the increased ground only a slight benefit can be derived fiom in- roll, several levels of increased lift were in- creasing the amount of lift while keeping cluded in the study. thrust at a maximum. The key to reducingjet noise lies in thrust reduction.

Approach. Predictions were made using the ANOPP dual stream coannularjet noise mod- Status/Plans. It has been shown that re- ule for the AST-205-1 aircraft powered by duced thrust, used in conjunction with in- GE21WllB14A scaled engines. A power set- creased lift to compensate for increased run- ting of 100%thrust and a standard lift con- way distance, can reduce Effective Perceived figuration were used as a baseline for this Noise levels at the sideline and centerline study. Power settings varied from 100% to FAA noise measurement locations. Betterjet 80% maximum thrust, while lift was increased noise prediction modules being designed for in increments of 15% to a maximum of 60%. ANOPP will allow the noise impact from a A n angular rotation rate of 3 O per second and high-speed civil transport (HSCT)aircraft on a constant climb angle of 8" were used. Effec- community noise to be evaluated.

tive Perceived Noise Level (EPNL) values were calculated at the FAA prescribed noise Robert A. Golub measuring points. Noise levels were also Acoustics Division calculatedon a 1-by-6mile grid of 65 observer Langley Research Center positions, with each observer located 4 feet (804) 864-5281 above the ground. Noise contours were gener- ated from these observer positions.

Accomplishments. Contours for both the baseline case and the 80% thrust/30% in- creased lift case are shown in the figure. The reduced thrust case decreases the necessary ground roll distance by 646 feet and lowers the sideline EPNL level by 4 EPNdB from the maximum thrust baseline case. Although the centerline noise level is not of major concern, it too benefits by 4 EPNdB from the reduction in thrust and increase in l i f t . Additionally, as the figure shows, reducing thrust and in- creasing lift alter the shapes of the contours.

;As thrust is decreased and lift is increased, the contours tend to shorten in length and become more compact.

, Scale Model of McAir 279-36 Installed for Hot-Gas Ingestion Testing (top); emperature on Measured &e 1 Sound Pressure Levels Measured at Three

9-2 ASTOVL Acoustic Loads Test

Objective. To characterize the acousticsand and impingement tones were observed, and dynamic structural loading of an advanced their frequencies were well predicted by cur- short takeofVvertical landing (ASTOVL) air- rent jet noise theories.

craft operating in ground effect and at el- evated jet exit temperatures. This research ce. Acoustic loads on ASTOVL, was part of a Lewis/Langley cooperative pro- configurations can potentially attain suffi- gram, complementary to in-house research ciently high levels to cause structural failure.

aimed at improvingthe understanding of the No satisfactory theory exists that can predict impingementacousticsofmultiple heatedjets.

the acoustics of multiple heated non- axisymmetric jets operating in ground effect Approach. A 9.2% scale model of the McAir in an off-design condition. The acquired data 279-3C aircrafk was tested in the Lewis 9- by will be valuable in understanding both the 15-Foot Low-Speed Wind Tunnel. The model fundamental physics ofjet-impingement and (see accompanyingfigure)was built to evalu- the configuration issues in ASTOVL aircraft ate the hot gas ingestion (HGI) characteris- acoustics.

tics of the design at temperatures up to 1000" F, at nozzle pressure ratios (NPR) of 4 . The Status/Plans. This test served as Phase I1 of model geometry parameters varied included Lewis' hot gas ingestion work with this de- forward nozzle splay angle (from 12" inward sign. Additional acoustic testing is planned to 18" outward) and L i f t Improvement Device for Phase 111, includingthe placement of trans- (LID) configuration (on, off). Acoustic data ducers on the underside of the fuselage be- were obtained from a linear array of five tween the nozzles. This will allow for the freestream microphones located parallel to investigation of the dynamics of fountain im- the model centerline. Model dynamic loads pingement on the model, and the propagation were measured with eight watercooled dy- of noise from the fountain to the farfield.

namic pressure transducers located on the underside of the canard, wing, and aft fuse- L. Kerry Mitchell lage.

Acoustics Division Langley Research Center Accomplishments. Data were taken for (804) 864-3625 over 700 combinationsof model geometryand test conditions. The data shownillustratethe minimal effect ofjet temperature on the noise level (OASPL) as measured on the canard, the wind, and in the freestream, while the jets were operating at the design condition. As shown by the data, levels up to 160 dB were observed on the canard (which was closer to the nozzles than the wind) and up to 140 dB in the freestream. At other conditions, levels as high as 180 dB with discrete tones 160 dB in amplitude were measured. Tones were gen- erated at a wide variety of conditions (model height, nozzle pressure ratios, jet tempera- tureg) and geometries. Both free-jet tones 17 1

0 Vane 1 - stowed, -IOo

Vane 2 - deployed, +20°

A Vane 3 - deployed, +20°

1 m 5

Aeroacoustic I -

I

loads, psi 1.0

ep

F-15, B-1B nozzle flap loads

-

m 5

0O0

O m 1

, Figure 9-3. Aeroacoustic Loads on Thrust Vectoring Vanes of a Model F-18 High Alpha Research Vehicle (HARV) 9-3 Aeroacoustic Loads on F-18 HighAlpha Research

Vehicle (HARV)

Objective. To provide a benchmark data- corded on previous F-15 and B-1Bouter diver- base on the nearfield acousticloads generated gent flap studies. The F-15 and B-1B loads weredominatedbyatwinjet resonancemecha- by supersonicjet plumes of high-performance nism, which is absent in the present research aircraft. High aeroacoustic load levels due to twin jet resonance have caused component when the vanes are fully deployed. Ground failures on F-15 and B-1B aircraft. A model tests of a full scale F-18 HARV at Dryden also and full-scale static study was conducted to confirmed high levels of near-field loads. The evaluate aeroacoustic loads associated with removal of the divergent nozzle flaps on the the thrust vectoring control system (TVCS)of baseline F-18 to accommodate the TVCS re- the F-18 High Alpha Research Vehicle sulted in a convergent nozzle configuration.

Convergent nozzles produce significantly (HARV).

greater shock noise. At the maximum after- Approach. A 7.4% aft-end model of the F-18 burner power setting, a condition not tested HARV/TVCS was designed, constructed,and in the model study,anintense tone was present in the near-field spectrum. This tone lies near instrumentedwith water-cooleddynamicpres- the predicted frequency for twin supersonic sure sensors to measure aeroacousticloads on the three thrust vectoring vanes, Dynamic plume resonance-a phenomenon known to loads were measured on the model for vane contribute to the high acoustic loads for the convergent nozzle configuration.

angles from stowed (-10") to fully deployed (+20"). Only military power nozzles were studied with nozzle pressure ratio's ranging Significance. The database obtained in the from 2 to 5 and jet total temperatures from present research effort provides structural 104"to680"F. Inaparallel effort, researchers design guidelines for future application of a TVCS. The removal of the engine nozzle flaps at NASA AmesDryden measured the near- and installation of the TVCS confirms prior fieldacousticcharacteristicsofthe F-18HARV with TVCS installed on the twin F-404 en- predictions that significantincreasesin plume gines and on an unmodified F-18. The diver- shock noise will interact with the aircraft gent nozzle flaps on the F-404 engines were structure, which could possibly reduce the expected life cycle for components on the F-18 removed on the F-18 HARV, leaving essen- tially convergent nozzles. Near-field sensors aircraft.

were positioned near the aircraft in a similar fashion to one used in the model scale study. StatusPlans. The results of research will be reported with NASA AmesDryden.

Accomplishments. Overall dynamic pres- sure loadsvariedconsiderablywithvane angle John M. Seiner position and nozzle pressure ratio, as shown Acoustics Division Langley Research Center in the figure. The peak pressure value of 1.5 psi at a nozzle pressure ratio of 2.8, recorded (804) 864-6276 for thrusting vanes at 20", translates to an estimated full- scale dynamic force load on a vane of between 300 and 400 lbs. Stowed vanes (-1O")experiencemuch smaller dynamic loads, 0.4 psi rms. At higher pressure ratios, the initial spread rate of the jet is less, and leys jethrane interaction occurs. These levels are significantly higher than previously re- 2.5 h -w- X U

Near Field Vorticity

ar ressure Fluctuati~~s Figure 9-4, Direct Computation of Aerodynamic Sound Generation. Figure 1: Scattering of Plane Sound Waves by a Compressible Vortex; Figure 2: Sound Generated by Co-Rotating Vortices

9-4 Direct Computation of Aerodynamic Sound Generation

Objective. The prediction of the far-field Significance. Successful simulation of the sound produced by a turbulent flow requires simple test problems mentionedabove, includ- a detailedknowledgeof acousticsource terms. ing excellent agreement with theories and In many theories these terms are crudely experiments, shows that accurate direct approximated, based on empirical correla- computation of fluid flow with sound sources tions. Computationofboththeacoustic sources and *-field sound, whose energy is 10 orders and far-field soundusing the unsteadyNavier- of magnitude smaller than the flow energy, is Stokes equations allows the physics of the feasible.

sound generation process to be studied di- rectly. Such a knowledge will allow direct Status/Plans. The codes developed will be validation of aeroacoustic theories and help used to compute the sound generation in two develop noise control strategies. and three dimensional turbulent shear flows.

Approach. The Navier-Stokesequations are Sanjiva K. Lele solved numerically. The infinite domain is Department of Aeronautics and Astronautics truncated and newly developed obliquely Stanford, University nonreflecting boundary conditions are em- (415) 723-7721 ployed at the computational boundaries.

Parviz Moin Accomplishments. To address the feasibil- Ames Research Center ity and accuracy of direct computation of (415) 604-5127 aerodynamic sound generation, severalmodel problems have been considered, including: Tim Colonius, Brian E. Mitchell the scattering ofsound wavesby a two- dimen- Department of Mechanical Engineering sional (2-D) compressible vortex, and sound Stanford, University generation by a 2-D co-rotating vortex pair. (415) 723-9602 Figure 1 shows (1) the amplitude of the acous- tic field generated by the interaction of inci- dent planar waves with the 2-D compressible viscous vortex, and (2) a comparison of scat- tering amplitude directly measured in the computations, with a theoretical prediction based on an acoustic analogy. Their agree- ment verifes the accurate representation of the acoustic sources by the numerics: discretization does not act as a significant source of sound in the computations. Figure 2 shows a preliminarycomputation of the far- field pressure fluctuations of the sound field, which is generated by a co-rotating vortex pair, and vorticity contours in the vicinity of -the vortex pair.

,

- Boundary element method

0 Experimental data Insertion loss (dB) 10

-5

-1 0

-3 -2 -1 0 1

Dimensionless height, Y Figure 9-5. Comparison of Boundary Element Method with Experimental Results I

9-5 Computational Model for Long-Range Acoustic Propagation

. T h i s research has shown that

Objective. To develop a method capable of the Boundary Element Method can be used to predicting the collective innuence of the im- study long-range propagation. Several inno- portant physical features oflong-rangepropa- gation. vations have made it possible to apply the method to frequencies and distances never Approach. Two-dimensional sound propa- before possible with numerical methods. Re- gation over an arbitrarily shaped ridge situ- sults also show that the method compares well with measured results.

ated on a locally reacting infinite boundary in a homogeneous medium is treated using the StatudPlans. The method will be used to Boundary Element Method. The discretiza- help analyze data from field experiments.

tion procedure provides a set of linear alge- braic equations which are solved for the un- knowns overthe boundary surfaceofthe obsta- Willie R. Watson Applied Acoustics Branch cle. The solution of the boundary unknowns Langley Research Center is used to compute the sound field at any point (804) 864-5290 of interest in the domain.

Accomplishments. Acomputationalmethod capable of studying long-range propagation was developed. An effective scheme to reduce computational time and storage while retain- ing reasonable solution accuracy using cubic isoparametricelements, approximate polyno- mials, and asymptotic expansions of the Hankel function was implemented. A n out- of-core equation solver for the large, fully populated unsymmetric algebraic equations produced by the Boundary Element Method was developed, and computed results com- pared with measured data.

h truts and Sensors for nd Tunnel Acoustic Studies ,

9-6 Quiet Struts and Sensors for Wind el Acoustic Studies

Objective. To develop the technology and CFDsimulationshave identifiedseveral strut/ design methodology for quiet struts and sen- brace junctions that have well-behaved flow sors to be used in subsonic airflow: primarily fields and are potentially quiet components closed-return wind tunnels.

for microphone struts. A wide range of impor- tant geometric parameters have been identi- Approach. Computational fluid dynamics fied for aeroacoustic testing in the 7- by 10- (CFD) will be used to analyze flow fields foot tunnel.

around struts and junctions so as to identify low-dragconfigurations,which should thereby S i g n i f i c a n c e . Improved acoustic sensors have low levels of self noise. Selected shapes and support struts will result in significantly and mountingconfigurations will be tested in lower background noise in closed-returnwind the Ames 7- by IO-foot wind tunnel. Acoustic tunnels, without any modification to the facil- sensors will be tested with nose cones de- ity as long as wind noise is the dominantnoise signed to generate low levels of wind-induced source, as it is in the 40- by 80-foot tunnel.

noise. Investigations will be made of signal Thisis equivalent to addingexpensive screens processing methods for combining signals from to the circuit for turbulence control. The microphone pairs such that uncorrelated quality of acoustic data acquired from low- (wind)noise will be attenuated.

noise Advanced Ducted Propellers, HSR jet suppressors, and similar devices will be sig- Accomplishments. A 40 x 80 wind tunnel nificantly improved.

test of special acoustic sensors developed by Northrop Corporation has shown that nose Statufllans. Quiet struts have been de- cones shaped like Pitot probes round noses, signed for testing in the 7- by 10- foot wind and with the sensor openings several probe tunnel. Fabrication of the struts and associ- diameters downstream from the nose shoul- ated hardware will commence in the near der, generate 6 dl3 to I O dB less wind noise future. Discussions concerning future coop- than do the commercial microphones with erative research on sensor development with nose cones used at Ames. Further reductions Northrop Corporation are being planned.

in turbulence-inducednoise were achieved by summing the signals from two sensors in the Paul T. Soderman same probe or by summing signals from two Fixed-Wing Aerodynamics Branch sensors in adjacent probes. Uncorrelated Ames Research Center (415) 604-6675 sound was thereby cancelled.

Chapter 10

High-speedResearch

The High-speed Research Program seeks to provide the technology answers for an environmen- tally safe design of a supersonic transport. In aerodynamics the technologies focus on efficiencies needed to provide a low noise configuration,both in climb out and cruise, The specific objectives of the program are to provide enabling technologies for supersonic transports to achieve in a practical configuration:acceptableoverland sonicbooms, efficient second stage climb-outfor low noise, and supersonic laminar flow control for peak cruise efficiency. The laminar flow control will significantly reduce the aircraft gross weight, hence the noise signatures. The sonic boom element w i l l be evaluated in FY 1992 to determine whether the promise of low boom is great enough to continue research.

There is a great body of data from the previous supersonic cruise research program that ended in the early 1980s. However, prediction methods, optimization techniques, and experimental testing have improved for complex configurations. In addition,new ideas for high lift devices and subsoniclaminar flow control are now candidates for investigation-for supersonicconfigurations.

These developments have given hope that supersonic aircraft efficienciescan be improved to help the environmental and economic realities of the coming age. Based upon systems studies these efficiencies translate into goals for lift-to-dragratios of 10 in climb, 15 in transonic cruise, and 10 in supersonic cruise.

The High-speedResearch Program is focused on( 1) CFD modeling andvalidation of aeroacoustic analysis including measurement techniques and atmospheric modeling; (2) development of innovative high lift devices to provide increased lift-to-dragratios on climb out; (3) design, wind tunnel test and analysis of low sonic boom configurations; (4) large-scale flight evaluation of supersoniclaminar flow control including design codes for transition prediction, and the practical installation of suction systems.

Program Manager: Benjamin Neumann OAST/RF Washington, DC 20546 (202) 453-2795 , BLACK AND WHITE PtlOTOGRAPt-1 30- by 6O-Foot Tunnel Complete Configuration, LEF=30 deg, TEF=20 deg, Q=7.0 psf Fineness Ratio (FR) measured ahead o f wing apedfuselage juncture

Alpha - degrees

Figure 10-1. Effect of Forebody Length on Directional Stability b

10-1 Effect of Fuselage Forebody Fineness Ratio on HSCT High-

Lift Directional Stabilty

Objective. To determine the effect of in- reductions in wing area result in correspond- creased forebody fineness ratio (1engtWdiam- ing increases in relative fuselage forebody eter) on the directional stability characteris- lengthandconsegluentlyincreasesinfbrebody tics of high-speed civil transport (HSCT) con- finenessratio. Previous experiencewith mili- figurations under high-lift conditions.

taryvehicles has indicaM that high-fineness ratio forebodies can significantly affect ve- Approach. An existing model, representa- hicle directional stability characteristics.

tive of current HSCT designs and having a forebody finenessratio of 5.0, was modified to These results indicatethat advances in HSCT permit testing with alternate forebodies hav- high-lift systems may be used to accomplish ing fineness ratios of 6.5 and 8.0. Tests were reduction in wing size without adversely af- conducted in the 30- by 60-foot tunnel to fecting directional stability characteristics.

identifj. potential vehicle integration effects which may possibly limit the utility of HSCT Status/Plans. No future work is planned in high-lift concepts. this area.

Accomplishments. Results of theinvestiga- E. Richard White tion showed only minor effects of forebody Flight Dynamics Branch fineness ratio on configuration directionalsta- Langley Research Center bility for the angle-of-attack range correspond- (804)864-1147 ing to normal low-speed operating conditions.

For higher angles-of-attack,increasing fine- ness ratio was found to have a positive stabi- lizing effect on directional stability.

Significance. Projected advances in HSCT high-lift system effectiveness will permit sig- nificant reductions in vehicle wing size. Since the fuselage is sized by passenger/volume requirements (which w i l l remain unchanged) , I Figure 10-2. Computed Streamlines for Various Fence Heights on a NACA 4412 Airfoil

10-2 High-Lift Systems Research for HSCTApplication

Objective. To develop advanced high-lift Significance. The theoretical and water systems that will help reduce the community channel results are very encouraging. Aprac- noise impact and contribute to the economic ticalsystem basedonthe trappedvortexshould success of new high-speed civil transport air- eliminate the need for complex leading and craft. trailingedgeflaps, whileproviding sufliciently high lift coefficients to allow safe, controllable Approach. A combined computational and landings. This concept can generate signifi- experimentalpmgramisbeingpursued. Based cant lift coefficients at low angles-of-attack, on limited small-scaletesting, theoreticaland reducing the t a i l strike problem at takeoffand computational analyses, and a literature touchdown. Reducing the lift-off angle-of- search for suitable high-lift devices, the attack by 4" saves at least 2500 pounds in trapped vortex was selected as the most prom- landing-gear weight. In addition to lower ising technology. landing gear weight, the wing area required for takeoff is reduced. Overall, a significant Accomplishments. Potential flow analyses aircraft weight savings can be achievedwhich have demonstrated the stability conditions will result in lower engine thrust require- required for trapping a vortex on a wing. The ments and safer, quieter operation.

need for two fences to trap a vortex with very little suction at the vortex core was a signifi- Status/Plans. The computational studies cant discovery. Preliminary water-channel have shown the trapped vortex to be a viable and wind-tunnel tests have shown the effec- technique for generating high-lift at low tiveness of the two-fence configuration for angles-of-attack. A small-scale test is under simple geometries. These results also vali- way in the Ames 7- by 10-foot wind tunnel dated the computations. Additional Navier- which will define the effect that various geo- Stokes computations have shown that the metric parameters (fencespacing and height, vortex trapping process is not sensitive to etc.) have on the vortex trapping process.

Reynolds number. These results will be applied immediately in another cooperative test of the Boeing low- Cooperativewind tunnel tests were performed speed model at the Ames 7- by 10- foot wind with Boeing Commercial Airplane Company tunnel.

at the University ofwashington Aeronautical Laboratory. These tests showed that for a C. Ross James representative high-speed civil transport Fixed-Wing Aerodynamics Branch model, a double-fence configuration gener- Ames Research Center ated a tighter vortex than did a single-fence. (415) 604-6722 , Figure 10-3. Integration of the VMS and ANOPP for High-speed Civil Transport (HSCT) L Community Noise Prediction

10-3 Integration of Flight Simulation With Aircraft Noise

Prediction

Objective. To provide a tool to evaluate the Pratt & Whitney VSCE-516 engines. The effects of high-lift aerodynamics and aircraft set of aerodynamic and propulsion complete operating procedures upon high-speed civil data available for this supersonic aircraft transport (HSCT) community noise.

provides a validation for the simulator. The ANOPP code has had upgrades performed to Approach. The approach combines the flight the flight dynamics, atmospheric absorption simulation capabilities of the Langley Visual andjet mixing noise modules for better super- Motion Simulator (VMS) and the noise predi- sonic aircraft noise prediction. The VMS cation capabilities of the Aircraft Noise Pre- simulator has been operated for trial flights diction Program (ANOPP) to predict ground and the output has been utilizedin the ANQPP contours for HSCT community noise evalua- code to predict contours.

tion. The VMS is a ground-based flight mo- tion simulator with six degrees-of-freedom. It Significance. A successful high-speed civil features a variable atmosphere model and transport must be environmentally compat- computer-generated out-the-window visual ible. Advanced operating procedures, ad- scenery. It has an advanced transport-type vanced propulsion systems, high-lift aerody- cockpit equipped with a full complement of namic concepts,and acousticsuppressionhave aircraft flight and engine-thrust controls as all been proposed as techniques for commu- well as flight instrumentation displays. Aero- nity noise reduction. The effects of using dynamic control forces on the sidestick con- these new technologies must be quantified to troller and rudder pedals are provided by a ensure that a proposed high-speedcivil trans- hydraulic system coupled to an analog com- port can meet acceptable community noise puter. The simulator provides the pilot with limits with minimum economic penalty.

realistic variable-feel characteristics of air- craft stiffness, damping, coulomb friction, Status/Plans. A baseline community noise breakout forces, and inertia. The ANOPP evaluation of the AST-105 configuration w i l l code consists of dedicated noise prediction be completed to assess the impact of the ad- modules forjet powered aircraft. This predic- vanced engines and advanced piloting proce- tion program requires a detailed description dures. The simulator and noise prediction of the aircraft position and engine operating code will be updated to simulate a more cur- conditions as a fhction of time. With the rent proposed HSCT aircraft.

flight trajectories and engine propulsion re- quirements provided from the VMS, ANOPP Robert A. Golub is capable of predictingthe ground contours of Acoustics Division the noise levels associated with simulated Langley Research Center takeoff, landing, or level flight. The accompa- (804) 864-5281 nying figure conceptualizes the relationship between the two software-based prediction systems.

Accomplishments. In the initial evaluation study, the VMS has been reprogrammed to simulate the NASA designed AST-105 air- craft. This aircraft has a cruise Mach number of 2.6, carries 273 passengers, and has a range of 4500 nautical miles. It is powered by four Blended wing-bdy amfigumtion Specified target signature - I D

-_ f-

I I 1 I I I I -I .5 -4 .a 1-2 1.6 - 2 0 - 1 6 -1.2 - 8 -.4 0 t S i g n a t u r e ~ t u a l l y o b t a i n a l a c c o r d i n g t ~ d ~ ~ ~ Figure 10-4. Computed Signature Results for Low Boom Configuration

10-4 Design System for Low Sonic Boom Configurations

Significance. High-speed civil transport Objective. To facilitate and accelerate the design and analysis of low sonicboom blended (HSCT)milestones for assessment of the sonic wing-body configurations. boom problem can be met more easily ifmodel design can be accelerated. Configurations Approach, Existing codes were modified so with acceptable sonic boom characteristics would open the possibility of overland super- that all input and output geometry and array sonic flight.

formats were compatible. New codes were written and incorporated as required. One new code automatically adjusts the thickness Status/Plans. Experience in the use of the distribution to obtain the geometry required system is being gained. Some minor tailoring of the codes occurs as a result of this experi- by low-boom constraints. Computerized loft- ence.

ing was incorporated to reduce drawingboard work. U s e of a blended wing-body concept Raymond L. Barger, M a r y S. Adams eliminates the tedious problem of accounting Fluid Mechanics Division for the wing-fuselage gap that occurs when Langley Research Center they are treated as separate components.

(804) 864-2315 Graphics codes were incorporated for rapid display of relevant geometries and distribu- tions.

Accomplishments. Some sample designs have been completed. Ground-level signa- ture results for one case are shown in the accompanying figure.

B/A = .25

B/A = .50

.I B/A = .75

m B/A = 1.0

2~ 0 1

1 2

Rise time, ms

Figure 10-5. Benefit of Boom Shaping

10-5 Sonic Boom Shaping For Reduced Loudness

Objective. To quantify the effects of sonic Significance. These results indicate that boom signature shaping on subjective judg- significant reductions in boom loudness can ments of boom loudness. be achieved by detailed shaping of the boom signatures. They also provide data that can Approach. A new Langley Research Center be used in the selection of a preferred boom sonic boom simulator was used to obtain sub- shape.

jective loudness judgments h m a group of local residents of a wide range of candidate StatusPlans. Future studies willintroduce boom shapes. The shapes were front-shock a "house filter" into the boom simulation pro- minimized and covered a range of initial to cess in order to assess indoor effects. Asym- peak werpressures (0.25 to 1 . 0 0 1 , initial rise metrical boom signatures will also be investi- times (1 to 4 msec), and secondary rise times gated to determine whether asymmetry may (20 to 50 msec). Duration for all booms was provide additional reductions in perceived held constant at 300 msec. loudness.

Accomplishments. Subjectiveloudness rat- Jack D. Leatherwood ings were obtained for the range of shapes Acoustics Division shown at the top of the figure. These shapes Langley Research Center were characterized by an initial rapid rise to (804) 864-3591 an intermediate overpressure (level B) fol- lowed by a more gradual rise to peak overpressure (level A). The mean loudness ratings, as a fkction of initial rise time for four ratios of intermediate to peak overpressure,are shownin the bottom figure.

These data show that loudness, for constant peak overpressure,decreased withincreasing rise time and, for a specific rise time, in- creased with intermediate to peak pressure ratio.

Rise

SuJJective

loudness 90

level, dB

70 80 90 100 110

Perceived level, dB

F’igure 10-6. (1) Sonic Boom Simulator (L-90-5755); (2)Subjective Loudness of Sonic Booms a s a Function of Perceived Level (PL).

c -3

10-6 Prediction of Subjective Response to Sonic Booms

Objsctive. A noise metric that adequately Signi€icance. The results of this study indi- predicta the subjectiveloudnessofsonicbooms cate that a noise metric exists that effectively that have Merent physical characteristics is predictsthe subjectiveresponse to sonicboonrs

needed for the assessment of sonic boom sig- . Thus, a validated method

natures predicted for candidate high-speed comparative assessmentof the transport configllcEtions. To address this subjective acceptability of sonic boom signa- need, the sonic boom simulator shown in the tures predicted for candidatehigh-speed trans- upper figure was developed to study subjec- port configurations.

tive response to sonic b m s . This investiga- tion was aimed at assessing the effects on StatudPlanS. The sonicboom simulator will subjective loudness of the overall duration, continue to support efforts to develop a high- rise time, overpressurepand wave shape of speed transport having sonic boom signa- sonic booms. tures that permit overland supersonic flight.

Relative acceptability of various proposed "minimized" signatures will be investigated Approach. Loudness judgments of selected sonic boom signatures were obtained from 36 in future studies.

test subjects using the method of numerical category scaling. Each test subjectjudged the David A. McCurdy loudness of each of 150 simulated sonicbooms Acoustics Division on a d e from 0 to 10. The set of 150 sonic Langley Research Center booms represented different combinations of (804) 864-3596 overall duration, rise time, overpressure,and wave shape. Overall duration ranged from 25 to 425 msec with rise times of 1, 2, 4, and 8 msec. These ranges of overall duration and rise time easily encompass the values pre- dicted for high-speed civil transports. The overpressures for the different wave shapes, N-wave and"minimized,"varied from approx- imately 0 . 6 to 1.6psf. The loudness judg- ments were averaged across test subjects and converted into decibel-like subjective loud- ness levels for each of the 150 sonic booms.

Accomplishments. Analyses of several dif- ferent noise metrics indicated that perceived level (PL) was the best predictor of subjective response to sonic booms. The lower figure shows the subjective loudness levels plotted against PL for each of the four rise times.

Perceived level adequately accounted for the effects of rise time, overpressure, and wave shape. However, a small effect of duration on subjective response(-1 dB)overtheverywide duration range considered was found, which is not accounted for by PL.

N"

=r

c3 ........................

.....

.... .....

......................................

...... S ........

m - L c, _..~ .........................

. . C.

II .... ary

a

L ....,....... } .............. .........

.....

a a a

a

Figure 10-7. IndoodOutdoor Sonic Boom Simulation Facility

10-7 Indoor/Outdoor Sonic Boom Simulation Facility

Objective. To determine an acceptable level sonicboom simulation facility will provide the of sonic boom exposure for a high-speed civil unique capability of exposing a house to re- transport, human response to booms heard alistic sonic boom waveforms. It is known both indoors and outdoors is to be examined. that human response to sonic booms heard The first step in this study is the development indoors is quite different from outdoor re- of a suitable sonic boom simulator. sponse. This is due to a variety of factors such as acoustic transmission loss and noise-in- Approach. A simulator is required which duced vibration and rattle. This facility will can provide realistic sonicboom signatures at enable a systematic investigation t o be made the facade of a typical residential structure.

of these indoor/outdoor differences.

Since the fkequency content of sonic booms extends to very low frequencies,conventional Status/Plans. The simulator is currently loudspeakerarrangements are unable to gen- operational and human responsetesting, both erate realistic sonic boom waveforms. The indoors and outdoors, will begin in fall 1991.

Georgia Tech Research Institute (GTRI) was given the task to design and fabricate an Kevin P. Shepherd array of acoustic drivers that could cover the Acoustics Division frequency range from 3 H z to 4 kHz, and Langley Research Center generate an N-wave having an amplitude of (804) 864-3583 approximately 2 pounds per square foot.

Accomplishments. A sonic boom simulator has been designed that consists of 25 acoustic horns and loudspeakers. The custom-de- signed, very low frequency horns are servo- motor driven and radiate soundover the range 3 H z to 30 H z . The other, conventional loud- speakers radiate sound from 30 Hz to 4 kHz.

The loudspeakers are arranged in a vertical array adjacent to an uninhabited house lo- cated at the GTRI laboratory.

Significance. When fully operational, this @ Objectives:

0 Develop sonic boom response

Surveyed areas

questionnaire

0 Provide preliminary data on

extent of sonic boom annoyance

@ Sonic boom exposure: a Long term SR-71,0.5 to 1.0 psf,

1 per week

0 Exposure ceased 6 months prior

to study

@ Findinas:

0 Little to moderate annoyance

tartle reaction frequently noted

0 Vibration frequently noted, some

damage attributed to sonic booms

Figure 10-8. Preliminary Sonic Boom Survey ,

10-8 Preliminary Sonic Boom Survey

Objective. To develop a response question- brated, shook, or rattled during a boom. The naire for an extensive survey and to provide the most frequently mentioned items were some data on the extent of annoyance to sonic whole house, dishes, pictures, and knick- booms in areas which have been exposed for knacks. About one-quarter of the respon- many years to sonic booms with intensities dents associated sonic booms with damage to comparable'to those from proposed future their house, although only eight people in the supersonic civil transports. entire sample thought that sonic booms were an issue in their community.

Approach. "he present study developed a questionnaire for subsequent and more ex- S i g n i f i c a n c e . This preliminary survey has tensive surveys through interviews in two provided information for final development of communitiesthat had experiencedsupersonic a questionnaire for a major sonicboom annoy- overflightsofSR-71airplanesfor severalyears. ance survey. In addition, the insight on the It was estimated, based on analysis of USAF extent of annoyance due to long- term sonic files of supersonic flight operations, that the boom exposure has provided valuable but SR-71 overflights occurred on average about statistically limitedinformation as to whether once per week and produced sonicbooms with people will accept sonic booms with intensity peak overpressures of 0.5 to 1.0 psi, although on the order of those predicted for future exposures had ceased about 6 months prior to supersonic transports.

the survey. A total of 23 respondents livingin Status/Plans. A major survey of human the border area of Washington and Idaho and 22 respondents living in central Utah were annoyance is planned to be conducted in con- junction with a USAF investigation of the interviewed during July 1990. Questions that addressed the following topics were in- extent and prediction of sonic boom exposure cluded terminology for sonicbooms, extent of in theNellis Military OperatingAreaofsouth- vibration, activityinterference, links to previ- ern Nevada. The NASA portion of this study oussurveys(e.g.,OklahomaCity, 1964),startle will provide estimates of the relationship be- reaction, and extent of annoyance with links tween sonic boon exposure and community tocommunityannoyanceby conventional air- response.

craft noise.

Kevin P. Shepherd Acoustics Division hmplishments. This effort has produced a draft questionnaire that appears to meet Langley Research Center most of the goals established for this prelimi- (804) 864-3583 nary survey. In addition, the preliminary survey provides a number of findings related to the extent of annoyance from and percep- tionofsonicboomsinthe surveyareas. People were aware of sonic booms and on average gave ratings of "a little annoyed" to "moder- ately annoyed" and about equal to some other noise source. Almost all of the respondents reported that they had been startled by sonic booms and nearly half reported that their sleep had on occasion been disturbed. A high proportion of people reported that things vi- pagation - Effect of Atmosphere on h r r m I

10-9 Minimum Sonic Boom Rise lEme Determined by Absorption

Objective. T o quantify the effects of atmo- S i g n i f i c a n c e . The prediction of rise time spheric absorption and turbulence on sonic provided by the new theory is considerably boom wave forms. This research is in support better than that given by previous theory of the High-speed Research Program, which which neglected molecular absorption. Fur- is addressing the feasibilityof supersonicflight ther, the new theory provides a conservative over land.

estimate of sonic boom rise time.

Approach. Predictions of a recently devel- Status/Plans. The new absorption theory oped theory describing the effects of molecu- will be integrated into currently available lar absorptionandviscous dissipationonsonic sonic boom prediction codes. Additional com- boom rise times were compared with rise parisons with other sonicboom databases are times determined from measured sonic boom planned. Further work on atmospheric ef- signatures generated by F-16 and SR-71 air- fects on sonic booms will concentrate on the craft. A total of 116 sonic boom signatures development of the tools required to predict were analyzed in this study. The Mach nun- the effects of turbulence.

ber range was from a low of 1.14 to a high of 3.0. Flight altitudes varied between 3.84 km Gerry L. McAninch and 22.3 km, thus providing a wide range of Applied Acoustics Branch peak overpressure, which is the variable of Langley Research Center primary importance for predicting sonic boom (8O4)-864-5269 rise time. This work was carried out by Dr. A.

Pierce of Pennsylvania State University un- der a grant to consider atmospheric effects on sonic boom wave forms.

Accomplishments. The accompanying fig- ure provides comparisons of two theoretical rise time curves as a function of shock overpressure, and rise times as determined from measured sonic booms. The compari- sons show that the rise time predicted by the new theory is consistently less than that de- termined from the measured sonic boom sig- natures. Since atmospheric turbulence could not be contblled during the measurement perid, and atmosphericturbulence is thought to be a major factor in sonic boom rise times, most of the difference is currently attributed to atmosphericturbulence. This is a tentative conclusion, however, as the two data points below the predicted curve on the right-hand figure attest to the possibility of exceptional cases.

,

DAC F-16XL-2 SLFC DESIGN FEASIBILITY STUDY

-

........ -,................I-.....-

d''s' .....

24.7% ------I-. _ - - , - - r r - r - . . - r - - - - - ~ . - * - - a a 31.8 Yo

-- ____------------ F-:, 46.9%

LFC Glove ....................

..-

- -.::::- 66.0 Y O

Outline r i Mach = 1.4 Alt = 50,000 ft Relft = 1.7 million 0.30 0.3 Llmited Suction 0.2s 0.2s @ . . a 0.28 Turbulent Tnndllon o.,5 0.15 Lacailon (uc) T 0.10 0.10 0 1 0.05 0.00 0.00 4 8 I 1 1 4 I .I I I f I I Alpha, dag. Alpha, deg.

Figure 10-10. Supersonic Laminar Flow Control Program ,

10-10 Supersonic Laminar Flow Control Program

Objective. T o develop and validate technol- Significance. Industry is preparingpmpos- ogy for practical, reliable, and maintainable ds in response to the RFP. Data obtained supersonic laminar flow control concepts for from the F-16-1 flight tests are providing future high-speed civil transports (HSCTs). inputs for code calibration and transition pre- diction methodology, as well as improvingour Approach. A balanced program involving understanding of laminar flow/transition NASA and industry has been structured to physics over a highly swept supersonic wing.

carry out the supersonic laminar flow control Data firom ground facility tests, to date, are program. The program utilizes a mix of com- verifying the design methods and will provide putationalefforts, groundfacilityexperiments, data for CFD calibrationandcomparisonwith and flight testing. Advanced computational flight tests. The F-16XL2 feasibility study fluid dynamics (CFD) methods and boundary verified that achievingextensivelaminar flow layer stability codes are being developed to with a suction glove on the F-16XL-2 is fea- guide the design and for code validation. sible, but illustrated that special treatment of Swept-wingmodel experiments are under way the wing/fuselage area would be required to in low disturbance supersonicwind tunnels to achieve laminar flow while not producing provide data on leading-edgetransition phys- stability and control problems.

ics. Ongoing and planned precursor flight tests on the F-16XL-1 and -2 will obtain de- Status/Plans. Award of the F-16XL2 con- tract is anticipated for March 1992. Flight sign criteria and code calibration data to reduce the risk for the suction panel tests on testingofthe F-16XL-1is scheduledfor comple- the F-16XL-2. Flight tests of suction panels in tion at the end of calendar year 1991. Precur- FY 1995 will achieve extensive laminar flow sor tests with the F-16XL-2will be initiatedin over a range of flight conditions and provide about March 1992.

data for code validation and development of design methodologies . Michael C. Fischer Flight Applications Division Accomplishments. The RFP for the design, Langley Research Center fabrication, and installation of suction panels (804) 864-1921 and a suction system on the F-16XL-2 was released on July 29. Data from the Rockwell suction panel on the F-16XL1 are being used to calibrate CFD codes. Laminar flow to about 8.6% chord was obtained in initial flights at M=1.4 and an altitude of 50,000 feet both with low suction and without suction. Recent flights have demonstrated laminar flow to about 25%chord with full suction for similar flight conditions. The F-16XL-2 supersonic laminar flow control feasibility study was completed. Technology studies with industry were initiated. A non-suction swept wing pressure model was tested in the LaRC low disturbance supersonic tunnel.

, 20 1 DC 10-Degree Transition Cone

10-11 Flow Quality Study in the Unitary Plan Wind Tunnel

Objective. To determine the flow quality in the NASA Langley Unitary Plan Wind Tun- Si cance. The current emphasis on nel (UPWT) based on measurement of cone HSCT studies has reiterated the need to have transition Reynolds number. well-defined wind tunnel flow quality levels.

The AEDC 10" Transition Cone, which has Approach. The potential deterioration of been testedin numerous wind tunnels around the UPWT nozzle and test section wall sur- the world, provides a common reference to face finish over the past several years has led compare flow quality between facilities. The to a reevaluation of the tunnel flow quality present results indicate the effectsof poten- using the Arnold Engineering Development tial deteriorations in the UPWT flow quality Center (AEDC) 10" Transition Cone model. and/or the test model surface finish.

This model was last tested in the UPWT in 1974. The present study was conducted at the Status/Plans. Requests have been made to same Mach number and Reynolds number m D C to permit Langley Research Center to test conditions as the previous test as well as refinish the surface of the AEDC 10" Transi- tion Cone. Once refinished the model will be several additional Mach numbers concen- trated in the region of interest for high-speed retested in the UPWT.

civil transport (HSCT) studies.

Jeffrey D. Flamm Applied Aerodynamics Division Accomplishments. The AEDC 10" Transi- tion Cone model was tested in test section two Langley Research Center of the UPWT at Mach numbers of 2.36,2.86, (804) 864-5955 3 . 5 1 , and 4.60 and over a Reynolds number range of 1.0 x 106 to 5 . 0 x lo6 per foot.

Generally the current levels of transition Reynolds number were slightly lower than those obtained in the 1974 tests. This reduc- tion in transition Reynolds number may be a result of increased tunnel turbulence levels causedby deterioration of the nozzldtest sec- tion wall f i s h or perhaps other flow distur- bance producing mechanisms (e.g., turning vane induced disturbances). However, the AEDC 1 0 " Transition Cone model surface finish was found to have numerous imperfec- tions and may have caused the relative reduc- tion in transition Reynolds number.

, ,

Chapter 11

Aerothermodynamics Research and Techndogy

The Aerothermodynamics Research and Technology program focuses on advancing both our understandingof and ability to address the issues associated with high-temperaturegas effects as they impact the aerodynamic and heating environments encountered by vehicles and spacecraf% for both Earth and planetary missions. The problems of high Mach number flight throughout the continuum, transitional, and free molecular flow regimes are of primary importance to the program. For the Earth-to-Orbit (ETO) vehicles, which encounter Mach numbers as great as 25, a significant amount of dissociation and chemical nonequilibrium exists at the high altitudes (i.e., above 50 km). Also, because many ET0 vehicle concepts must return to Earth for landing, the relationshipbetween high-speed aerothermodynamicefficiency and low- speed flight performance must be investigated for each cofiguration. For more energetic missions that may involve probes or aeroassist space transfer vehicles (ASTVs), flight Mach numbers as large as 50 are encountered, and ionization, radiation, and thermochemical nonequilibrium can be significant to the vehicle designs.

The Aerothermodynamics R&T program addresses the issues through emphasized research in four primary areas: computational tool development; experimental research and computational validation; facilities research and development; and configuration assessment. The computa- tional tools developed include not only computational fluid dynamics methods but also computa- tional chemistry methods which provide a unique ability to calculate high-temperature gas properties. The vehicle synthesis engineeringtools developed are becoming industry standards for preliminary design and analysis. Application of aerothermodynamics technology to vehicle designs will result in reduced flight environment uncertainty, optimized configurations, and improved performance margins.

Program Manager: Jim Moss OAST/RF Washington, DC 20546 (202) 453-2820

AFE Equivalent Sphere Radiative Flux

cv

E

0 3

\

Tangent Slab Transport (I-D)

. 1 1 . 1 1 1 1 1 ( 1 MDA Transport (3-D)

-3

- 2

odified Differential Approximation

(MDA) Shows Potential For

Solving 3-0 Radiative Transport

a

I 5 10 15 20 25 30 35

s, cm

Figure 11-1. Advancement in Radiative Transport ,

11-1 Advancements in Radiation Transport

Objective. To develop improved methods to and the developmentof the 3-D MDA d e are predict radiative heating over three-dimen- an excellent start to predicting and interpret- sional configurations in thermochemical ing Aero-Assist Flight Experiment results, nonequilibrium flows.

and in the application to aerobrakes in the Space Exploration Initiative.

Approach. A radiation transport code was developed that minimizes the spectral array StatudPlans. Efforts are continuing to im- required for adequate resolution of the radia- prove the accuracy and efficiency of the re- tive flux. Techniques were defined to opti- centlydeveloped 3-D radiativetransport meth- mize the couplingof the radiative transport to ods. The methods will be further demon- a flow-field code. A method was developed to stratedthough applicationsat additionalflow efficiently account for 3-D radiation effects conditions and vehicle configurations.

over complex configurations.

Accomplishments. A nonequilibrium ra- Lin C. Hartung diative heating prediction method, the Lan- Aerothermodynamics Branch gley Optimized Radiative Nonequilibrium Langley Research Center (LORAN) code has been developed. Numeri- (804) 864-4371 cal studies have been performed to determine the minimum spectral points required for adequate predictions of the radiative flux.

The results have been compared with ground- based and flight measurements, and with the detailed Park NEQAIR code, with good agree- ment. I n i t i a l studies have demonstrated the feasibility of using the LORAN code coupled with the flow-field code LAURA. A Modified Merential Approximation (MDA) method has been adopted for 3-D radiative transport, and has been applied successfully for axisymmetric, blunt-body cases.

Significance. High-temperature radiative transport at equilibrium conditions over axisymmetric blunt bodies at zero angle-of- attack with the one-dimensionaltangent-slab approximation is reasonably advanced over earlier studies on Earth entries and the work on the Pioneer Venus and Project Galileo probes. However, for aerobraking applica- tions, the radiative transport is expected to be at thermochemical nonequilibrium conditions over complex 3-D configurations at non-zero angles-of-attack. Current advancements in the development of the nonequilibrium LO- R&$ code, coupled to the LAURA flow field,

z"

a

0 W lL

E

P= W

e

0 0 0 0 0 0 r: 9 v! 3 T c'!

0 0 0 0 0 0 n > v a m IC.

z"

v) VI

z

Y c- 01 I I

0 j

Q)

0 o j $

0 , ' I

_____------

0 0 0 0 0 0 L? 3 c ! c'! T : 9 0 0 0 0 0 0 n > W n U

. Energetics of Gas-Surface Interactions in Transitional Flows at

Entry Velocities

11-2 Energetics of Gas-Surface Interactions in Wansitional Flows

at Entry Velocities

Objective. To provide fundamental infor- StatudPlans. Additional studies to M y mation on the velocities and energies of mol- characterize the translational velocity distri- ecules incident on a surface behind a shock butions throughout the flow field are planned layer in transitional flows at entry velocities for the purpose of evaluating possible means and to provide data to support experiments on of coupling DSMC and Navier-Stokes solu- proposed tethered satellites. tion methods.

Approach. The direct simulation Monte Carlo (DSMC) method was used to model the Richard G. Wilmoth, James N. Moss flow about a 1.6-m diameter sphere in the AerothermodynamicsBranch altituderange of 130 to 90 km for E a r t h entry Langley Research Center at 7.5 km/s. Samples of the velocity and (804)864-4367 translationalenergydistributionswere taken for gas molecules strikingthe surface at vari- Virendra K. Dogra ous points on the sphere and for each gas Vigyan Research Associates species. (804) 864-2957 Accomplishments. Velocity distributions were computed for each of five gas species at five different altitudes as a function of the circumferential location on the sphere. The computations required the sampling of more than one million molecules at each condition.

The results will be published in AIAA Paper 91-1338.

Significance. Proposedexperiments in tran- sitional,entry flows will require sophisticated instrumentation to study the details of gas- surface interactions. These DSMC results show the wide range of incidentvelocities and energies that must be considered in designing such experiments. The results also demon- strate the high degree of translational nonequilibrium present behind the shock layer.

, 1 k 0 2 k A 3 k A 4 k CPU time, s Figure 11-3. Convergence History: Six Tasks Adapting Partition Boundaries ,

11-3 Asynchronous Macrotasked Relaxation Strategies for the

Solution of Viscous Hypersonic Flows

Objective. T o exploit parallel, multitasking S i g n i f i c a n c e . Simulations of moderately capabilities of CRAY-class supercomputers to complexentryvehicles(i.e., Hd20, aerobrake- improve efficiency of numerical simulations payload assemblies) have large memory re- of viscous, hypersonic flows in thermochemi- quirements, sometimes requiring dedicated cal nonequilibrium. use of a supercomputer. The present asyn- chronous,macrotasking strategy enables com- Approach. an upwind-biased, point-implicit plete utilization of all processors in the dedi- relaxation of the governing equations was cated mode with speedup proportional to the implemented. Relaxation factors were inde- number of processors for coarse p a i n ma- pendently defined for the viscous and inviscid chines like the Cray-2. Furthermore, while a contributions to the solution. Computational single Cray Y-Mp does not have sufficient work was dividedinto several, independently memory to execute these simulations, net- executed tasks through partitioning of the working several such machines to provide computational domain using CRAY even faster turnaround is theoretically pos- macrotasking. The point-implicit relaxation, sible.

which limits the computational stencil, in combinationwith the appropriate relaxation Status/Plans. These strategies will be ap- factors eliminated the need for synchronized plied to STS and HL-20 flow-field simula- communication between the tasks for steady tions.

flow problems.

Peter A. Gnoffo AerothermodynamicsBranch Accomplishments. Solutions for hypersonic flows over aerobrakes, including the wake Langley Research Center flow, have been obtained using as many as (804) 864-4380 eight tasks executing asynchronously. Sacri- ficing synchronization on coarse grain ma- chines has shown no adverse effects on total CPU time required to obtain a solution, and reduces turnaround time for a given job.

Adaptive partitioning provides further, mod- erate improvementin required CPU time by moving partition boundaries to concentrate task activity in areas where the contribution to the solution error is the highest. The accompanying figure shows the convergence history for each of six tasks applied simulta- neously and asynchronously to the solution of hypersonic, reacting flow over an aerobrake.

The pattern is somewhat chaotic, because of the lack of synchronization and moving parti- tion boundaries, but the overall trend is de- crasingby approximatelya factor of six times faster in turnaround time than a single task application.

, 21 1 3.

CL

u

b

n

M CL pc

E

u

n

0 0 0 W D Q W N - ~ ' D V N O _ - _ - 0 0 0 0 Figure 11-4. Particle Simulation in a Multiprocessor Environment

11-4 Particle Simulation in a Multiprocessor Environment

Objective. To develop a particle simulation proachandthescalabilitytolargenumbersof code that is portable to a wide class of vector processors enables larger-scale applications and multiprocessor supercomputers and ca- of direct particle simulations than previously pable of addressing large-scale rarefied flow possible.

problems.

Approaeh. Direct particle simulation meth- Status/plans. The current version of the ods model flows as a collection of discrete developedparticle simulationcode will port to particles that interact with each other and several parallel MIMD and vector machine with boundaries through collisions. The colli- architectures. Grid adaption techniques will sion-selection rule used is highly compatible be employed to increase resolution in areas of with the requirements forvectorization.These the flow having larger gradients. More ad- methods have been implemented on the 128 vanced models for thermal and chemical re- processor Intel iPSC/860, and a dynamic do- laxation, now under development, will be in- main decomposition is utilized to distribute corporated in the current code.

the simulation across processors and to bal- ance the load.

Jeffrey D. McDonald Accomplishments. A new code has been Eloret Institute developed that can take advantage of me- AerothermodynamicsBranch dium-grain parallelism through a domain Ames Research Center decompositionwhile retaining the vector com- (415) 604-1140 patibility of previous codes. This provides a capability to examine three-dimensional flow problems involving gas mixtures in therm+ chemical non-equilibrium. The code is por- table over a wide class of Multiple Instruction MdtipleData(MIh4D)machine architectures.

Speedup as additional processors are devoted to the simulation remains linear to the f u l l 128processorsof the iPSC/860 tested to date.

Significance. With continuing interest in hypersonicflight, a need exists for a computa- tional capability to predict the aerodynamic and thermal environment found around ve- hicles such as the National Aero-spacePlane (NASP)or the Aero-AssistFlight Experiment (AFE). Because of the difficulties in applying the continuum equations to the very low- density hypersonicregimesthese vehicles will encounter, application of direct particle simu- lation methods is an appropriate consider- ation. The numerical efficiency of this ap- h 300 Experiment cp % Numerical (continuum) c c .

a

* e I I I 1 1 I

0.0 0.2 0.4 0 . 6 0.8 1 .o 1.2 1.4 1.6

Radial Distance (cm j 11-5, Experimental and Numerical Analyses of Small Rockets (Transverse Pitot , Pressure Profiles at Nozzle Exit)

11-5 Numerical and Experimental Analyses of Small Rocket Flows

Objective. To provide experimental and nu- StatdPIans. The particle method computa- merical data for accurate prediction of rar- tions will be extended further in the trans- efied flow of small, low-thrust rockets used for verse direction and into the backflow region control of spacecraft. behind the rocket. Furthermore, solutions will be generated through the same nozzle for Approach. Expansion of diatomic nitrogen the expansion of argon for comparison with through a low thrust rocket was considered. additional experimental measurements.

A small nozzle flow provided conditions that pass &om continuum to rarefied. To assess Iain D. Boyd rarefaction effects, the nozzle flow was com- Eloret Institute puted with both a traditional continuum CFD Aerothermodynamics Branch technique and a stochastic particle simula- Ames Research Center tion method. The plume was also calculated (415) 604-4907 with the particle method. The computations were compared with new experimental data Paul F. Penko for Pitot pressure and flow angle. Lewis Research Center (216) 433-2404 Accomplishments. Comparison of pitot pres- sure at the nozzle exit reveals that the par- ticle method predictions agree very well with the experimental data, whereasthe continuum solutions overpredict the pressure by 25%.

The particle method solutions continue to agree with experimental data further out into the expansion plume. The numerical efficiency of the particle method has been improved substantially by restructuring the algorithms to suit the architecture of vector supercom- puters. A typical run of the particle method code required 3 CPU hours and 24 Mwords of central memory on a Cray Y-MP.

Significance. The dserences in the con- tinuum and particle method solutions will increase as the plume expands. The present study reveals that significant errors will be incurred through the prediction of such flows with continuum CFD. To obtain accurate so- lutions of small plumes, and hence accurate assessment of the interaction between the plume and the spacecraft, it is necessary to employ the particle simulation technique.

35" Ramp, p cx) = 10.72 x 106-65 kg /m3

M 00 = 24.26, h 00 = 0.230 mm, T, = 394.2 K

DENSITY (plp,) "'-

x = 71.4 mm Y 9 m m x, m m x, mm

STREAMLINES MACH NUMBER R 2 2 0

19 0 10 16 0 13 0 9 40 10 0 40 8 7 30 7 0 Y , 30 6 K 4 0 2 0 mm 20 m m 2 0 1 0 3 10 0 5 2 0 2 1 0 x, m m x, mm Figure 11-6. Hypersonic Rarefied Flow About a Compression Corner ,

ll-6 Hypersonic Rarefied Flow About a Compression Corner

Objective. To numerically simulate the phe- . The capability to accurately

nomena of shockhundary layer interactions calculate the phenomena associated with and compare the numerical results with the shockhoundary layer interactions is critical experimentally measured data obtained in for the design of future space transportation the DLR wind tunnels at Gottingen, Ger- vehicles. The present study gives added con- many. fidence in the ability of the DSMC method to simulate such flows for hypersonic rarefied conditions.

Approach. The focus is on the merged-layer interactions that occur for hypersonic (M, = 2O)nitrogenfbwover a two-dimensionalmodel Status/Plans. Results of the study have consiBting of a flat plate followed by a com- been published in AIAA Paper 91-1313.

pression ramp. The ramp comer is located 71.4 mm from the plate leading edge, and James N. Moss, Joseph M. Price ramp angles of 0 " , 15", 25", and 35" are consid- Aerothermodynamics Branch ered. Freestream Reynolds number based on Langley Research Center model length ranges from 6,000 to 39,000. (804) 864-4379 The calculations are made with a 2-D version of the direct simulation Monte Carlo (DSMC) Ch.-H. Chun method. DLR Institute for Experimental Fluid Mechanics Accomplishments. Calculations have been Gottingen, Germany made for four test conditionsusing the DSMC method. Results of these calculations have been compared with various experimental measurements to provide both quantitative and qualitative comparisons. The experi- mental and computational findings agree on the combination of flow conditions and ramp angles necessary to produce separation.

,

DSMC I

Free Molecular Limit \ = - = g = = = = - ~ - - = - = = Continuum viscous \ niversity Of

0. - \

Maryland

- - -A- - Continuum Viscous

A%. %% With Bridging Formula

! %%

0.

f

5 '

f '.

.4

.2

e 11-7. Rarefied Flow gime Hypersonic Waverider Configuration, I

11-7 Advancement in Developing an Efficient 3- DSMC Code

Significance. The computational efficiency, Objective. To significantly improve our present capability to calculate rarefied flows code setup procedures, and memory manage- ment of the F3 code are far superior to our over complex configurations at low altitudes.

earlier 3-D DSMC code, with no loss in accu- racy. Our present simulation for the delta Approach. The capability of a new direct wing with the F3 code was set up in a few days simulation Monte Carlo (DSMC) code, de- noted as F3, was further developed, demon- and required 1 to 2 days on a 32-bit worksta- strated, and validated and flow fields were tion as contrasted with several months to simulated in the rarefied flow regime about setup and one Cray CPU day with the earlier complex configurations in three dimensions. code. The present results with the F3 code show that the high values of lift-to-dragpre- dicted for a waverider concept are signifi- Accomplishments. Significant code im- cantly reduced at high-altitude,rarefied flow provements have been made in the areas of surfacedefkition and grid generation for code conditions.

setup, graphical displays for diagnosis proce- dures, and post-processingand analyzing re- Status/Plans. Future improvements to the code will include adaptive gridding, a vari- sults, and in data compressing techniques to able inner mesh resolution,improved memory reduce memory size. F3 was used to simulate management, and optimization for vector and the flow field over a delta wing at angles-of- parallel computing. Code will be demon- attack from 0" to 30" for a Mach 20 flow at a strated for selected 3-D configurations.

Knudsen number of 0.016. The aerodynamic coefficientswere in excellent agreement with ground-based experimental data, thus pro- Didier F. G. Radt viding partial validation of the code. Solu- Aerothermodynamics Branch Langley Research Center tions had been obtained over a hypersonic (804) 864-4388 waverider configurationproposed by the Uni- versity of Maryland that was optimized for Mach 25 at an altitude of 90 km. Also, solu- tions were obtained using an approximate bridging technique to correct for rarefied flow effects at higher altitudes. Finally, solutions were obtained with the current F3 DSMC code at altitudes from 97 to 145 km. The F3 lift-to-drag ratios were much lower than those predicted by the University of Maryland.

Q) 1L.

! 2

e

at

e

<D d a cu

d"

d

I1

Figure 11-8. Shuttle Orbiter Pressure ,

11-8 Flow-Field Computations for Shuttle Orbiter

Objective. To obtain flow-field solutions for StatWIans. Viscous flow solutions with the shuttle orbiter adequate for comparison non equilibrium chemistry will be obtained at with the flight data h m the Shuttle Infrared actual shuttle flight conditions. The compu- Leeside Temperature Sensor (SILTS) experi- tational results will be compared with SILTS ment. flight data and data h m the Developmental Flight Instrumentation flights.

Approach. Improved gridding techniques were used to define the surface and volume W i l l i a m L. Kleb, K. James Weilmuenster grids. The Langley Aerothermodynamic Up- Aerothermodynamics Branch wind Relaxation Algorithm ( L A W ) code Langley Research Center was used to obtain the flow-field solutions. (804) 864-4364 Accomplishments. A grid and flow-field solution has been obtained for inviscid flow at a40° angle-of-attack for Mach 7.4 flow, awind tunnel case. The computed surface pressures are in excellent agreement with the experi- mental data on both the windward and lee- ward surfaces. The results identifieda small region of non-smooth grid surface near the aft end of the vehicle. This s m a l l grid problem has been resolved, and a grid for a viscous solution has been completed.

Sisnificance. The first computational solu- tion has been obtained for the complete or- biter configurationat a hypersonic flow condi- tion. The analysis of the results for the wind tunnel case provides confidence that the grid technique and the LAURA code can provide excellent results; thus, we can proceed with more computer-intensive cases and analyze viscous flow with nonequilibrium chemistry at actual shuttle flight conditions. T h i s solu- tion also adds significantly to our database of computational results for complex configura- tions such as the HG20, National Aero-Space Plane, and the Aero-Assist Flight Experi- ment vehicle.

I "

Figures 11-9. Approximate Flow Field and Heating Analysis

U-9 Approximate Heating Analysis

Objective. To develop rapid but reliable S i g n i f i c a n c e . Successful demonstration of approximateflow-field balysis capabilityfor the approximate flow-field technique for 3-D application in conceptual design studies and configurations will significantly advance the preliminary aerothermal design. useabilityofengineeringmethods.The analy- sis would provide the aerothermal commu- Approach. Boundary layer techniques were nity w i t h a unique approximate method for incorporated in a versatile inviscid method computing the heating rates over 3-D blunt for heat-transfer predictions over 3-D sur- noses needed in thermal design studies. Cur- faces at various angles-of-attack. rently, more detailed methods require many hours of computation.

Accomplishments. An inviscid method has been developedincorporating the normal pres- Status/Plans. The boundary layer methods sure approximation of Maslen. The method are presently being included in the inviscid has been demonstrated to yield rapid but method for application to 3-D vehicles. Com- parison of results will be made with a detailed reliable inviscid solutions over axisymmetric and 3-D surfaces at angle-of-attack.The code Navier-Stokes method.

requires 10 seconds of computer time on a workstation to obtain a solution over a 3-D Christopher J. Riley blunt nose. Boundary layer approximations Aerothemodynamic Branch Langley Research Center have been included for axisymmetricvehicles (804) 864-4387 at angle-of-attackfor (1) perfect gas and equi- librium chemistry and (2) laminar and turbu- lent heating calculations.

am'

/ I

a n "ti,

$ 5

N k i

$ 5

Figure 11-10. Heating on Modified Shuttle Orbiter ,

11-10 Comparison of Heating Rate Calculations with Experimen-

tal Data on a Modified Shuttle Orbiter at Mach 6

Objective. To provide detailed information ward symmetry plane, in a circumferential on the ability of an "engineering code" based direction around the model, and in the high on combined inviscid and boundary theory heating region along the fuselage or wing AA3DBUHALIS (fisymmetric h a l o g for leading edge.

Xhree-Qimsnsional Boundary Layermgh a bviscid hlution) and a "benchmark" Significance. These results give cordidence Navier-Stokes code LAURA (Langley that both AA3DBb'HALIS and LAURA can

Aerothermodynamic Upwind &laxation a- be used to predict more accurately the heat-

gorithm) to predict heating rates from the ing environment of future entry vehicles and windward symmetry plane and in the vicinity should lead to less conservatism in design. In of wing leading edges on winged vehicles at addition, since AA3DBUHALIS require only 1% of the time required for similar LAURA large angles-of-attack.

calculations, it should be very useful for per- forming parametric studies.

Approach. Heating rate calculations from AA3DBL/HALIS and LAURA were compared with experimental dataobtained at Mach 6 on Status/Plans. AA3DBUHALIS andLAURA a modified Shuttle Orbiter model. This is the will be compared with wind tunnel data at other flow conditionsand with availableflight first step in a process designed to validate the codes for application to wingedvehiclesin the data.

flight environment.

H. Harris Hamilton 11, Francis A. Greene, Accomplishments. Comparisons have been K. James Weilmuenster made at 30" and 40" angle-of-attack for condi- Aerothermodynamics Branch tions where the flow over the model was Langley Research Center completely laminar. Both codes compare well (804) 864-4365 with the experimental data along the wind- Figure 11-11,Computed Meridional Pressure Contours from Mach 5.04 Flow of Premixed ' Hydrogen-Air Over a Blunt Body Using a 181x127 Grid (4 Frames of an Oscillation Cycle).

11-11 Numerical Simulation of Unsteady Shock-Induced

Combustion

Objective. To develop a computational lation increases the accuracy of numerical method that will lead to the simulation and calculations in flows with exponentiallyvary- understanding of the unsteady shock-induced ing species concentrations.

combustion observed in ballistic range ex- periments. Significance. The physical mechanism caus- ing unsteady combustionin supersonicballis- Approach. Ballistic range experiments pro- tic range experiments is now better under- duce a periodic, unsteady flow when a blunt stood. This w i l l lead to improved simulations projectile is fired supersonically into a of combustion processes for hypersonic sys- premixed hydrogen-air mixture (Frame 1 in tems.

the figure). This unsteady behavior is caused by the combustion of the mixture, which has Status/plans: The current formulation will been heated by the bow shock in front of the be applied to other problems in which shock- projectile. Observations show that the heat induced combustion or ignition is important.

release due to combustion does not occur im- Examples are flows in shock-tunnelsor ram- mediately behind the bow shock but is de- accelerators. Application of this work to other layed by the finite rate of the chemical reac- types of flows with exponentially changing tions in the high-speed flow. The region species concentrations, such as ionization between the shock and the heat release is processes, may also be valuable.

called the induction zone. There is little or no heat release in this zone, but the concentra- tions of chemical species important for igni- Gregory J. Wilson, Myles A. Sussman tion increase exponentially. Previous work Aerothermodynamics Branch hasshown that thechemistryin theinduction Ames Research Center zone must be modeled accurately to simulate (415) 604-4228 these flows. The grid spacing requirements for sufficient accuracy are so stringent that simulating these unsteady flows has not been feasible. A trans-formation of the species continuity equations has been found to reduce grid requirements and make such simula- tions practical. This trans-formation is com- patible with current numerical methods.

Accomplishments. This work represents the first known numerical simulation to in- corporate a detailed chemical reaction model capable of simulating the unsteady shock- induced combustion phenomenon seen in bal- listic range experiments (Frames 2 and 3 ) . A new formulationofthespeciescontinuityequa- tions has been developed and incorporated in a proven numerical scheme. This new formu-

Airfoil fins

UD

0 4 8 12 16 20 24 28

a deg

Figure 11-12. Subsonic W D Improvements of HL-20

11- 12 A e r o d ~ a ~ c / A e r o t h e ~ o d ~ a m i e C harac teristies

of HL-2QMZ-20 A/s Lifting Body Configurations

Objective. To experimentally determine (lift-to-drag ratio increased approximately aerodynamidaerothermodynamic character- 20%); at subsonic conditions, the effect of istics ofanHG20liftingbodyconfigurationat Reynolds number vanished for values in ex- hypersonic to subsonic flow conditions (i. e., cess of 8 million per foot in the LTPT. Aerody- entry to landing), and to evolve optimum namic performance of HG20s free of real-gas lifting body configurations from aerodynamic effects primarily because of flat windward perspective via modification of aerolines and surface (unlike the shuttle orbiter, which ex- control surfaces. perienced significantreal-gaseffectson pitch- ing moment). The HG20A concept provides Approach. Force and moment measure- improved subsonic aerodynamicperformance, ments were performed on the HL-20 configu- but must reenter at high incidence to match ration at subsonic(Low-TurbulencePressure 1iNdrag (i.e., cross range) of the HL-20 and Tunnel [LTFTI, 7- by 10-ft.Tunnel),transonic the will experience significantly higher heat- (Calspan Transonic Tunnel, 8-foot Transonic ingin the nose region as a result of steamlining Tunnel), supersonic Unitary Plan Wind Tun- of the nose.

nel [VPUrrl, and hypersonic flow conditions (HFC) to determine the effects of compress- Statufllans. We plan to complete wind tunnel testing of HL-20, HL-BOA, and HG ibility and viscosity and to simulate real-gas effects. Thermal mappings and flow visual- 20B concepts; reduce and analyze data; and disseminate results. Will initiate CFD code izations were performed at hypersonic condi- tions to determine aerothermodynamic char- calibration study at hypersonic conditions for HL-20 configuration.

acteristics. HL-20 aerolines, including con- trol surfaces, were modified to enhance aero- dynamic performance with no sacrifice in Bernard Spencer, Jr., George M. Ware, N.

Ronald Merski aerothermodynamic performance.

Experimental Hypersonics Branch Accomplishments. HL-20, HL-BOA series Langley Research Center (fourdifferentconfigurationshavinga stream- (804) 864-5245 lined upper surface including canopy, wind- wardsurfacechamber, andreducedbase area) configurationswere testedextensivelyat sub- sonic, supersonic, and hypersonic conditions.

Tests of the HG2OB series (second iteration; flat-like upper surface and wings faired into body) and HG2OI"itan 4 ascent configuration were initiated in the MSFC 14-inch tunnel.

S i g n i f i c a n c e . Unprecedented ground-based data have been established for HL-2OIHL- 20A lifting body configurations, clearly sur- passing previous phase A studies. Aerody- namic characteristics were determined from Mach 20 to 0.1 (entry to landing). Aerody- namic performance of HL-20 at subsonic speeds has improved significantly via modifi- catiod of the wing fi-omcylinder-slab to airfoil

Chapter 12

Aerobraking

Aerobraking can provide a substantial increase in payload or decrease in initial mission mass for missions that require a trajectory deceleration in the vicinity of an atmosphere-bearing planet. The objective of the Aerobraking Technology program is to develop and validate key supportingtechnologies which allow the effective use of aerobraking for manned lunar missions, robotic and manned round trip missions to M a r s , and other planetary exploration missions. The term aerobrakingin this context includes 1) capture into a planetary orbit, 2) reduction of orbit altitude, 3) direct entry from a hyperbolic trajectory, and 4) entry from orbit. The key supporting technologies include aerothermodynamic modeling; high temperature thermal protection sys- tems; adaptive guidance, navigation, and control; and lightweight structures that minimize or eliminate in-space construction and/or assembly. Aerobraking systems are strongly affeGted by each of these supporting technologies. System analysis that addresses missions, vehicle concepts, and operations is used to integrate the discipline areas, to assess technology require- ments and trades, and to ensure that system requirements are satisfied.

Program Manager: Jim Moss OAST/RF Washington, DC 20546 (202) 453-2820 ..

.- 8

a , --

c .

v)

.-

2 s

t

I 2

0 II II t 0 0 O 0 0 0 0 T- cu

(D m v) *

, Figure 12-1. Additional Post-Aerocapture AV Requirements Resulting From Multiple Off-Nominal Effects

12-1 Effect of Atmospheric and Aerodynamic Uncertainties on

Manned-MarsAerobrake Feasibility

0bjective.To develop astate-of-the-& M a r s total AV budget. The algorithm was also used aerobraking guidance algorithm capable of in support of the Mars Atmospheric Knowl- managing off-nominalatmospheric and aero- edge Working Gmup (MAKWG) to help char- dynamic conditions. acterize the significance of off-nominal atmo- spheric conditions.

Approach. A predictor-corrector guidance algorithm with deceleration feedback was Si ce. With the use of an adaptive developed and implemented into a three-de- guidance algorithm, atmospheric and aerody- gree-of-freedomsimulation to assess the im- namic uncertainties can be adequately man- pact of uncertainties in the M a r s atmospheric aged and do not pose mission failure concerns density profile and vehicle aerodynamics on that would preclude the use of aerobrakingin mission feasibility. Additionally, because the a manned M a r s mission.

mission is manned, a load-relief capability was incorporated such that the maximum Status/Plans. The results of this study have sensed deceleration did not exceed 5 g's (1 g is been documented in AIAA Papers 91-0058 theaccelerationofabodyat restontheEarth's and 91-2873 as well as the final report of the surface). MAKWG. Further refinement of this analysis includes incorporation of the algorithm into a Accomplishment. The predictor-corrector six-degree-of-freedom Monte Carlo simula- algorithm has demonstrated the capability to tion to better understand the coupled effect of accurately guide a Mars aerobraking vehicle various off-nominal conditions and identify through a wide range of off-nominal condi- the vehicle control implications.

tions (whileprovidingload-relief) to the proper exit orbit. The off-nominal conditions simu- R. D. Braun, R. W. Powell lated include errors in the atmospheric den- Vehicle Analysis Branch sity profile(bothvertical and horizontal struc- Langley Research Center ture) as well as a mispredicted trim angle-of- (804) 864-4507 attack. The figure shows the amount of addi- tional AV required to insert the vehicle into the proper parking orbit as a result of various combinationsof off-nominaleffects. The cases illustrated in this figure are representative of several worst-case scenarios;however, in each case, the mission was completed successfully.

N o t e that a contingency AV budget on the order of the nominal AV (75 d s e c ) would be more than adequate to account for post- aerocapture orbital trim after atmospheric flight through an off-nominal environment.

In comparison to the M a r s departure propul- sive requirements (on the order of 2-3 km/ sec), this is a relatively minor increase in the

w

m

C G

c

c L 4 v1 L;;1

E

h

L -

l W I Figure 12-2. Heat Shield Erosion in a Dusty Martian Atmosphere

12-2 Heat Shield Erosion in a Dusty MartianAtmosphere

Objective. To determine the extent of heat quency of particle impacts were calculated at shield erosion from dust particle impacts that several points along atmospheric trajectories occur on a large, blunt aerobrake configura- with entry velocities of 7 and 9 Msec and tion during aemcapture maneuvers at M a r s vehicles with ballistic coefficients of 30 (see Figures. 2b, 2c) and 40 kg/m2.

and the M a r s Environmental Survey probes entering the atmosphere along ballisticflight paths. Significance. The presently available ero- sion results for t h e manned vehicle indicate Approach. An explicit Navier-Stokes code that the design of the heat shield can be was used to compute the flow field about the significantly affected by atmospheric dust.

vehicle shapes shown in Figures la and 2a.

For example, it appears that glass-coated Calculationsusing a particle size distribution tiles are unsuitable heat shield materials for ranging from 1 to 19 pn were made. The high-speedMars entries unless coated with a paths, deceleration,heating, and sublimation protective material. For AVCOAT, the TPS of silicate (SiO2) dust spherules which enter mass may have to be increasedby 25% to 29%, the forebody shock-layer at various locations or about 1% of the vehicle mass, for entry at were calculated and traced to surfaceimpact.

8.6 km/sec. For the MESLJR probe the results Two heat shield materials were considered: a clearly show that the surfaceinsulationglassy glass surface that is representative of Shuttle tile, which has a 0.35 mm glass coating, re- ceramic tiles and a low-density ablator, quires a protective layer of TUFI (matrix of AVCOAT, that was used on the Apollo cap- silica and alumina fibers). For the AVCOAT sule.

ablator, the heat shield mass must be in- creased by about 30% to account for erosion Aecomplishments. The dustless flow field during or shortly after a dust storm. There- was computed at a number of points along fore, the vehicle’s heat shield must be de- several entry trajectories for the manned ve- signed to account for dust particle impact hicle and the NLESUR probe. Subsequently, damage.

dust particles were introduced into the shock layer and their trajectories were traced. The Status/Plans. The previous research will be approximate surface erosion is shown in Fig- extended to implement a coupled flow model ures l b and 1c for the manned vehicle as a to assess gas-solid chemical coupling and ex- function of radial distance from the stagna- amine the effect of micron-sized particles on tion point. For a vehicle with m/CDA = 200 kgl the vehicle heat shield for differentinsulation m2, over 9 mm of glass can be destroyed,while materials. The latest results will be reported for ablating AVCOAT the erosion exceeds 10 at the AIAA Aerospace Science Meeting in mm in the stagnation region.

January 1992.

This research was extended to examine the Periklis Papadopoulos, Michael Tauber heat shield erosion caused by atmospheric Aerothermodynamics Branch dust for small, unmanned, probes entering Ames Research Center the Martian atmosphere along ballistic flight (415) 604-1146 paths. Heat shield erosion, energy and fre- ,

Previous

Studies

Lower ET0 Launch Costs Lower Development,

(BasedonWlb) . 4 - 1

Operations, and Production Costs

This

MDSSC

Lower ET0 Launch Costs Lower DDT&E Costs

Study

Less LLQ Propellant Boiloff Net Savings No Aeropass Software Heritage for Mars A/B Less Mission Analysis AI8 Pro~uctio~As~mbJy = Extra Tanks Figure 12-3. Aerobraked Lunar Transfer Vehicle (LTV) Cost and Operations Study I

12-3 Aerobraked Lunar Transfer Vehicle Cost and

Operations Study

Objectives. T o use cost and risk analyses to Significance. Viability of the aerobrake for compare all-propulsive versus aerobraked the lunar mission has been examined in terms concepts for a lunar transfer vehicle (LTV) of life cycle costs. T o date, estimates had and to use resultant cost and risk knowledge shown that the mass savings from aerobrakes to define technology development roadmaps. were significant, but the question remained whether resources invested into aerobrake Approach. Develop life cycle cost (LCC) development programs exceeded these sav- models with sufficient depth of detail in all ings. After in-depth estimates of major phases of LTV operations and in technology aerobrake DDT&E and operations issues, development programs to capture differences aerobrake options have been shown to be between aerobraked and all-propulsive L W competitive with all-propulsive approaches concepts. U s e risk analysis to identify critical for the lunar missions. New, larger launch technology test requirements and risk-reduc- vehicles or fewer missions could alter the ing design changes. results.

Accomplishments. Relative life cycle cost Statdlans. Additional analysis and cost- estimates show a 7%-10% (inside uncertainty ing studies will address single-use ablator band) cost advantage for an aerobraked LTV designs and aerobrake disposal, and identifi- when compared with an all-propulsive con- cationoflunar aerobrakedevelopment spinoffs cept based on the 90-Day Report Option 5 for future Mars missions will be made.

lunar mission (25 years, one mission per year, originating in low-Earth orbit). Cost savings Lawrence F. Rowell stem hmreduced Earth-to-orbitlaunch costs, Space Systems Division reducedlunar orbit propellant boiloff,reduced Langley Research Center gravity losses, and fewer on-orbit operations (804) 864-44502 tasks (marginal advantage; assumes three- piece erectable aerobrake). Cost increases stem fkom Design DevelopmentTest & Evalu- ation (DDT&E) programs, aeropass software production and maintenance, and mission operations. Test and production facilities and space station (or free flyer) support facilities have been considered. Cost and risk analysis results have led to several design modifica- tions (relative to the baselined MSFC con- cept), and technology development programs have been defined (for costing purposes) in major discipline areas.

VE= 8.6 kmlsec, 5g llmit

.25 - Nonstruct. Mass

0 20% Winged 0 40% A Therm. Prot. System - Shape .20 U D = 0.85 AFE Shape

a

C U D = 0.3

-

s.35 M

r 8

v) - 3.10 z h I I I I 0 100 290 300 400 Ballistic coeff., (m/C,A, kg/m2) Figure 12-4. Mass Fraction of Aerobrakes for Manned Mars Mission \

12-4 Aerobrake Design Studies for Manned Mars Missions

Objective. To systematically access the im- flightpath angle corridor width and pmvides pact of aerobrake 4D and ballistic coefficient an order of magnitude increase in cross-range on aerobrake mass fiaction and mission tra- capability duringdescent&om orbit. Also, the jectory and verify the advantages of using winged vehicle can be launched into Earth aerobrakes over using chemical propellants orbit M y assembled. The blunt shapes will for Mars entry.

probably have to be assembled in orbit, which is complex and potentially hazardous.

Approach. The vehicles’ mass fractions that must be devoted to the aerobrakes, including

. The aerobrakes’ mass frac-

the heat shields, have been computed for a tions range from less than to slightly over the high-speed manned M a r s entry of 8.6 km/sec 15% value that is considered to make and with an Earth 5 g deceleration limit.

aerobraking indisputably superior to chemi- Blunt, low liflt-drag ratio (ID), configura- cal propulsive braking. In fact, using the best tions with ballistic coefficients (m/C ) of current chemical propellants yielded a pro- 100 and 200 kg/m2 were studied. In a g i o n , pulsion system mass fradion that exceeded a delta-winged vehicle, with a LID of 0.85 and the aerobrakes’ values by factors of 4 to 5.

a ballistic coefficient of 375 kglm2, was stud- ied. The convective and equilibrium radiative Future Plans. The results of the study were heating and the pressures were computed at reported at an AIAA meeting. In addition, the selectedbody locations for the Martian atmo- analysis procedures that were developed w i l l spheric gases along the overshoot and under- be applied to minimize the aerobrake masses shoottrajectorieS.Bothinsulative,heat shields for the unmanned probe/landers being stud- and ablators were considered. Polyimidef ied for the Mars Environmental Survey graphite(PVGr)primary structural members (MESUR) missions.

were assumed for structural analyses. How- ever, the basic shell strudure consisted of an M. Taubr, M. Chargin, W. Henline, alnminumhoneycomb sandwich core between K. R. Hamm, Jr., H. Miura PI/Gr face sheets; the shell was stiffened by I- Ames Research Center beam frames. Multi-variable optimization was (415) 604-6086 used to minimize the structural weights.

A. Chiu and L. Yang Accomplishments. After addingheatshield- Sterling Software ing and optimizing the structure, the Palo Alto, CA aerobrakes’ mass fiactions (defined as heat shield plus aerobrake mass divided by the vehicle’s t o t a l mass at entry) varied from about 15% to 13% for ballistic coefficients of 100 and 200 kg/m2, respectively, for the blunt shapes and was slightly under 17% for the winged vehicle (see figure). The higher mass fraction for the winged-vehicle resulted from its much higher ballistic coefficient which created more intense heating and a higher mass fraction for thermal protection. How- ever, the increase in ID from 0 . 3 to 0.85 for the winged vehicle nearly doubles the entry T, = 1785OK, Reuseable Tile Slumping Starts and Thermocouple Temperature Limit Reached Radiative Heating . + '

I_E+I VSL - Without Slip

E

A NS - Without Slip

, + NS - With Slip

s

Y n Q) c,

a

K S .I w

a

s 100

Time, sec

Figure 12-5. Stagnation Point Heating Rate on AFE

12-5 Heating Rates forAeroassist Flight E

Objective. To define the heating rates and The heating rates, supplied flow-field characteristicsover the Aemassist by Langley, are based on the best techniques Flight Experiment (AFE) vehicle for the pur- available, and are the most complete data set pose of vehicle design, experiments’ design, of accurate calculations. The AFE project has and scientific pre-flight predictions.

relied heavily on these results and our assess- ment of the heating environment in the ve- Approach. The convective and radiative hicle design. Our results have shown that the heating rates and the flow-field characteris- heating environment for the current trajec- tics were calculated with several Merent tory yield heat shield surface temperatures computational codes at flight conditions from that are near the limit of the tile material.

currently considered trajectories. Any changes to the trajectory that result in increased heating rates could seriously jeop- Accomplishments. The stagnation-point ardize the survival of several experiments heating rates have been calculated at flight and acquisition of quality data.

conditions along the current trajectory. The heating rates over the complete vehicle, StatudPlans. The plan is to remain continu- forebody and afterbody, have been calculated ously involved in the calculation and assess- at two flight conditions - time of peak heating ment of the heating for the AFE. The impact and beginning of prime data period. The in uncertaintiesof several criticalparameters calculations are based on the best estimate of on the heating rates is being assessed.

physical input data and also consider uncer- tainties in critical values such as catalytic H. Harris Hamilton I1 wall coefficients and variations in radiation Aerothermodynamics Branch methods. The results have been supplied to Langley Research Center key organizations within the AFE project. (804) 864-4365 The results are being used in the heat shield design and in the design of the experiments.

The stagnation-point heating rates are used to define correction factors to the heating method used in the heat shield design, and the heating rates over the complete configu- ration are used as anchor points in the design process.

L 24 1 Form Approved REPORT DOCUMENTATION PAGE OM8 No. 0704-0188

I

u b k reoonmq aurden 'or i h n ~ollertion of ,-fi)rmdtiv 81 ?s:imatea i o arerage 1 hour oer resmrse rncluarno the lime for reviewing mstructtons. searching existing data Murces.

athem^ m a m a r ~ i a m n g in- data needed and Corroletini ana r e w e w q the L311eqton of infcrmafion Send comments re araing this burden esrlmate or any other arpecl of thlr ollemo: ct mtvmation mciuaing iuggesiioni tor reouctnb this o w m n i o warnnngton Heaaauariers lervices. Direclotate 30, mformation Owraiioni and Rewns. 12 IS Jetfenon iavv, tiqhway. Suile lZC4 Lrlmgior IL. 22202S302 and to in- Officer)? Management ana Budget Paperworx Reduction Projen(0704-0188). Warhmgton. DC 20503 1 2 . REPORT DATE 1 3 . REPORT TYPE AND DATES COVERED . AGENCY USE ONLY (Leave blank)

I A p r i l 1992 I Technical lemorandum

I. TITLE AND SUBTITLE 5. FUNDING NUMBERS NASA Aerodynamics Program Annual Report 1991 NASW 4430 Louis J. Williams, Kristin A. Hessenius, Victor R. Corsiglia, Gary Hicks, Pamela F. Richardson, George Unger, Benjamin Neumann, and Jim Moss . PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 8. PERFORMING ORGANIZATION REPORT NUMBER NASA Office of Aeronautics and Space Technology Aerodynamics Division (OAST/RF) 10. SPONSORING lMONlTORlNG t. SPONSORINGlMONlTORlNG AGENCY NAME(S) AND ADDRESS(ES) AGENCY REPORT NUMBER National Aeronautics and Space Administration Washington, DC 20546 NASA TM-4368 11. SUPPLEMENTARY NOTES This report is an annual accomplishments review for the Aerodynamics Division (Code R F ) . The information contained herein covers FY 91.

I2a. DISTRIBUTION I AVAILABILITY STATEMENT 12b. DISTRIBUTION CODE

Unclassified - Unlimited

Subject Category Vf I 13. ABSTRACT (Maximum 200 words) This report is the annual accomplishments review for the Aerodynamics Division (Code RF) during FY 1991. The program includes both fundamental and applied research directed at the full spectrum of aerospace vehicles, from rotorcraft to planetary entry probes. This report contains a comprehensive review of the following aerodynamics elements: computational methods and applications, CFD validation, transition and turbulence physics, numerical aerodynamic simulation, test techniques and instrumentation, configuration aerodynamics, aeroacoustics, aerothermodynamics, hypersonics, subsonics, fighter/attack aircraft and rotorcraft.

Unlimited ISN 7540-01-280-S500 Standard Form 298 (Rev 2-89) Prexribed b v ANSI Sid Z39-18 198 102