Euler Technology Assessment for Preliminary Aircraft Design ......preliminary design tools in the...

NASA / CR-1998-206947

Euler Technology Assessment for

Preliminary Aircraft Design-Unstructured/Structured Grid NASTD

Application for Aerodynamic Analysis of

an Advanced Fighter/Tailless

Configuration

Todd R. Michal

Boeing Company, St. Louis, Missouri

National Aeronautics and

Space Administration

Langley Research Center

Hampton, Virginia 23681-2199

Prepared for Langley Research Centerunder Contract NAS1-20342

March 1998

Available from the following:

NASA Center for AeroSpace Information (CASI)

800 Elkridge Landing Road

Linthicum Heights, MD 21090-2934

(301) 621-0390

National Technical Information Service (NTIS)

5285 Port Royal Road

Springfield, VA 22161-2171

(703) 487-4650

TABLE OF CONTENTS

SUMMARY ..................................................................................................... 2

INTRODUCTION ........................................................................................... 2

APPROACH .................................................................................................... 4

GRID GENERATION .............................................................................................................. 4

FLOW SOLUTION METHODOLOGY AND PERFORMANCE CHARACTERISTICS ......................... 6

RESULTS ......................................................................................................... 8

BASELINE CONFIGURATION ................................................................................................. 8

CAMBERED WING CONFIGURATION .................................................................................. 10

LATERAL-DIRECTIONAL CHARACTERISTICS ...................................................................... 10

CONTROL DEVICES ............................................................................................................ 10

CONCLUSIONS ........................................................................................... 11

ACKNOWLEDGEMENTS ......................................................................... 13

REFERENCES .............................................................................................. 13

FIGURES ....................................................................................................... 15

SUMMARY

This study supports a NASA project aimed at determining the viability of using

Euler technology for preliminary design use. The primary objective of this study was to

assess the accuracy and efficiency of the Boeing, St. Louis unstructured grid flow field

analysis system, consisting of the MACGS grid generation and NASTD flow solver codes.

Euler solutions about the Aero Configuration/Weapons Fighter Technology (ACWFT)

1204 aircraft configuration were generated. Several variations of the geometry were

investigated including a standard wing, cambered wing, deflected elevons, and deflected

body flap. A wide range of flow conditions, most of which were in the non-linear regimes

of the flight envelope, including variations in speed (high subsonic/transonic, supersonic),

angles of attack, and sideslip were investigated. Several flowfield non-linearities were

present in these solutions including shock waves, vortical flows and the resulting

interactions. The accuracy of this method was evaluated by comparing solutions with test

data and Navier-Stokes solutions. The ability to accurately predict lateral-directional

characteristics and control effectiveness was investigated by computing solutions with

sideslip, and with deflected control surfaces. Problem set up times and computational

resource requirements were documented and used to evaluate the efficiency of this

approach for use in the fast paced preliminary design environment.

The use of unstructured grids was found to significantly decrease the cycle time of

NASTD applications primarily through a reduction in grid generation time. The efficiency

and robustness of this method, while still too slow for generating an entire aerodynamic

database, are sufficient to provide data at a large number of points across the flight

envelope. The accuracy was generally sufficient for preliminary design use up to moderate

angles of attack (~15 degrees). The prediction of aerodynamic effects due to control

surface deflections were of mixed accuracy. Aerodynamic predictions of the elevon

control effectiveness and the lateral-directional characteristics due to asymmetric control

deflections were accurately predicted while the control effectiveness in pitch was

consistently over-predicted. Less accurate aerodynamic predictions were obtained for

control devices that generate a large amount of wake like separation such as the body flap.

Euler technology has strong relevance to preliminary-design applications. This

technology provides a means of predicting non-linear aerodynamic effects that previously

could only be obtained in the wind tunnel. This study has indicated that an un-exploited

potential exists for development of a lateral-directional design tool. However, further work

is needed to determine the parts of the flight envelope where this technology should or

should not be used. Additional work may also be required to develop empirical calibration

for some applications. The greatest benefit of this technology will be realized when it is

tied to advances in multi-disciplinary design tool development.

INTRODUCTION

Over the past decade, great strides have been made in the development of

preliminary design tools in the airplane radar signature and structural analysis disciplines.

In contrast, aerodynamic analysis in preliminary design continues to rely primarily on

linear tools developed several decades ago. The limitations of these methods are becoming

increasingly apparent with the advent of low observable and unmanned aircraft technology.

2

Thesetechnologieshaveled to non-traditionalvehicleshapesand control surfacedevicesthat exhibit highly non-linearaerodynamicbehavior. Aerodynamictools basedon linearaerodynamicmethodsare inadequatefor thesetypesof aircraft,particularly at theedgesofthe flight envelopesuch as high anglesof attack. For such applications,wind tunneltestingand/ornon-linearanalysisis required.

Until recently, Computational Fluid Dynamics (CFD) Euler or Navier-Stokesmethods have been used very little in the preliminary design environment. This is

primarily due to long cycle times and unvalidated accuracy levels. Computational

hardware improvements and CFD technological developments, such as the advent of

unstructured grids, have greatly reduced cycle times making Euler CFD methods a viable

candidate for preliminary design use. Incorporation of these methods into the preliminary

design process enables analytical determination of non-linear aerodynamic properties that

currently can only be assessed in the wind tunnel. These methods could therefore

potentially reduce the amount of costly wind tunnel testing. Another benefit of using CFD

in preliminary design is the potential for the development of a multi-disciplinary design

tool. Such a tool would allow the aerodynamic design to be tightly integrated with the

design of other disciplines.

Despite these potential advantages, CFD methods have not found their way into the

preliminary design environment. One reason for this may be the uncertainty associated

with using a new technology. There are several risks involved in using Euler CFD

methods for preliminary design. Errors are introduced into the analytical predictions by

neglecting the effects of viscosity. The significance of this error varies with the problem

geometry and flow conditions. The ability to account for this error is well documented for

many problems such as attached flows at low to moderate angles of attack, however, for

some cases the consequences of neglecting viscosity are difficult to predict. In addition,

the ability of Euler CFD methods to predict lateral-directional characteristics and the

effects from non-traditional control devices are not proven. It is also unclear whether the

improvements that have been made in CFD cycle time are sufficient to meet the needs of

the fast paced preliminary design environment.

A few years ago, NASA Langley Research Center initiated a project (Ref. 1-6) to

evaluate the viability of a series of Euler CFD methods for use in preliminary design. This

report summarizes the assessment of the Boeing, St. Louis developed CFD tools, MACGS

and NASTD, for use in preliminary design. These tools provide for rapid analysis of

complex configurations using either structured or unstructured grid techniques. This study

was focused on the assessment of the unstructured grid Euler capability of these tools.

NASTD/MACGS applications are performed within an integrated process whereby the

grids are generated directly on the CAD model. A common database file is carried

throughout the process from grid generation through post processing. To avoid the

requirement for a mainframe- or super-computer, which often is not available in the

preliminary design environment, solutions are computed in parallel on a network of

workstations. This provides rapid turnaround and low memory usage. These tools have

been used extensively on Boeing, St. Louis production programs such as the F/A-18 E/F,F-15C and AV-8B.

There were four primary objectives in this study. These were to assess the effects

of viscosity over a range of Mach number and angle of attack, assess aerodynamic

predictions for non-traditional control devices, evaluate lateral-directional analysiscapabilities,anddocumentconvergence-performancecharacteristics.

APPROACH

MACGS and NASTD were applied to the analysis of the AeroConfiguration/Weapons Fighter Technology (ACWFT) 1204 configuration shown in

Figure 1. This configuration was tested extensively (Ref. 7) in the NASA Langley

Research Center 8 ft. Transonic Pressure Tunnel. The ACWFT configuration is

representative of advanced preliminary designs. It is a tailless aircraft with a chined

forebody. Two wing geometries were tested with the ACWFT model. The baseline wing

has +/-30 degree leading and trailing edge sweeps, an aspect ratio of 2.65, and a taper ratio

0.132. The baseline wing cross section consists of a modified NACA 65A004 airfoil with

a sharp leading edge. The alternate wing model consisted of the same planform as the

baseline wing with the addition of camber and twist. The test model contained several

non-traditional control devices including elevons, and body flaps. A flow through duct

connecting the inlet and nozzle was incorporated into the test model. Test results included

force and moment measurements and pressure tap data over the wing surface. In addition,

pressure sensitive paint was used to obtain the global surface pressure distribution over theaircraft.

For this study, CFD solutions were computed on the ACWFT vehicle in six

different configurations as shown in Figure 2. These configurations included the baseline

configuration, baseline fuselage with a cambered/twisted wing, baseline configuration with

aflerbody flap deflected 90 degrees, baseline configuration with symmetric elevon

deflections of-20 degrees, baseline configuration with asymmetric elevon deflections of

+/- 20 degrees, and the baseline configuration with sideslip. The baseline, cambered wing,

and symmetric elevon configurations were modeled assuming symmetry about the fuselage

centerline. While developing the CFD model of the cambered wing configuration, we were

unable to locate the geometric definition of the cambered wing/fuselage interface that was

used on the wind tunnel test model. For this study an interface was made up by blending

the cambered wing geometry into the baseline wing root section. Unfortunately, the results

presented below show that this geometry modification may have influenced the resulting

CFD drag predictions.

A summary of the CFD run matrix is shown in Figure 3. For each configuration,

Euler solutions were computed at flow conditions of Mach 0.6, angles of attack of 1O, 15,

and 20 degrees, Mach 0.9 at angles of attack of 10 and 15 degrees, and Mach 1.2 at 10

degrees angle of attack. In addition to the Euler solutions, Navier-Stokes CFD solutions

were computed about the baseline and cambered/twisted wing configurations at the same

flow conditions to isolate the aerodynamic effects due to viscosity.

Grid Generation

Grid generation was performed using MACGS (Refs. 8,9), which is a general

purpose, arbitrary topology grid generation system developed at the McDonnell Douglas

Corporation. It supports the generation of multi-zone structured and/or unstructured grids.

MACGS is comprised of three modules: ZONI3G, GMAN, and GPRO. ZONI3G is used

to generate structured and/or unstructured surface grids. GMAN provides the capability to

generatevolume grids, specifyboundaryconditions and generatecoupling informationbetweenstructuredand/orunstructuredgrid zones. GPROmanipulateszones (suchastransforming, splitting, and combining) and supportsinputting and outputting files invariousformats. Theinteractive,graphicaluserinterfacesof ZONI3GandGMAN supportboth thenoviceandexpertuser.

A dual grid approachwasusedin this studywherealI viscouscomputationswereperformed on a pre-existing structuredgrid and all Euler solutionswere computedonunstructuredgrids. Unstructuredgrid generationwasperformedusingthefour stepprocessshownin Figure4. In the designenvironment,the geometryresidingin the CAD systemoften containsdetailedgeometrycomponents or surface gaps that the CFD user does notwant to include in the analysis. The first step in the grid generation process is to modify

the surface geometry to remove these unwanted geometry components and fill any

remaining gaps or holes. An unstructured surface grid is then interactively generated on

the clean surface representation. Grid resolution is controlled by the user through

specification of boundary edge distributions for each surface patch and through several line

and point source options. A tetrahedral volume grid is then generated within MACGS

using a Delauney point insertion approach. Grid swapping and smoothing are used to

ensure the quality of the final grid. Resolution of the volume grid is set by the surface grid

spacing and by two user specified parameters that control the global cell spacing and the

amount of clustering near the geometric surface. The resulting grid is partitioned into

multiple blocks with the METIS algorithm (Ref. 10). The sizes of each block are selected

to balance the solution load on parallel computational systems.

The structured and unstructured surface grids for the baseline ACWFT

configuration are shown in Figure 5. The multi-zone structured grid was generated in

MACGS during a previous study funded by Wright Laboratory (Ref. 7). The inlet and

nozzle ducts were treated differently in the unstructured and structured grids. To simplify

grid generation, the inlet and nozzle ducts were faired over in the structured grid. For all

but two of the unstructured grids, the inlet duct was modeled up to the compressor face

(where a mass flow boundary condition was specified) and the nozzle duct was faired over.

For the symmetric and asymmetric deflected elevon unstructured grids, a flow through duct

was modeled that connected the inlet and nozzle faces. This was done to capture the

effects of the nozzle flow washing over the deflected elevon surfaces.

The sizes of the resulting unstructured surface grids are shown in Figure 6. The

surface grids ranged in size from 150,000 to 300,000 triangles. Cuts through the structured

and unstructured volume grids are shown in Figure 7. In the first of the cuts shown in the

top of Figure 7, the lower surface of the structured grid differs from the unstructured

surface due to the inlet duct fairing. In Figure 8, the surfaces of the unstructured baseline

grid are shown after grid partitioning. In this example, the unstructured grid has been

partitioned into ten equal size zones.

The sizes of the structured and unstructured grids are summarized in Figure 9. The

sizes of the unstructured grids ranged from just over 1 million cells for the baseline

configuration to 2.7 million cells for the asymmetrically deflected elevon grid. The viscous

structured grids contained about 2.7 million nodes. Labor hours for the unstructured grid

generation are shown in Figure 10. The baseline unstructured grid required 30 person

hours to generate. Grid generation for the other five configurations examined in this study

were made by making minor modifications to the baseline surface grid. Thesemodificationsrequiredonly afew personhoursfor eachconfiguration. Thecomputationaltime requiredto generateeachunstructuredgrid is shownin Figure 11. The CPU timesgivenare for a Silicon GraphicsR10,000processorandrangefrom 1.5hoursto almost3hours.

Flow Solution Methodology and Performance Characteristics

The solution computations in this study were performed using the NASTD flow

solver. This flow solver was developed by the McDonnell Douglas Corporation, and can

run on structured, unstructured or a combination of structured and unstructured grids. It

supports multi-block and overlapping (chimera) grids. It runs in serial or parallel on a wide

variety of machines. A complete description of the NASTD structured grid solution

algorithm is given in Reference 11. The structured grid algorithm solves any subset of the

full Reynolds averaged Navier-Stokes equations. Options include Euler, thin layer,

parabolized Navier-Stokes and full Navier-Stokes calculations. Turbulence can be

modeled by a variety of algebraic, one- and two-equation turbulence models. The solution

algorithm can be selected zonally by the user. The default time integration scheme is a

first-order, approximately factored implicit scheme. For inviscid flows (or, under the thin

layer approximation, for directions without viscous terms) the implicit operator is

diagonalized, providing a significant speed-up. Explicit Runge Kutta options of up to

third-order are also available for time accurate flowfields. For steady-state flows, variable

time steps based on local eigenvalues are used to speed convergence. Grid sequencing is

available to speed convergence on large grids. The default explicit spatial operator is a

second-order flux difference splitting scheme, also known as Roe's scheme. The standard

upwind operator has been replaced by a mixed scheme which retains the upwind scheme

stability properties with reduced numerical dissipation. Optionally, the scheme may be

switched to various first- through fifth-order schemes and total variation diminishing

(TVD) limiters may be activated. Other available schemes include standard second-order

central differencing with added second- and fourth-order dissipation.

A complete description of the NASTD unstructured grid algorithm and the parallel

implementation is given in References 12 and 13. This algorithm is a node-based upwind

finite-volume unstructured grid algorithm. The implementation used for this study solves

the Euler equations on tetrahedral cell grids. Higher-order computations are achieved

using a least squares reconstruction scheme with flux limiting. The numerical flux values

are computed at the mid-point of each edge using Roe's approximate Riemann solver.

Flowfield variables are stored at grid nodes and flux computations are performed at each

grid edge. This results in relatively low storage requirements and run times. Time

integration is performed using an explicit point Jacobi or Runge-Kutta algorithm for each

node.

The following NASTD options were used in the computations for this study. The

fluxes were computed using a second-order accurate Roe's scheme with a Total Variational

Dimensioning (TVD) limiter in the case of the structured solutions and a monotone limiter

for the unstructured cases. In the structured grid Navier-Stokes computations, turbulence

was modeled with the Spalart-Almaras turbulence model. An implicit approximate

factorization algorithm was used for the time integration of the structured grid cases and an

explicit Runge-Kutta scheme was used for the unstructured grid cases. Solu.tionconvergencewasdeterminedby monitoringtheintegratedlift, dragandpitching moments.For the viscouscasesthe friction dragwasalso monitored.No attemptwasmadeto findthe maximum CFL numberfor thesecases. Insteada "safe" CFL numberwas selectedbasedon past experience. CFL numbersranged from 0.3 to 0.7 for the unstructuredcomputationsand 1.0to 3.0for thestructuredgrid computations.

Two representativeexamplesof Euler solution convergencefrom this study areshownin Figures 12 and 13. Theseexamplesrepresentthe best (Figure 12), and worst(Figure 13)Euler solutionconvergencehistoriesfrom this study. In thesefigures, the liftcoefficient is plotted versusthe solutioncyclenumber. In Figure 12the convergenceforthebaselineconfigurationat Mach1.2, I0 degreesangleof attackand5 degreessideslipisshown. This solution was restarted from the 0 degree sideslip solution at 700 cycles. Thelift coefficient reaches a steady value after an additional 1800 cycles (for a total of 2500

cycles). In Figure 13 the lift coefficient versus cycle number is shown for the baseline

configuration with -20 degree symmetric elevon deflections at Mach 0.6 and 20 degrees

angle of attack. This solution was run with the first-order accurate scheme for 1800 cycles

and then switched to the second-order scheme. The lift coefficient did not converge to a

steady value but instead oscillated about an average. This behavior was observed in all of

the 20 degree angle of attack cases and is most likely due to a non-steady behavior in theflow field.

The convergence properties of the NASTD Navier-Stokes solutions about the

baseline configuration at Mach 0.6, 10 and 20 degrees angle of attack are shown in Figures

14 and 15. These solutions were run for 680 cycles on a sequenced grid (every other grid

point removed in all three directions). The solution was then switched to the full grid and

reached convergence after another 500 cycles. The Navier-Stokes solutions did not

experience the oscillatory behavior at the high angles of attack seen in the Euler solutions.This could be due to the viscous terms which add additional diffusion to the flow.

Run times for the Euler solutions are summarized in Figure 16. The minimum,

average, and maximum run time are shown for each configuration. The numbers shown

represent the total CPU time and were obtained by multiplying the CPU time/iteration/cell

by the number of cells and the number of iterations. The times do not include the savings

obtained by running on a parallel computational system. For instance the average baseline

configuration solution required 90 hours of CPU time. When this solution was run on a

cluster of ten workstations, the actual clock time was a little under 10 hours. The memory

requirements for each of the Euler solutions are summarized in Figure 17. These memory

requirements are presented as though the solution were run as a single zone on one

processor. The actual memory requirements per machine varied depending on the number

of grid partitions. For ten equal size partitions, the memory requirements are 1/10 the total.

Solution run times and memory requirements for the Navier-Stokes solutions are

summarized in Figures 18a and b. Once again these numbers are given for a single zone

solution on a single processor. The actual requirements were much lower when running in

parallel.

RESULTS

Baseline Configuration

Euler and Navier-Stokes solutions were generated on the baseline ACWFT

configuration. Comparisons of the Euler and Navier-Stokes solutions were made to

identify the error introduced in the Euler solutions by neglecting viscosity. Further

comparisons with test data were made to identify the accuracy of the CFD methods. In

addition, differences between the baseline and other configuration results were used to

measure the incremental effects of each configuration.

Contours of the predicted surface pressure coefficient and traces of the streamlines

for the Euler and Navier-Stokes results at Mach 0.6, and angles of attack of 10, 15 and 20

degrees are shown in Figure 19. In this Figure, the Euler solution is shown on the left half

of the aircraft and the corresponding Navier-Stokes solution is shown on the right half of

the aircraft. At 10 degrees angle of attack, the Euler and Navier-Stokes results are very

similar. Both solutions predict a vortex separating off of the chined forebody and another

off of the wing leading edge. At 15 degrees angle of attack the surface pressures are once

again very similar. Both solutions indicate vortices similar to the 10 degree angle of attack

case. At 20 degrees angle of attack the differences between the Euler and Navier-Stokes

solutions are more noticeable both in terms of surface pressure distribution and particle

traces. In the Euler solution the wing leading edge vortex appears to have burst. This

results in significant differences in the surface pressures between the Euler and Navier-

Stokes solutions on the upper surface of the wing.

Total pressure contours at fuselage stations of 260 in., 420 in. and 510 in. are

shown in Figures 20 and 21. The locations and strengths of the predicted vortices for the

Euler and Navier-Stokes solutions are very similar at I0 degrees angle of attack. At 20

degrees angle of attack, however, the Navier-Stokes solution indicates a larger total

pressure loss in the vortex cores and the structure of the vortices are considerably different.

The effect of viscosity on the predicted solutions at different Mach numbers is

presented in Figure 22. In this figure, surface pressure coefficient contours and streamline

traces from the Euler and Navier-Stokes solutions at 10 degrees angle of attack and Mach

numbers of 0.6, 0.9, and ! .2 are compared. The streamline patterns predicted by the Euler

and Navier-Stokes solutions are similar at all three Mach numbers. However, there are

several differences in the predicted surface pressures at Mach 0.9. As expected, the Euler

solution indicates a strong shock over the wing at about 75% chord, while the Navier-

Stokes solution predicts a more diffused footprint of the shockwave that is slightly further

fornvard. This is probably due to shock boundary-layer interaction effects that the Euler

solution is missing. At Mach 1.2, the shock has moved aft of the wing trailing edge and

the Euler and Navier-Stokes solutions agree fairly well.Pressure coefficient contours on the lower surface for the Euler and Navier-Stokes

solutions are compared in Figure 23. The results agree well except near the inlet where the

two solutions have a different treatment of the inlet geometry (faired over for Navier-

Stokes and flow through for the Euler). While having little influence on the upper surface

solution, the different inlet models significantly change the lower surface results.

A comparison of surface pressures from the CFD results and pressure sensitive

paint (PSP) test data is shown in Figures 24 and 25 for flow conditions of Mach 0.6, and

8

Math 0.9 at 15 degreesangleof attack. The PSPdatawas not calibratedto provide aquantitativevaluefor eachcolor. Instead,the color mapusedfor plotting theCFD resultswasselectedto attemptto matchthe colors of the PSPdatathus providing a qualitativecomparisonof the flowfield structuressuchas shockwaves. At Mach 0.6, the Euler,Navier-Stokes,and PSPcontoursarevery similar. At Mach 0.9 the test data comparesfavorably with the Navier-Stokes results while, as expected, the Euler results clearly miss

the shock boundary layer interaction on the wing.

In addition to the PSP pressure data, surface pressure taps were placed at four

spanwise locations on the test model. Comparisons of the CFD surface pressures with

measurements taken at the pressure taps are made in Figures 26 through 31. In Figures 26

through 28, results from the solutions at Mach 0.6, angles of attack of 10, 15 and 20

degrees are shown. At angles of attack of 10 and 15 degrees, the Euler and Navier-Stokes

results are similar with the exception of the suction peak at the leading edge. As expected,

the Euler solution over predicts the acceleration around the sharp wing leading edge. At 20

degrees angle of attack there are significant differences in the Euler and Navier-Stokes

surface pressures. This is consistent with the differences that were observed in the surface

pressure contour plots above. Comparisons of the CFD results with the pressure tap data at

Mach 0.9 are shown in Figures 29 and 30. As expected, the Euler solutions predict a shock

location that is slightly aft of that predicted by the Navier-Stokes solutions. In addition,

there is a considerable amount of smearing of the shock footprint evident in the Navier-

Stokes results and test data that, is not present in the Euler solution. In Figure 31 the CFD

surface pressures are compared with pressure tap data at Mach 1.2 and 10 degrees angle of

attack. At this flow condition, the shock has left the wing surface and the Euler and

Navier-Stokes solutions compare favorably with the test data.

Force predictions were obtained from the CFD solutions by integrating the surface

pressures (and skin friction for the Navier-Stokes solutions) over the aircraft surface.

Corrections were added to the Euler drag estimates to account for the skin friction drag.

The corrections were obtained using the following procedure. First, Euler solutions were

computed over the baseline configuration at angles of attack that resulted in zero lift, and

Mach numbers of 0.6, 0.9, and 1.2. Next, the zero lift drag predicted by each Euler

solution was subtracted from the zero lift drag measured in the test at the same Mach

number to obtain the skin friction contribution to the total drag. The resulting skin friction

estimates were then added to all the Euler estimates. For the Euler solutions that failed to

converge to a steady state, force and moment values were obtained by averaging the

integrated results over the last few hundred cycles of the solution. Error bars are drawn to

indicate the maximum and minimum oscillation about the average. The CFD force and

moment predictions are compared with test data in Figures 32 through 34. The lift and

drag from the Euler solutions match the test data very well with the exception of the Mach

0.9, 15 degrees angle of attack. This is probably due to the missing shock boundary layer

interaction effects in the Euler solution. Surprisingly, the Navier-Stokes force and moment

results are slightly worse than the Euler predictions. The most likely reason for this

discrepancy is the presence of the faired over inlet model used in the viscous computations.

Cambered Wing Configuration

An alternate wing was tested on the ACWFT geometry. This wing was similar to

the ACWFT baseline wing with the addition of camber and twist. Contours of the

predicted surface pressure coefficient and streamline traces are compared for the Euler and

Navier-Stokes cambered wing configuration solutions at Math 0.9, 10 degrees angle of

attack in Figure 35. The comparison is similar to the baseline wing comparisons showing

that the Euler solution is missing the shock boundary layer interaction effects. The force

and moment predictions are compared with the test data in Figures 36-38. Again these

comparisons are very similar to the baseline wing results. The most notable deviations

from the test data occur at flow conditions of Mach 0.6, 20 degrees angle of attack and

Math 0.9, 15 degrees angle of attack. Increments in the force and moment predictions

between the cambered wing and baseline wing configurations are shown in Figures 39-41.

The test data shows a slight increase in lift and decrease in drag with little change in

pitching moment. The Euler and Navier-Stokes results also predict a slight increase in lift,

however, both methods predict an increase in drag. This discrepancy from the test data

may be partially attributed to an increase in interference drag caused by the method

employed to attach the alternate wing in the CFD grid.

Lateral-Directional Characteristics

The ability to compute lateral-directional characteristics is essential for a

preliminary design tool. The baseline configuration was run at 5 degrees sideslip to

evaluate the lateral-directional characteristic prediction capability of the present Euler

method. Streamline traces and surface pressure predictions from the Euler solutions at

Mach 0.6, 15 degrees angle of attack with 0 and 5 degrees sideslip are compared in Figure

42. The sideslip has little effect on the surface pressures. The most notable effect of the

sideslip is the change in track of the vortices downstream of the aircraft with little effect

over the aircraft itself. Force and moment predictions are compared with test data in

Figures 43-45. Increments in the force and moment predictions between the sideslip and

zero sideslip cases are shown in Figures 46-48. The test data indicates large increments in

the lift, drag and moments while the CFD results indicate small changes in the forces and

moments. The test data trends are contrary to the expected behavior of a tailless aircraft.

We suspect that the location of the support strut on the model may have influenced the testdata.

Control Devices

One of the objectives of this program was to evaluate the ability of the CFD method

to compute the effects of control surfaces. Solutions were computed about three control

devices including a symmetrically deflected elevon, asymmetrically deflected elevon, and a

body flap.

The ACWFT solid surface model and a closeup of the surface grid about the

syrmnetrically deflected elevon is shown in Figure 49. The elevon was deflected up 20

degrees on the left and fight sides of the aircraft. Euler solutions were run at all six flow

conditions. A comparison of the surface pressure and streamline traces for the baseline

(undeflected elevon) and symmetrically deflected elevon cases at Mach 0.6, 15 degrees

angle of attack is shown in Figure 50. The elevon deflection primarily affects the solution

10

nearthetail andhaslittle effecton the solution over the wing. The predicted lift, drag andpitching moment from the CFD solutions is compared with test data in Figures 51-53.

Once again the CFD lift and drag predictions agree with the test data, however, the CFD

results over predict the effect of the elevon deflection on pitch up. This can be seen in the

incremental force and moment plots shown in Figures 54-56. The CFD and test data both

indicate a substantial reduction in lift with a slight decrease in drag. The CFD results

overpredict the effect of the elevon deflection on pitching moment.

The ACWFT solid surface model and a closeup of the surface grid about the

asymmetrically deflected elevon and deflected body flap are shown in Figure 57. Surface

pressure and streamline traces from the CFD solutions at Mach 0.6, 15 degrees angle of

attack, with the devices are compared with the baseline solution in Figure 58. The

asymmetrically deflected elevon has little effect on the solution other than in the tail

region, whereas the body flap has a larger influence on the surface pressure of the

surrounding geometry.

Force and moment predictions for the asymmetrically deflected elevon CFD

solutions are compared with test data in Figures 59-61. The comparisons are similar to the

previous results showing good agreement for lift and drag except at Mach 0.9, 15 degrees

angle of attack. Increments in the force and moment predictions with the baseline results

are shown in Figures 62-64. As expected the pitching moment increment is very small at

all three Mach numbers. At Mach 0.6, the CFD results indicate an increase in lift and drag

as the angle of attack is increased whereas the test data a constant increment in lift and a

decreasing increment in drag. These discrepancies may be due to the lack of convergence

of the Euler method at the high angles of attack. The incremental pitch, yaw and roll are

well predicted by the CFD method.

Force and moment predictions for the deflected body flap CFD solutions are

compared with test data in Figures 65-67. Once again the force comparisons are good with

the exception of the transonic cases at Mach 0.9. Increments in the force and moment

predictions with the baseline results are shown in Figures 68-70. The incremental data

indicates a slight decrease in lift and increase in drag (for angles of attack less than 15

degrees). The CFD results underpredict the lift decrease due to the flap particularly at the

higher angles of attack. The CFD drag increments are also much higher than the test data

increments at the higher angles of attack. The discrepancies may be due to lack of

convergence or poor modeling of the large wake like separated region aft of the flap. This

type of flowfield is largely dominated by viscous effects and is not well modeled with an

Euler method. The incremental pitch, roll and yawing moments due to the flap are well

predicted by the CFD results.

CONCLUSIONS

This study has provided an assessment of the viability of using the NASTD

unstructured grid Euler technology in preliminary design. Euler solutions about the

ACWFT 1204 configuration with several geometry variations including baseline wing,

cambered wing, deflected elevons, and deflected body flap were generated. A wide range

of flow conditions, most of which were in the non-linear regimes of the flight envelope,

were evaluated including transonic and high angle of attack flowfields. Several non-

linearities were present in these solutions including shock waves, vortical and separated

11

flows. Comparisonswith testdataandNavier-Stokessolutionswereusedto evaluatetheaccuracyof this Euler methodand to identify viscouseffects for selectedconfigurationsandconditions. Solutionswith sideslipanddeflectedcontrolsurfaceswerecomparedwithtest datato evaluatethe ability to accuratelypredict lateral-directionalcharacteristicsandcontrol effectiveness.

The unstructuredgrid approach facilitated rapid modeling of the ACWFTconfiguration and its variations. Geometryvariations such as flap deflections weremodeledin only a few hours. The methodologyproved to be very robust generatingsolutions for various surfacecontrol devicesand variousflow conditionswith very fewproblems. Run timesweresufficiently fast for usein thepreliminarydesignenvironmentwith overnightrun timespossibleonparallelcomputationalsystems.Thesecycletimesaresufficient to generatedataat a largenumberof pointsacrossthe flight envelope,however,they are still too slow to generatean entire aerodynamicdatabasetypically developedduring thewind tunneltest.

The Eulerresultscapturedseveralnon-linearaerodynamiccharacteristicsof thetestdataat high anglesof attack. Forceandmomentpredictionsweregenerallysufficient forpreliminarydesignuseup to moderateanglesof attack(~ 15degrees)acrossthe examinedMach numberrange. Lessaccuracywasobtainedat high anglesof attackandfor controldevicesthatgeneratedalargeamountof separationsuchasthebody flap. Onesurprisingresult of this studywasthatNavier-Stokesforceandmomentresultswerenot appreciablybetterthan theEulerpredictions. This indicatesthereis little benefit in steppingup to thelonger run times and complexities of a Navier-Stokesmethod for these types ofpredictions.

OnecasewhereNavier-Stokespredictionsmay haveprovidedbetter resultsthantheEuler solutionsis for thedeflectedafterbodyflap. Thiscontroldevicegeneratesa largeseparatedregionaft of theflap that interactswith thefuselagesurfaceto generatechangesin theforce andmomentdistributions. TheEuler methoddid a goodjob of predictingtheflap control effectson theroll, yaw,andpitchingmomentsbut significantlyoverpredictedtheeffect on lift anddrag. Betterpredictionswereobtainedfor theelevoncontrol deviceeffectiveness.The lateral-directionalcharacteristicsdueto asymmetriccontroldeflectionswere accuratelypredictedwhile the control effectivenessin pitch wasconsistentlyover-predicted. With further work, empirical correlations to compensatefor the pitcheffectivenesscouldbedeveloped.

Euler technologyhas strong relevanceto preliminary-designapplications. Thistechnologyprovidesa meansof generatingnon-linearand lateral-directionalaerodynamicdatathat previouslycouldonly beobtainedin thewind tunnel. The ability to analyticallygeneratelateral-directionaldataprovidesanun-exploitedpotentialfor the developmentofa lateral-directionaldesigntool basedonexistingEulcr technology.

It is likely that linearaerodynamictoolswill continueto beusedto developa largeportion of the aerodynamicdatabase.Furtherstudy is necessaryto determinethepartsofthe flight envelopewhere linearand non-linearmethodsare bestapplied. Comparisonsbetweenlinearandnon-linearresultsalthoughnot apartof this study,couldhelpguidethisdetermination. Furtherstudy is also necessaryto developempiricalcalibrationsof Eulerresults for someapplications. This wasevident m the elevoneffectivenesspredictionsobtainedin this study.

12

Oneof the greatest potential benefits of Euler technology may be the ability to tienon-linear aerodynamic data into a multi-disciplinary design tool. The ability to share data

with other disciplines and use a common geometry database is essential if this technology

is to be used successfully in the preliminary design environment.

ACKNOWLEDGEMENTS

This effort was sponsored by NASA-Langley Research Center under Contract

NAS1-20342. Farhad Ghaffari was the contract Technical Monitor and provided

invaluable technical guidance over the course of this study. His assistance is gratefully

acknowledged.

REFERENCES

1.) Finely, D.B.,"Euler Technology Assessment Program for Preliminary Aircraft Design

Employing SPLITFLOW Code With Cartesian Unstructured Grid Method," NASA

CR-4649, March 1995.

2.) Kinard, T.A., Hams, B.W., and Raj, P.," An Assessment of Viscous Effects in

Computational Simulation of Benign and Burst Vortex Flows on Generic Fighter

Wind-Tunnel Models Using TEAM Code," NASA CR-4650, March 1995.

3.) Treiber, D.A. and Muilenberg, D.A.," Euler Technology Assessment for Preliminary

Aircraft Design Employing OVERFLOW Code With Multiblock Structured-Grid

Method," NASA CR-4651, March 1995.

4.) Finely, D.B. and Karman, Jr., S.L.,"Euler Technology Assessment for Preliminary

Aircraft Design - Compressibility Predictions by Employing the Cartesian Unstructured

Grid SPLITFLOW Code," NASA CR-4710, March 1996.

5.) Kinard, T.A. and Raj, P.," Euler Technology Assessment for Preliminary Aircraft

Design - Compressibility Predictions by Employing the Unstructured Grid USM3D

Code," NASA CR-4711, March 1996.

6.) Kinard, T.A., Finley, D.B., and Karman, Jr., S.L.,"Prediction of Compressibility Effects

Using Unstructured Euler Analysis on Vortex Dominated Flow Fields," AIAA Paper-

No. 96-2499, June 17-29,1996.

7.) O'Neil, P.J., Krekeler, G.C., Billman, G.M.. and Creaseman, F.C., "Aero

Configuration/Weapons Fighter Technology (ACX_,'FT) - Summary Technical Report,"

WL-TR-95-3002, December 1994.

8.) Gatzke, T.D., et. al., "MACGS: A Zonal Grid Generation System for Complex Aero-

Propulsion Configurations," AIAA-91-2156, June 199 I.

13

9.) LaBozzetta,W.F., Gatzke, T.D., Ellison, S., Finfrock, G.P., and Fisher, M.S.,"MACGS - Towards the Complete Grid Generation System," AIAA 94-1923, 12thAIAA Applied Aerodynamics Conference, June 20-22, 1994.

10.) Karypis, G., and Kumar, V.,"METIS, Unstructured Graph Partitioning and Sparse

Matrix Ordering System", Users Manual August, 1995.

11 .) Romer, W.W., and Bush, R.H., "Boundary Condition Procedures for CFD Analysis of

Propulsion Systems - The Multi-Zone Problem," AIAA-93-1971, June 1993.

12.) Michal, T., and Halt, D., "Development and Application of an Unstructured Grid Flow

Solver for Complex Fighter Aircraft Configurations," AIAA 95-1785, June, 1995.

13.) Michal, T., and Johnson, J., "A Hybrid Structured/Unstructured Grid Multi-Block

Flow Solver for Distributed Parallel Processing," AIAA 97-1895, June, 1997.

14

Figure1.

r

ACWFT 1204 Three View Drawing.

CASE 1._ Baseline Configuration CASE 3.0: Symmetric Elevons

CASE 1.b: Cambered Wing CASE 3.b: Asymmetric Elevon

CASE 2: AfteYoody Flop CASE4: Basellnewlth Sideslip

Figure 2. ACWFT 1204 Configurations Modeled in Euler Technology Assessment Study.

15

c_

(z

10°

150

200

100

150

200

Case la: BaselineMach 0.6 Mach 0.9 !Mach 1.2

I,V I,V I,V

I,V I,V

I,V

Case 2: After Body FlapMaeh 0.6 Maeh0.9 Mach 1.2

I I I

I I

I

Case 1b: Cambered/Twisted WingMach 0.6 Mach0.9 Mach 1.2

I,V I,V I,V

I,V I,V

I,V

10°

£_ 150200

c_

Case 3a: Symmetric Elevons

100

150

200

Mach 0.6 Mach 0.9 Mach 1.2

I I I

I I

I

Case 3b: Asymmetric Elevons

10°

15o

20°

Mach0.6 Mach0.9 Mach 1.2

I I I

I 1

I

I: Inviscid V: Viscous

Case 4: Baseline with Sideslip

o_

10°

150

200

Mach0.6 Mach0.9 Mach 1.2

I I I

I I

I

Figure 3. Euler Technology Assessment CFD Run Matrix.

Surface Grid Generation

Volume Grid Generation

Figure 4. MACGS Unstructured Grid Generation Process.

16

35O

_' 300

250

200

¢_ 150

"6_oo

_ 5o

0

Figure 6. Unstructured Surface Grid Sizes, Six ACWFT 1204 Configurations.

18

IF l 111111i_lfl/I III UlliliIfJIIlklI III lllitlllllillllll Illlr)[_i IIilllllUllmlllllll

_llIIIYl,!lll r I f\_ I It Ill 11IIIIINnlllllll II_[llJfJt_/ll.l(l I\l I I IllUIIItmmHIIIII II

iiiiiiiiiiiiiii,_1 lIII] llll II '_l_llit_Nllttlllll IIIILllllllllnlllllII__/HIIIllllllllllllllll II_ii_jiiiiVlllllllh'ltmillllllldiilllllllll II ""_I]ITTlll f_ILII ] flfdlllllllllllll]llllll111II II

_ I _iI IIIIIIl[[llll_llllllIII I illlll

_q_l_'_llJ]]JiBI]llllllllllli[llll [JillIII I........ I.III

]I I

l_,i i

i

Figure 7. Cuts Through Volume Grids, ACWFT 1204 Baseline Configuration.

19

\ \

Figure 8. ACWFT 1204 Unstructured Surface Grid After Partitioning.

2O

3000 I

2500

2000

1500

1000

50O

• Cells •Nodes I

30

Figure 9. Volume Grid Sizes.

25

20

"J 10

5

Figure 10. Unstructured Grid Labor Hours.

21

3

2.5O

® 2

¢_ 1.5

m o.5

Figure 11. Unstructured Grid Computation Time.

"E

o

o68

0 67

0_16

Oe6

O64

063

5OO

Figure 12.

I ! | i n * * | i

I tOO 1500 2000 2500 3000 _7 4_3: 45[]0 SOOt] SSO0

Cycte

Lift Convergence, Unstructured Grid Euler Solution, Mach 1.2,

10 degrees angle of attack, 5 degrees sideslip.

22

3D

25

20

ou,

15

1

Figure 13.

148

i i i

1000 2000 3000 4000 5000 8000 7000

Cycle

Lift Convergence, Unstructured Grid Euler Solution,

Mach 0.6, 20 degrees angle of attack.

148

138 i i i i i i i840 720 800 080 gso _040 1120 1200

Cycle

Figure 14. Lift Convergence, Navier-Stokes Solution,


3_

3oo

®5

2_

292 I i i I i i i ,_0 720 000 O_ gSD _._ 1120 1200 12_C 1&60

Cycle

Figure 15. Lift Convergence, Navier-Stokes Solution,


23

1000

900

800

700

600

500

400'I,.,.

I_ 300

200

100

0

Figure 16.

5O

45

4O0" 35x

300E• 25

o 20

_ 150

_ 10

5

Figure 17.

um• average •maximum

NASTD Euler Solution Computation Times.

NASTD Euler Solution Memory Requirements.

24

I minimum• maximum200

1000ILl

800

600

400

II:--- 200t_¢n

0

• average140

120

100

:E 80

0E 6O

411

2O

0

a.) Single Processor Run Time b) Single Processor Memory

Requirements

Figure 18. NASTD Navier Stokes Solution Time and Memory Requirements.

25

M = 0.6, _ = 100 M = 0.6, _ = 15 ° M = 0.6, c_= 20 °

Cp

I .5• 0.0

-0.5-1.0

15

Navier-Stokes Euler

Navier-Stokes Euler

Navier-Stokes

Figure 19. Viscous Effects Versus Angle of Attack, ACWFT Baseline Configuration,

Mach 0.6, NASTD Euler and Navier-Stokes Solutions.

Euler Navler-Stokes Pl/Pto

I 1.0

0.9

0.8 260 .............

0.7

420 ...........................

Euler, Navier-Stokes

Comparison

Figure 20. Euler and Navier-Stokes Comparison, Total Pressure Contours, ACWFT

Baseline Configuration, Mach 0.6, 10 Degrees Angle of Attack.

26

Euler Navier-Stokes Pt/Pto

I 1.0

0.9

0.8 260 ............

0.7

420 .............................

Euler, Navier-Stokes

COMPARISON

Figure 21. Euler and Navier-Stokes Comparison, Total Pressure Contours, ACWFT

Baseline Configuration, Mach 0.6, 20 Degrees Angle of Attack.

Cp

0,0

-0.5_'_-1 0

M = 0.6, _ = 10 0

Euler_| Navier-

okes

M = 0.9, _ = 10 0

Euler

M = 1.2, a = 10 0

EulerNavier-Stokes

Figure 22. Viscous Effects Versus Mach Number, ACWFT Baseline Configuration, 10

Degrees Angle of Attack, NASTD Euler and Navier-Stokes Solutions.

27

Lower Surface

M "0.9, a = 10° M= 1.2, ='= 10°

Invlscld Viscous Invlscld Vbcous

Cp

0.30

-0.25

-0.80

Figure 23. Pressure Coefficient Contours on Lower Surface of ACWFT Baseline

Configuration, Euler and Navier-Stokes Solutions.

Experimental - Computational Cp Comparison

Euler Solution

(M 0.6, c_15 °)

Pressure

(M 0.6, a 16 °)

Nsvler-Stokes

(M 0.6, c_15 °)

Figure 24. CFD Surface Pressure Comparison With Pressure Sensitive Paint Test Data,

ACWFT Baseline Configuration, Mach 0.6, 15 Degrees Angle of Attack.

28

Experimental - Computational Cp Comparison

Euler Solution Pressure Sensitiv( Navier-Stokes Sol

(M 0.9, _ 15°) (M 0.9, a 16°) (M 0.9, _ 15°)

Figure 25. CFD Surface Pressure Comparison With Pressure Sensitive Paint Test Data,

ACWFT Baseline Configuration, Mach 0.9, 15 Degrees Angle of Attack.

i0

1o_5

g: 0

_'_? _ TestData

'_',,_, - _ NAS'T'DEulerSolution

--- NASTD Newer-Stokes Solution

_0% span

, , , , ,

¢_ cut at 70% spa_ o

_e t_ 20 21 zz

x/c

jo

ioiT

t 55% span

| o¢

2:b e_so

te tg _o Zl 22 z_

_c

i

cut at 88% span

Ie4 _ee !e2 _le l_o _ _$ _1_ _ e_c

Figure 26. CFD Surface Pressure Comparison With Test Data, ACWFT Baseline

Configuration, Mach 0.6, 10 Degrees Angle of Attack.

29

qt,,1-

!lb

-1 o

J . . .

i

Testo=a. -- NASTD 6.Jler Solutbon_, ....... N/L_TD N_vier.Slokes Solution

,,, !

,_e e- cutat40% _oan

m

|

d .

I

>o •

cutat88% s_n



_a

er

",. _ list U_ta: " I"¢_TD Euler Solutioni ",., ...... NASTD Nav_r-Stokes Solution

! \,,

p

c_

(_ °,7 /

__""-cut"" _t 88% span



30

'7't,

(_ lest Data_._..: N/kS'TO Euler Solution

...... NAS'T'D Navier-Stokes Solubon

•. i; t_ w m: _ ..... , ,,I

s_

• cutmtTO%spsn

• = 'D _: =t =

_ _ ,

!

'_ t_o c_ 55% _n

lo [

II.I; 11; l_.s la,4 1_+: z;'t, _t;= z'Ll ZZ4 =_Z Z_;

|=

cut _t 88% span



it •:._.,',+

i

.+f:

F

._ i+1:

(_ T_ E)alaf'_ I_ASTC Euler Solution

_-_-__..... ....... I_A_ IIS NaVl_r-_StoltesSolut=on

° ":,"'m'-- m _

"% _I

11 tI0,+ ,_+ ,i_ ,! ,I :) 21 LI l,; II =; _/,. _; III I II 14 •+i_ /,;_ L; ,I_.| ;11.| gg" +:+.Z Z_L+'

O 1_ 10

(]I

(-


Configuration, Mach 0.9, 15 Degree Angle of Attack•

31

¢._

Zl

_ " cut at 40% span

¢_ T,,_._ _ iNASTD Euler Solution

...... NASTD Na,ttet-Stokes Solution|

_4

)c

J_

)4

, t .....

"*--Z" "--_ "°_'_" "--''_" _ _ ",

! cut at 56% span

_--o_ T .... _..Q__.._..._.._=_._

cut at 70% span

• i

r _ _ a

-_ c4

_"- t" cut at 88% spmn

i



1.2 1.2

1.0

0.8

"6

I

o,4

0.2 !

0.0 ,

0 5 10 15 20

Angle of Attack (deg)

1.2

1.0

o.B

_ 0.6

_ 0.4]

00

0.0 0.1 0.2 0.3 0.4 0.5

Drag Coefficient

1.0

0.8eO

0.6L)

=t.I 0.40.2

I

0.0

0.20 0.15 0.10 0.05 0.00

Pitching Moment Coefficient

Figure 32. CFD Force and Moment Comparisons With Test Data, ACWFT Baseline

Configuration, Mach 0.6.

32

"6

1.2

1.0

0.8

0.6

0.4

0.2

0.0

1.2 1.2 ................................................ _-_

1.0

"6 "6

""i -t 0.4

lestda_ 0.2

H I o.o0 5 10 15 20 0.0 0.1 0.2 0.3 0.4 0.5 0.20 0.15 0.10 0.05 0.00

Angle of Attack (deg) Drag Coefficient Pitching Moment Coefficient



1.0

0.9

0.8

0.7

0.6

0.5

= 0.4-J

0,3

0.2

0.1

O0

5 10 15


1.0

0.9

0.8

0.7

0.6

0.5

I 0 0.4! =I :I O,3

2O

0.2

0.1

0.0

0.0

m

E

0,1 0.2 0.3 0.4 0.5

Drag Coefficient

1.0

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0.0

0.10 0.05 o.oo -o.o5 -oJo




33

Cambered Wing

M = 0.9, _ = 10°

Cp

Inviscid Viscous

0.5

I0.0

-0.5

-1.0

-1.5

Figure 35. Cambered Wing Surface Pressure Coefficient and Streamline Traces, Mach

0.9, 10 Degrees Angle of Attack.

1.2 1.2 1.2 ....... =.........

1.0 1.0 1.0

_ 0.8 _ 0.8 _ 0.5

o, _ o., -_ o,

0.2 0 2 ! 0.2

0.0 0.0 t _ , 0.0

0 5 10 15 20 0.0 0.1 0.2 0.3 0.4 0.5 0.20 0.15 0.10 0.05

Angle of Attack (deg) Drag Coefficient

0.00


Figure 36. CFD Force and Moment Comparisons With Test Data, ACWFT Cambered

Wing Configuration, Mach 0.6.

34

t.2

1.0

0.8.2

0.6

0.4

0.2

0.0

5 10 15 20

Angleof Attack(deg)

1.2

1.0

0.8"6

_ 0,6

_ 0.4

0.2

0,0

0,0 0.1 0.2 0.3 0.4 0.s

Drag Coefficient

1.2

1.0

- 0.8

mo

0.6

0.4

0.2-

0.0

0.20

J

0,15 0.1o oo5 oo0


Figure 37. CFD Force and Moment Comparisons With Test Data, ACWFT

CamberedWing Configuration, Mach 0.9.

1.2

1.0

i 0.83Z

0.6

0.4

0.2 =

0.0

0 5 l0 t5 20

Angle of Attack (¢leg)

,2]i

1.0 _-

/

ool II • _ro.,_= I0.0 0.1 0.2 0.3 0.4 0.5

Drag Coefficient

1.2 _ ...............................................

1.o

._ 0.8

==8 0.8¢J

=.J

0.4

0.2

0.0 ' P

0.10 o,05 0._ -o.os -o.I0


Figure 38. CFD Force and Moment Comparisons With Test Data, ACWFT Cambered

Wing Configuration, Mach 1.2.

0.30

. 0.20e=JE

0.10

=_-_ 0.00

P

-0.10- -0.20

-0.30

IN.I clara

• f,_STD Euler

• N_SI"D Namer.Stokes

0.10

0.08

i |0.08_ 0.0,1

0.02

_ -0.02-0.04

-0,1_

-0.08

-0.10

0

0.4

0.3

i 0.2

i 0.1

0.0

-0.1

E_ -0.2

! -0.3

-0.4

5 10 15 20 S 10 15 2O 5 10 lS 20

AngleOfAttack ((:leg) Angleof Attack(deg) Angleof Attack(dog)

Figure 39. CFD Incremental Force and Moment Comparisons With Test Data, ACWFT

Cambered Wing Configuration, Mach 0.6.

35

0.30

. 0.20c

_o

0.10

0.00m

c

_ -0.100

-0.20

-0.20

i 0.10]

F 0.08

_o 0.06

( ) 0.04

0.02a_ 0.00

E =) -0.02E

,i°°' tT

i _-o.08-0.08i

-0.I0

0

0.4 _r"..................................

0.3

0,2

)(oo_" o-0.1

6

G_=

-0.2

-0.3 t

-0.4I r5 10 15 20 5 10 15 20 0 5 10 15 20

Angle of Attack (deg) Angle of Attack (deg) Angle of Attack (deg)



0.30

- 0,20co

0.10¢,.)

_ 0.00

-0.I0

= -0.20

-0.30

0.10

0.08

0.06u

_= 0.04

0.02

_=_0.00'D'_ -0.02=

-0.04®

-0.08

-0 08

-0.100 5 10 15 20

Angle ol Attack (deg)

0'l- 0.30.2o i

!)o,0.0°-0.1

I_ -0.2

-_ -o.3

-0.4

T..... m

5 10 15 20 0 5 10 15 20

Angle of Attack ((leg) Angle of Attack (dell)



36

Beta = 5° Beta = 0°

(M = 0.6, ct = 150)

Cp

0.5

0.0

-0.5-1.0

-1.5

Figure 42. Surface Pressure Coefficient and Streamline Traces, ACWFT Baseline

Configuration, Mach 0.6, 0 and 5 Degrees Side Slip Angles, 15 Degrees Angle of Attack.

12 t2 12

10

--O8c

¢:

U

.-J04

00

10

_08"

/m _ o_/ (J

£_ (]4.

02,

00'

0 5 10 15 2O O0


mI

J

10

" O8c

04

02

O0

!ov 02 03 04 05 020 0_o oco -olo

Drag Coefficient Pitching Mommnt Coefficient


Configuration, 5 Degrees Sideslip, Mach 0.6.

37

,2F..............................................1.o i.o | I.O _I_

I!°'I'°"0.6 0.6 _ 0.6

0.4 _ 0.4 _ 0.4

02 021 02oo_/ I 00I i .=.,_, il 000 s 10 15 2o 0.0 0.1 0.2 0.3 0.4 o.s 0.10 o.os 0.00 -o.os -0.10

Angle of Attack ((leg) Drag Coefficient Pitching Moment Coefficient



1.2 " 1.2

1.0

0,8-

0,6

¢.)E

0,4-

1.0

i q i

5 10 15


1.2

1.0

0.8

0.6

I0,4 t

0.2

0.0

0.0 0.1 0.2 0.3 0.4 0.5

DrxgCoeff_lem

0.8JR

0.6

0

E04

0.2 " 0.2

0.0 0.0

0 2O 0.10

i

0.05 O.OO -O.(Y3 -0.10




38

0.30 ....................................

= 0.20

0.10¢.)

0.00 ___m

|_ -0.I0

.._ -0,20

-0.30

5 10 15 20


0.10

0.08

0.06 I0.04 +

0.02

=0.04 t

=0,06 1

0 5 10 15 20

Angle Of AtWck (aeg)

0,4

0.3

_o.1

o.o-0.1 _•0.2

--_ -0.3

.0.4

0 5 10 15 20

Angle of AIlaCk ((leg)

I 0.05 i ! 0.05I

0.04 II 0,04

0.03 J 0.03

"°" foo,I 0.01 ]•0.0, _ .0.0,_ _,

z -o.o2 z _o.o21 i

i-0.05 .............. J .0.05 '- J

5 10 15 20 0 5 10 15 20

AIlgle of Atlack (deg) Angle of Attack (de.g)


Baseline Configuration, 5 Degrees Sideslip, Mach 0.6.

0.30

i 0.200,10

;5 0.00 _--_- -

]

i .0,10

--= -0.20

-0.30

0 5 10 15 20

Angle of Attack (deg}

0.05 ...........................

_ 0,03U

0.02

oo,0

.0,01 • •

.002

-0.03.0_0.4

-0.05 ..............

5 10 15 20


0_10 1

_ o.o8io.o6

_ o.ooL--.-_'-I-_"-0.02_

-0.06 _tmdm

.0.10

0 5 10 15 20

Angle of _tack (Oeg)

0.05 ,

0.02

o.0_ Jr

•0.02 _

.0.03 -_ . ,

•0.05L ........ --_0 5 10 15 20


0,4 ........................................

E 0.3

I 0.2Z

i Q.10.0 _- 1"T-

-0.1

-0,2

_-- "0.3

-O.l

5 10 15 20




39

0.30

i 0,200.10

0.00

]

-0.10- -0.20

-0.30

0.10

. 0.08

i 0.06

0.04 J

0.00

-0,02

-0.04

-0.06

-0.08

-0.105 10 15 20


0.05 |

j°o:t

i -0,OlI --0.02

-0.03

0 5 10 15 20

Angle of/Umck (deg)

0.4

" 0,3

0.2

_| o1: |oo

_o -0.1Z

_am,am i -02

"_'_" I -_ -0._

-0.45 10 15 20

Angle of Attack (dell)

0.05 t ..................................0.04

0.03 ÷

0.02

0.01 I0

-0.01-0.02

I-o.o4 , _'_ e_-0.05 L

0 5 10 15 20


II

i5 10 15 20

Angle of AUack (dell)



40

-20 ° (up) Elevon Deflection

ACWFT With Symetric Elevon Deflection

Figure 49. Unstructured Surface Grid About Deflected Elevon.

Symmetric Elevon: 5 E = -20 ° Baseline: 6 E = 0 °

(M = 0.6, (_ = 15 0)

Cp

Figure 50. Surface Pressure Coefficient and Streamhne Traces for ACWFT

Configuration with Symmetric Deflected Elevon l'k.flect|ons of 0 and -20 Degrees, Mach


41

1.2 1.2

1.0

0.8

J_

0.6O

-J 0,4-

0,2

1.0

5 10 15 20

Angleof Attack(deg)

_E 0.8!o

0.6

E_ 0.4

0.2

0.0 0.0

0 0.0 0.1 0.2 0.3 0.4 o.s

Drag Coefficient

1.2

1.0

._ 0.8

i 0,6

_ 0.4

0.2

0.0 p

0.40 0.35 0.30 0.25 0.20


Figure 51. CFD Force and Moment Comparisons With Test Data, ACWFT With -20

Degree Symmetric Elevon Deflection, Mach 0.6.

1.2

1.0

0.8

o

0.60

=0.4

0.2

0.0

5 10 15

Angle of Attack(aeg)

1.2 ]i

1,0 i

-- 0.8

Jfa

0.60E

0.4

0.2

0.0

20 0.0 0.1 0.2 0,3 0.4 0.5

Drag Coefficient

1.2 _-.................................... ]

1.0 _-

oeJR

°61S 0,4

0.2-

0,0 , _

0.40 0.35 0,30 0.25 0.20




1.2 1.2 1.2

1.0

0,8e.,

_o

=_ 0.6

=:0.4

0.2

0.0

1.0 1.0

I

__ 0.6i

I

0.6

0.4

0.2 0.2

m t

0.0 T 7 - 0.0 _ i

0.0 0.1 0.2 0,3 0.4 0.5 0.30 0,25 0.20 0,15 0.10

Drag Coefficient Pitching Moment Coefficient

O.B

0.6

0

0.4

5 10 15 20

Angleof Attack(deg)



42

0.30 0.10 ] 0.4 ]

-- 0.08 / 0.3 t0.04 / i 0.2

0.10 0.02 _ / "_ ,.,t_ 0.1

ooo o.oo[ .._, _ " "=._=_o.o

°°21 \ I _o-o.1_-o.1o .0.041 \ T t-O2= -o.o6t i"1-0.20 -008_ ! ,,'_m_ I j _ -o3-0.30 -0.10 I -0.4

0 5 10 15 20 0 5 10 15 20 0 5 10 15 20

Angle of Attack ((:leg) Angle of Attack (deg) Angle of Attack (deg)


With -20 Degree Symmetric Elevon Deflection, Mach 0.6.

0.30 0.10 i 0.4

0.08

= I"n ._ o.os_ 0.10 _ 0,040.02 :_ 0.2

.._ 0.00 ' _ 0,00 EmI' -0.02

-0,10 ,_ .0.1

! ioo, ...--= -0.20 • • _ .0.06 [ _uam o

-- • NASTOEu_ c•0.08 1 -- -0.3

•0.30 -0.10 | -0.4

0 5 10 15 20 0 5 10 15 20 0 5 10 15 20

Angle of Attack (deg) Angle of Attack (deg) Angle of Attack (dog)



0.30

0.20o

0.10

,1=,,_ 0.00

|-0.10

¢: -0.20

-0.30

m111

B

0.10 |

0.081

0.06 l0.04

0.02

ooo- -.0.02

-0.04-

.0.10

0

0.4

_ 0.1

°iE. oo-0.1

i .0.2

-_ .0.3

-0.4

5 10 15 20 5 10 15 20 5 10 18 20

Angle of Attack (deg) Angle ol Attack (deg) Angle of Attack (deg)



43

(down)leftelevon

-20 o (up) right elevon deflection

ACWFT With

Figure 57. Unstructured Surface Grid About Asymmetrically Deflected Elevon andDeflected Body Flap.

Asymmetric Elevon:

8E= +20°/-20 °

(M = 0.6, = 15 °)

Baseline Afterbody Flap: 81== 90°%

0.5O.C

Figure 58. Surface Pressure Coefficient and Streamline Traces, for ACWFT Baseline,

Asymmetrically Deflected Elevon, and Deflected Afterbody Flap Configurations, Mach


44

1.2

1.0

0.8

_ 0.6

•,J 0.4

0.2

0.0

.................................... "1

5 10 15 20

Angle 04' Attack (deg)

1.2

T.O

._ 0.8

_ 0.6

_ 0.4

0.2

0,0

0.0 0.1 0.2 0.3 0.4 0.5

Drag Coefficient

1.2

1.0

._ 0.8o

_ 0.6

_ 0.4

0.2

0.0

0.20

i

o.15 o._o 0.05 o.oo

Pltchklg Moment Coefficient

Figure 59. CFD Force and Moment Comparisons With Test Data, ACWFT With

Asymmetric Elevon Deflection, Mach 0.6.

¢J

E.,..i

1.2 t'

1.0

0.8

0.6

0.4

0.2

0.0

1.2

1,0

0.8

_ 0.6

g0.4

1.2 ,I

1.0 i

0.8

_ 0.6

_ 0.4

o.2I ' _"_--0.0 [ '"

0.0 0.1 0.2 0.3 04 0.5

0.2

0.0

5 t0 15 2O 0.20 0.15 0.t0 0,05 0.00




1.2

1.0

0.8

i 0.6

_ 0.4

0.2

o.o 10 5 10 15

Angle of Attack (dog)

1.2

1.0 ¸

i _ 0.8

i _ o.8

_J 0.4

o.2t

J0.0 I

20 0.0

Eu_

01 0.2 0.3 0.4

Drag Coefficient

0.5

1.2

1.0

0.8

06

0.4

0.2

o.o I0.10

I

0.05 0.00 -0,05 -010




45

0.30]-........................................i! o.oa°"°0.20 4 i E

0.10 0.04

__ 0.02

0.00 0.00

-0.10 -0.02

-0.10

0 5 10 15 20

Angle of Auack (deg)

0.05

i 0.040.03

- 0.02

0.01

.0.0t

-0.02

j .0.03.

_ -0.04

.0.05

0 5 10 15 20

Anglo of AtBck [deg)

0.05

i o.o410.03

0.02

I 0.01

g o

_ .0.01E

_ -0.o2

i .0,1_

4?.05

0.4

0.3

0.2

• i _=E o.1

i -0.1

!• w_sraeB - -0.3

-0.4

5 10 15 20 0 5 10 15 20

Ar_le of A_ek (¢_) _ of _et=¢=(¢_)

5 10 15 20

Angle of ,4Lttack (deg)


With Asymmetric Elevon Deflection, Mach 0.6.

o.3o ...........................................!

0.20

¢J

0.10

._ 0.00

-0 10

c .020

.030

010

i 008

i ¢, ._ 0.06

i _ o,o4I _ 0.02

ooo! i -002! -004

-0,08

-010

5 10 15 20

Atigle of Attlck ((leg]

005 7¢

004+

003 •

-- 0 32 •

001

I

-001>.

-0.02

i -0.03

E

_ -0.04

-0.05

I

0 5 10 15 20

Angle of Attack ((:leg)

0 05

003 ;

O02JOOti .0o'i

I•005 '

5 10 15 20


0 5 10 15 20

Angle of AXlack (deg)

0.4

,_ 0.3 t

0.2

:E

_'_ oo_-01 I *E-0.2 i

¢ -0.3 j-04

0 5 10 15 20

Angle of Attack (cleg)



46

0.30

._ 0.20u

0.10o¢:

0.00

i -0.10

-- -0,20

-0.30

ig

!!

5 10 15 20

AngM of Attack (deg)

0.05 ................

| o,. iI 0.03 i

o 0.02 t

o.01

0 .... __--_t

-0,01

-0.02

-0.03

-0.04

-0,05

0 5 10 15 20

Angle of AtzJck (deg)

0.10

0.08

0.06

0.04

0.02

0.00

-0.02

-O,04

-0.06

-0.08

-0.10

I---I5 10 15 20

Angle of AttIck (dog)

0.05 !

0.04 t0.03

0.02 I •0.01

E

| -0.°21

I '03 t--'' IS -0.04 • NAS'rDEder•0.05 ' " I

0 5 10 15 20

Angle of Atlack (deg)

0.4 .................

°_02

io,

_ -0.2

--_ -0.3

-0.4

0 5 10 15 20

Angle ol Anlck ((leg)



1.2

1.0

. 0.8c.,t

0.6

O

0.4

0.2

0.0

o 5 10 15 20


1.2 ...................... _ 1.2

1.0 I i i 1.0o.5 i "E 0.8

!°°I °'-J 0.4 I 0.4

/

°21 I 02_dm

0.0 0,0

0.0 0.1 0.2 0.3 0.4 0.5 0.20

Drag Coefflckmt

o.15 o.1o o.o5 o.oo



Deflected Body Flap, Mach 0.6.

47

1,2 .......................................

1'0 B_0.8

0,6

"_ 0"4 t

0.2

0,0 ,

1.2

1.0

-- 0.8

_ 0.4-

0,2-

1.2 ............................

1.0

_ 0.8

i 0,6

0.4

0.2

0.00.0 L b I _ i

0 5 10 15 20 0.0 0.1 0.2 0.3 0.4 0.5 0.20 0.15 0.10 0.g5 0.00




1.2 ¸

1.0

_ 0.8c

0.6

S 0.4

0.2

0.0 ,

0 5 10 15

_gkD of Attack (dag)

1.2 ........................................ ,

20 0.5

1.0 _"

0.8 ÷

0.6

0,4

0.2 "t

0.0

0.0 0.1 0.2 0.3 0.4

Drag Coefficient

1211.0 I

0.8

0.6

0.4

0.2

0.0

0.10 0.05 0.00 -0.05 -0.10




48

0,30 ......................................

E 0.20

i 0.10o

-0.I0

--_ -02D

-0.30

5 10 15 20


0.05 ............................

0.04

0,03

0.02

0.01 ;0

>_ .0.01

"i -0.02

-0.03

-0.04

.0.05

0 5 10 15 20

Ang4e of _tack (oog)

0.10

0.08

0.06

0.04

0.020.00

.0.02

-0.04

o: I .,-°,,.I-0.10

5 10 15 20

Angle of Attack (de9)

0.05 |

::o°;I0.02

0.01

-0.01

.0.02 1

.o.o3I-- .0.05

0 5 10 15 20

Ang_ of Attack (deg)

0.4 ...............................

i 0.30.2

=_ o.1

_ o.o_o -oij.o:

--_ -0.3

-0.4

t

5 10 15 20



With Deflected Body Flap, Mach 0.6.

0,30 ................

0.20

0.10t.J

0.00 ------4-------+---------

-0.30 .......................

5 10 15 20


0.05=

0.04

:_ 0.03o- 0.02

o.01

_ -0.01

_ -O.02

-0,03|.0.04

!;

-0.055 10 15 20

Angle of Attack (dog}

0.10 ,

0.08 _'

oo,t:.°0:L_o:oo

..0.06 1 _*"'_'-- .0.08 ; ! I _' I

0 5 10 15 20

Angle of Atlack (deg)

0.04 ÷

0.03 t

°°2I Io.o,_ /

-0.01 _ I

•0.02 4

•0.03 Jl

_lWtdatl I.0.04 i II _110 F.uler

-0.05(0 5 10 15 20

Angkl of AnaCk (dog/

0,4 ,

j:::I

I! 0.1

0,0

-0,1

-0.2

-- -0.3 ¢

-0,4 [0 5 10 15 20

Angle Of Atlack (deg)



49

0.30

0.20

0.10cJ¢=

..= 0.00

i -0.10

--_-0.20

-0.30 ...................................

5 10 15 20

Angle of Aeack (e_g)

0,0S -

-_ 0.04.

0.03

0.020.01

J __ 0 .... _

_ 1.01

_ 41.02

-0.03-0.04

_).055 10 15 20

Angle of Allz_k (deg)

0.10

0.08

.._ 0.06

0.02

0.00 _

i -0.02

.0.04

•0.06 _u_== I

•0.08 , I _,,¢Io =''-' I-0.10

5 10 15 20

Angle of Anack (deg)

: 0.05 ]-i oo4t l--u== I

io.o3t I " N_Eu=I;oo t

o.ol _

_ o_

-0.0t [

-0.02 1

-0.05 I

0 5 10 15 20


0.4

=_ 0.1

E_ o.o

O .0.1-0.2

--_ .0.3

-0.4

0 5 10 15 20




50

REPORT DOCUMENTATION PAGE For,.Ap_o,.dOMB No. 0704-0188

Public reporting burden for this collection of information is estimated to average I hour per response, including the time for reviewing instructions, searching existing.data sources,

gatherin 8 and maintaining the data needed, and completing end reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this

cot|ection of information,including suggestions for reducing this burden, to Washin_on Headquarters Services, Directorate for information Operations and Re'ports, 1215 Jefferson

Davis Highway, Suite 1204, Arlington, VA 22202-4302, and to the Office of Management and Budget, Paperwork Reduction Project (0704_0188), Washington, DC 20503

! 1. AGENCY USE ONLY(Leave blank) i 2. REPORT DATE 3. REPORT TYPE AND DATES COVERED

I March 1998 Contractor Reporti4. TITLE AND SUBTITLE 15. FUNDING NUMBERS

Euler Technology Assessment for Preliminary Aircraft Design -

Unstructured/Structured Grid NASTD Application for Aerodynamic

Anal_'sis of an Advanced Fighter{Tailless Configuration

6. AUTHOR(S)Todd R. Michal

7. PERFORMING ORGANIZATION NAME(S) AND ADORESS(ES)

Boeing Company

St. Louis, Missouri

9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES)

National Aeronautics and Space Administration

Langley Research Center

Hampton, VA 23681-0001

NAS1-20342, Task 13

WU 522-22-11-01

8. PERFORMING ORGANIZATIONREPORT NUMBER

10. SPONSORING/MONITORINGAGENCY REPORT NUMBER

NASA/CR-1998-206947

11. SUPPLEMENTARY NOTES

Langley Technical Monitor: Farhad Ghaffari

12a. DISTRIBUTION/AVAILABILITY STATEMENT 12b. DISTRIBUTION CODE

Unclassified-Unlimited

Subject Category 02

Distribution: Standard

Availability: NASA CASI (301) 621-0390

13. ABSTRACT (Maximum 200 words)

This study supports the NASA Langley sponsored project aimed at determining the viability of using Euler

technology for preliminary design use. The primary objective of this study was to assess the accuracy and efficiency

of the Boeing, St. Louis unstructured grid flow field analysis system, consisting of the MACGS grid generation andNASTD flow solver codes. Euler solutions about the Aero Configuration/Weapons Fighter Technology (ACWFT)

1204 aircraft configuration were generated. Several variations of the geometry were investigated including a standard

wing, cambered wing, deflected elevon, and deflected body flap. A wide range of flow conditions, most of which

were in the non-linear regimes of the flight envelope, including variations in speed (subsonic, transonic, supersonic),

angles of attack, and sideslip were investigated. Several flowfield non-linearities were present in these solutions

including shock waves, vortical flows and the resulting interactions. The accuracy of this method was evaluated by

comparing solutions with test data and Navier-Stokes solutions. The ability to accurately predict lateral-directional

characteristics and control effectiveness was investigated by computing solutions with sideslip, and with deflected

control surfaces. Problem set up times and computational resource requirements were documented and used to

evaluate the efficiency of this approach for use in the fast paced preliminary design environment.

14. SUBJECT TERMS

Computational Fluid Dynamics, Euler Formulation, Preliminary Aircraft Design,

Advanced Fighter Design, Vortical Flows, Unstructured/Structured Grid NASTD

17. SECURITY CLASSIFICATIONOF REPORTUnclassified

¼SN 1540-_)1-280-5500

18. SECURITY CLASSIFICATIOhOF THIS PAGEUnclassified

19. SECURITY CLASSIFICATIOI_OF ABSTRACTUnclassified

IS. NUMBER OF PAGES

55

16. PRICE CODE

A_420. LIMITATION

OF ABSTRACT

_tandard Form 298(Re¥. 2-89)PrescribedbyANSIStd Z39-18298-102

Euler Technology Assessment for Preliminary Aircraft Design ......preliminary design tools in the...

Documents

Transcript of Euler Technology Assessment for Preliminary Aircraft Design ......preliminary design tools in the...