[American Institute of Aeronautics and Astronautics AIAA Modeling and Simulation Technologies Conference and Exhibit - Montreal,Canada (06 August 2001 - 09 August 2001)] AIAA Modeling

c)2001 American Institute of Aeronautics & Astronautics or Published with Permission of Author(s) and/or Author(s)' Sponsoring Organization.

A01-37400

AIAA2001-4423Applications of a Simulation Environment DuringWind Tunnel Testing

Randy S. HultbergDavid R. GingrasJeffery W. Bell

Bihrle Applied Research Inc.Hampton, VA

Modeling and Simulation Technologies Conference & ExhibitIn Association with the Canadian Aeronautics and Space Institute

6 - 9 Aug 2001Montreal, Quebec, Canada

For permission to copy or to repuhlish, contact the copyright owner named on the first page.For ATAA-held copyright, write to ATAA Permissions Department,

1801 Alexander Rell Drive, Suite 500, Rcston, VA, 20191-4344.


AIAA 2001-4423

APPLICATIONS OF A SIMULATION ENVIRONMENTDURING WIND TUNNEL TESTING

RANDY S. HULTBERG*DAVID R. GINGRASf

JEFFERYW.BELL*Bihrle Applied Research Inc.

18 Research DriveHampton, VA 23666 USA

(757) 766-2416 FAX (757) 766-922

AbstractThere is a distinct parallel between software

requirements for hardware in the loop simulation anddynamic wind tunnel testing. Timing, hardwareinterface, and usability are key factors in overallsystem quality. A reconfigurable PC-Basedsimulation environment running under the Windowsfamily of operating systems met timing requirementsneeded for dynamic testing and was selected toperform data acquisition, processing, and displaytasks in two different efforts. As expected, theapplication of the software was a success. Thereconfigurable structure of the software allowed forthe quick design and implementation of software foreach test.

This paper provides a discussion of thesoftware chosen for use during the wind tunnel testingwith primary focus on simulation loop timing, andhardware interface. In addition, the paper presentsdetails pertaining to the specific application of thesoftware in the two wind-tunnel testing efforts.

IntroductionA typical wind-tunnel test to collect

aerodynamic forces and moments consists of a testarticle, force and moment measurement device, modelsupport (rig), signal conditioners, data acquisition andrecording devices, and data monitor. Additionalcomponents of wind-tunnel test apparatus can be rigcontrollers, model actuators, and a variety of tunnelstate monitors. As once would expect, the specificobjectives of test dictate the complexity of the test

setup. For example, dynamic wind tunnel testing requiresthe test article to be subject to a rotation and/or translatedmotion where position data must be synchronized with theacquired force and moment data. The resulting data mustthen be reduced and displayed for test engineers'inspection. The apparatus and data requirements fordynamic wind tunnel testing are not unlike those of real-time hardware in the loop simulations. Like real-timehardware in the loop simulations, overall system timingand communication are of the utmost importance. Oncereceived the data must then be processed and provided tothe pilot or test engineer. In hardware in the loopsituation, the pilot responds to stimuli such as graphicaldisplays or motion, whereas the test engineer andcontrollers respond to data reduced and provided in atimely manner.

Given the parallel between hardware in the loopsimulation and dynamic wind tunnel testing, commontools should be relevant to each application. Theremainder of this paper provides details pertaining to twosuch applications. A discussion of components andfunctionality of the selected simulation environment ispresented in addition to information about the specificapplication of the environment in each of the tests.

Symbols

Simulation execution time error.

Simulation clock time.

Simulation frame duration.

Reference simulation start time.

Reference clock time.

'ExF

Simulation

AT:Simulation

* Start

T•'-True

* Chief Test Engineert Senior Engineer, Senior Member% Chief Software EngineerCopyright © 2001 by Bihrle Applied Research Inc.Published by the American Institute of Aeronauticsand Astronautics Inc. with permission

Simulation Environment

OverviewThe simulation environment chosen for the wind

tunnel test efforts was the Bihrle Applied Research Inc.

1American Institute of Aeronautics and Astronautics


AIAA 2001-4423

(BAR) D-Six simulation environment. D-Six is areconfigurable PC-Based simulation environment thatruns in the Windows family of operating systems.Modular reconfigurability allows it to be extended toa number of applications (References 1 and 2).

D-Six uses a component-based architecture,such that the required core components are concernedprimarily with loading and executing simulationmodels. Additional functionality is integrated byloading individual component libraries that interactwith D-Six through one or more defined interfaces.Figure 1 contains a high-level diagram of the D-Sixsoftware structure and highlight key features in theenvironment.

The software consists of two major classesof components, project independent and projectdependent. Project independent components arereusable from project to project and include code thatmaintains the core simulation loop and controlstiming. Project independent components handle thegraphical user interface, modular hardware interface,the simulation database manager, and the simulationfunction library. D-Six architecture allows for projectindependent reconfigurability by the implementationof additional user developed modules or plug-ins.Project dependent components of the environmentare, as their name indicate, based on project specificinformation and consist of settings files and a projectdynamically link library (DLL). Contained in thisDLL, is specific project base code that is developedby the user. In flight model applications, user definedaerodynamics, propulsion, and control modelequations are implemented. This project-based codealso contains an interface to the project independentcomponents mentioned above in order to takeadvantage of the various resources provided.

As mentioned in the introduction, allhardware in the loop applications, timing, hardwareinterface, and user interfaces very importantcontributors to a successful system. In the D-Sixenvironment these duties are project independent andreusable. A detailed discussion of these tasks isprovided in the sections below.

TimingThe topic of timing and simulation can be a

confusing and controversial subject. The followingdiscussion is provided to explain the timing of the D-Six simulation loop and present evidence that it meetsthe timing needs of the simulation and wind-tunneltesting tasks for which it is applied. For the purposeof this discussion, a distinction is made between theterms "true time", "real-time" and "real-time system"."True time" refers to a reference time used to track

simulation timing. The use of the term "real time" in thecontext of this paper refers to timing of sufficient accuracyto solve the problem at hand. For the closed loop windtunnel testing efforts, a 10-millisecond accuracy indesired. The term "real-time system" refers to a commondefinition found in Reference 3, provided below.

"A real-time system is one in which thecorrectness of the computations not only depends uponthe logical correctness of the computation but also uponthe time at which the result is produced. If the timingconstraints of the system are not met, system failure issaid to have occurred."

Based on these definitions, with the exception ofWindows CE, the systems running any of the Windowsfamily of operating systems cannot be considered "real-time systems." However, this does not imply that softwarerunning under Windows operating systems cannot meet"real time" for a particular task.

With this in mind, since Windows is not a real-time system, there are no guarantees that a simulation timeframe can be executed within a fixed period of time.While there are no guarantees Windows will not preemptan application process and unnecessarily delay it's timeframe execution, there is nothing inherent in the operatingsystem that would cause this to happen. Timing errors arecaused solely by software installed on the system.Therefore, it is possible to determine the total timing errorfor each time step using the Windows performancecounter and use the information to detect errors greaterthan maximum values for a predefined task, thusproviding a degree of determinism. In cases where thesemaximums are not met, it is generally possible to identifyand correct the source of the error.

The D-Six simulation loop uses an errorcorrection model to maintain relative real-time operationas shown in the simplified state diagram (Figure 2).

During the begin state, the initial true time isdetermined, and the simulation time is set to zero prior tocalling the initialization routines for the simulation model.The wait state normally performs background operationsas well as effective delay prediction to reduce systemload, but for simplicity these operations are left out of theabove state diagram. The simulation holds in a wait stateuntil it is time to execute simulation frame. The executionstate increments the current simulation time by a singleframe length, then executes a single simulation step, uponcompletion it returns to the wait state.

In this loop, the Windows performance counterdetermines timer resolution and is dependent on theprocessor and clock speed. For example, an Intel PentiumIII-500MHz system has a performance counter frequency(PCF) of approximately 3MHz, where an Intel Pentium II233MHz system has a PCF of approximately 1.2MHz.

American Institute of Aeronautics and Astronautics


AIAA 2001-4423

The simulation frame execution time errorcan be computed at the start of the execution stateusing equation (1).

**ExF ~ True ~ * Start ~~ * Simulation ~ ̂ * Simulation {*•'

This value is a measure of the differencebetween a scheduled time for a simulation frameexecution and the actual time it is executed.

A series of tests were performed to evaluateD-Six timing. A D-Six simulation was run underWindows 2000 in conjunction with a CPU stress toolincluded with the Windows Platform SoftwareDevelopment Kit (SDK). The test configuration isprovided in Table I.

Table I. Timing test parameters.

Computer:

Operating System:

D-SixConfiguration:

Simulation Frame:

SimulationDuration:

Intel Pentium II 500 MHz

Windows 2000 Professional

• Six-Degree of FreedomFlight model. Includingnonlinear table look up.

• 600 x 600 Pixel DirectXGraphics Window

• A Single DirectX driven"real-time" strip chart.

• A Single Microsoft SideWinder USB game stick.

80Hz

10 seconds

The accuracy of these tests is based on theassumption that the processor clock is accurate. Thisassumption is valid based on publicly available data(Reference 4).

Simulations were executed at four differentlevels of system load defined by the Window Systemload tool, low, medium, high, and critical. The resultsare provided in Figure 3. Table II contains a summaryof test loads and maximum errors attained.

The test results show that under low systemload, the D-Six environment is capable of reasonableframe scheduling within 3 ms of true time and meetstiming requirements specified for this effort.Depending on task requirements, the environmentmay achieve sufficient timing under medium systemloading, but would not be recommended for use underhigh or critical loads.

Hardware InterfaceInterface with hardware and software

devices is handled by a core component in D-Sixnamed IOD (Input/Output Devices). Hardware

devices are integrated with D-Six through IOD by loadinga module that registers the hardware's capabilities (Figure4). Software devices also register their capabilities withIOD, and can be hardware emulators, or can represent auser interface, for example, a graphic of a switchboardthat responds to mouse clicks. IOD provides an efficientand consistent mechanism for passing informationbetween different D-Six components in a generic,hardware independent manner.

Devices are registered with IOD by passing adescription to a registration function. The description isessentially a flexible structure that declares channels ofanalog input and output, individual switches, and switchbit fields. IOD provides mapping between analog I/Ochannels and simulation variables with optional linearscaling. Channel scaling capabilities include setting aminimum and maximum, or setting an offset andmultiplier. Switches are used as triggers for softwareevents that are also registered in IOD. A software eventcan be any action carried out by a function call thatreceives a single flag that indicates if the event is beingtriggered on or off. Switches can be set to level or edgetriggered and can be set active when zero on non-zero.

Because IOD performs all of the common dataoperations most hardware I/O require, writing acomponent that registers a device with IOD is generally

Table II. Timing test results

Load

Low:D-Six +

MediumD-Six +

High:D-Six +

no other apps

1 Normal Priority Thread

3 Normal Priority Threads

Critical:D-Six + 4 Threads Above NormalPriority

ExecFrame

[ms]

2.78

19.71

252.06

617.34

trivial. Module wizards provide access to D-Sixfunctionality in addition to MFC and DirectXcomponents.

During execution, IOD polls each device beforeperforming operation on the input and output data.Components with a device that requires polling use thisexecution time to communicate with the device.Components that do not require polling must wait untilthey are polled to read or write data to IOD. IOD is notthread safe and will only poll a device before the data is tobe used, which is typically coincident with each



AIAA 2001-4423

simulation frame. However, users have the option ofslowing this rate to compensate for slow hardware.Hardware that requires faster polling times may behandled by a separate thread that communicates withIOD only during polling calls on the main thread.

User InterfaceD-Six contains a graphical user interface that

allows users to configure a workspace that is bestsuited for the tasks at hand. The interface is acollection of tools that provide the user a number ofoptions to access data before, during, or aftersimulation run and are listed in Table III.

Table III. D-Six user interface tools.Tool

DirectX Graphics

InstrumentationBuilder

Sound Utility

Strip Charting

Plot UtilityData Analysisand EditingProject Overdrive

Script Engine

DescriptionDirectX base 3-D graphics forout-the-window or externaldynamic views.Script controlled DirectX basedinstrumentation display packageto configure custom displays,gauges, or dials from bitmaps forgraphical display of information.DirectX based sound module thatallows volume and frequencyvariations of .WAV files to bedriven by simulation data andtriggered as mulation events.Multi-window real-time plottingtoolPost run data plotting utility.Reconfigurable post run dataeditor and analysis tool.Environment functionality to testproject algorithms by drivingsimulation with data that hasbeen imported from previoussimulations, flight, or previousexperiments.Script control of simulation tasksand data. The script engine alsoprovides easy access to thirdparty software for auto-reportgeneration.

Wind-Tunnel ApplicationsAs mentioned above, the D-Six environment

was included in the design of two wind-tunnelexperiments where the acquisition, processing, andfeedback of data were critical. The two wind-tunnelexperiments were associated with Small BusinessInnovative Research (SBIR) efforts for the US AirForce. The following sections provide detailspertaining to each of the applications focusing on the

integration of the simulation environment during the tests.

Formation FlightAs part of SBIR AF98-175 Phase II tasking,

Bihrle Applied Research developed a wind tunnel testingcapability to measure the effects of close formation flighton the aerodynamics of a trail vehicle in a two-shipformation (Figure 5). Because of the complex nature ofaerodynamic interference from flying in formation and theeffect of these interactions on control laws design,accurate physical modeling is important for the simulationfor automated refueling or station keeping tasks(References 5 and 6).

As part of this research, two advancedconfiguration subscale wind tunnel models were tested inthe Langley Full Scale Tunnel (LFST) in a number ofrelative formation positions. A majority of the data wascollected with no control surfaces deflected and produced"clean" configuration formation effect data. Part of theobjective in collecting the formation data was to assesscontrol power requirements to keep formation as well asto determine potential performance benefits.

To further investigate the effect of formation oncontrol power and trimmed performance, a model withactuated surfaces was placed in trail formation positions.The actuators were controlled by a real-time closed looptrimming control law during static formation points aswell as dynamic vertical translations through formationpositions.

Test DesignActive trimming portions of the wind-tunnel tests

relied on the use of the D-Six simulation environment fordata acquisition, monitoring, and processing. All datareduction, control-law implementation, control surfacecommand, and data recording were handled by projectdependent code loaded into D-Six. The function diagramin Figure 6, shows data flow for during the test. Data wereacquired from the six-component strain gauge balancemounted in the trail vehicle using the IOD component ofD-Six to interface with a National Instruments PCI Analogand Digital I/O board. This board was capable of 8differential analog inputs that could be processed with a16-bit analog to digital converter at a data acquisition rateof 20 kHz. Project code was run at a 50 Hz frame rate.

Once acquired, balance voltages were convertedto aerodynamic coefficients. Pitch and roll coefficientswere then used as feedback channels in a trim control law,Figure 7, which was a simple PI controller to maintain areference pitching moment and rolling momentcoefficient. Resulting control surface commands were sentto the digital output channels and fed to the surfaceactuators on the model.



AIAA 2001-4423

In addition to trim capability, the controllerwas configured to follow prescribed deflectionprofiles, like sine waves and ramps built using agraphical interface. The controller was also capableof following deflection commands fed by a controlstick interfaced through IOD.

User interface for the test consisted ofseveral graphical real-time strip charts displayingcontrol surface deflections and aerodynamicscoefficients. Dynamic digital displays of a number ofproject parameters were also used.

The use of the simulation environment andimplementation of control laws were a success. Staticdata were collected for each point in a grid offormation positions with active trimming on and off.Dynamic tests were also performed where theactuated model translated vertically through the leadvehicle wake. Forces and moments were measuredduring these tests with trim active and not active.Figure 8 contains sample plots pitching moment androlling moment coefficients from each of these cases.

Virtual Flight TestIn fulfillment of requirements for Phase I of

SBIR AFOO-297, Bihrle Applied Research was taskedto demonstrate concepts associated with VirtualFlight Testing (VFT)(Reference 7 and 8). VFT is aground testing method where a full-scale aerialvehicle, such as a missile, is placed in a wind tunnelusing minimal constraint and commanded to duplicatein-flight motions. The main objective of this testtechnique is to provide a control environment to testcontrol strategy, hardware, and other key systemcomponents.

To investigate test methodology and modelsupport geometry, a sub-scale wind-tunnel test wasperformed. The main objective was to excite vehiclemotions in order to analyze test article response andpredicting trajectories using data acquired during thetest. The data acquisition, processing, simulationtasks were performed using the D-Six simulationenvironment.

Test DesignThe main objective during the test design

was to minimize set-up cost and maximize efficiency.Because of its accessibility and sufficient size, theOld Dominion University three- by-four-foot lowspeed wind tunnel was chosen for the test. A custommodel support and measurement system was designedfor the test that allowed the required freedom ofmotion in pitch while measuring translational forces.

The measurement system consisted of twin strain gaugebalances mounted in the main rig supports. Figure 9contains a photo of the AIM-9 like article chosen for thistest mounted on support apparatus. The model wasdesigned and fabricated by Bihrle Applied Research andcontained actuated pitch control fins, a potentiometer tomeasure attitude, and a piezo electric gyro to measurepitch rate.

As part of the test design, a simulation of the testarticle was developed and used to determine loadconstraints for the test and tune control laws. Using themissile DATCOM (Reference 9), a nonlinear six-degree-of-freedom aerodynamics model was generated thenimplemented incorporated into a VFT simulation, whichcontained models of all key components for the VFT test.These models included, propulsion, control, equations ofmotion, weight and balance, and a wind tunnel rig model.

Since the simulation modeled the test articlemotions and loads in the wind tunnel, it was easilyadapted to model balance output and stimulate datareduction routines (Figure 10).

Once test software design was complete, the D-Six simulation environment was configured to providedata acquisition for force, attitude, and rate measurements,in addition to data processing and control of the missilemotion while providing information to the engineersmonitoring the test. Figure 11 contains a functionaldiagram of the integrated test arrangement. As thediagram indicates, large portions of code were identical tothose in the simulation study. The control lawimplemented in D-Six project code transmitted controlsurface commands, reflecting user control or predefinedmotions, to the test article. Balance and sensor voltageswere acquired through data acquisition hardware andprovided to data reduction routines that computed pitchattitude, pitch rate, and aerodynamic coefficients. Theseaerodynamics coefficients converted to forces using userdefine simulation conditions and summed with propulsiveforces from the test article simulation. These forces withaerodynamics moments computed by the simulation wereused in the simulation equations of motion to predicttrajectory.

The hardware configuration of D-Six was similarto its application during the formation wind-tunnel test.Additional signal conditioning was applied to the analogsignals from the strain gauge balances and the canard andpitch attitude pots located in the test missile. It includedamplification of the balance signals for optimumresolution during the analog to digital conversion as wellas filtering of any high frequency noise in the signal. Anaccurate power supply was used for power to the straingauge balance and position pots.

The user interface for the missile tests wasconfigured to display aerodynamic coefficients and pitch



AIAA 2001-4423

rates in real time using graphical strip charts. Adynamic three-dimensional graphical display of thetest article was driven with test data was used tomonitor test motions remotely. Digital displays ofproject parameters were also used.

ResultsDuring testing, data were collected from

static and dynamic conditions. Static test points wereset using the control surface to set the test articleattitude. Dynamic tests consisted of a number oframps, sine wave oscillations with prescribedamplitudes and frequencies, and a series of randommotions. For each of the dynamic test cases,simulation initial conditions for weight and balance,velocity, altitude, and thrust used in the trajectorypredictions. Figure 12 contains a screen shot of theuser interface during one of the dynamics tests. Plotsof predicted altitude, pitch angle, and surfacedeflection are plotted against time in the real-timeplot tool.

DiscussionThe efforts described above provide support

for the use of Windows based "real-time" simulationenvironments in time-critical applications. Timingtests with a low-stress system load revealed that asimulation environment with DirectX based graphicswill execute simulation frames well within threemilliseconds of scheduled times. These tests supportthe applicability of the environment for applicationssuch as hard-ware-in the loop simulation. Because ofvast similarities between hardware in the loopsimulation and interactive-dynamic wind tunneltesting, the simulation software's applicability wasextended to the experimental tasks of data acquisitionand test control in two different efforts.

The reconfigurable simulationenvironment's, D-Six, use in active trim testing information and virtual flight testing, allowed engineersto design and test reduction and processing code with

greater flexibility, in less time, than with conventionalmethods. To facilitate test execution, engineers configuredthe D-Six user interface to include "real-time" parametermonitoring in the form of graphical strip charts and digitalreadouts. The interface for the VFT effort was furtherenhanced by with the use of three-dimensional graphics todisplay test article motions.

Overall, the application of the D-Six simulationenvironment in the unconventional hardware in the loopscenarios was a success, saving each project time andmoney in the design and execution of the respective tests.

References

1 Bell, J.W. and O'Rourke, MJ. Application of aMultipurpose Simulation Design, AIAA 97-3798.

2 Gingras, D. "AV-8B 11+ Device S2F176 Flight ModelDevelopment Using A PC-Based SimulationEnvironment," AIAA 99-4191, August 1999.

3. Gillies,D, ([email protected]), FAQ newsgroup,comp.realtime, www.faqs.org/faqs/realtime-computing/faq/, July 2001.

4 Microsoft Developer's Network Library, April 2001.5 Blake, W.B and Gingras, D.R. "Comparision of

Predicted and Measure formation Flight InterferenceEffects," AIAA 2001-4136, August 2001.

6 Gingras D.R., Player J.L., and Blake W.B., "Static andDynamic Wind Tunnel Testing of Air Vehicles inClose Proximity," AIAA-2001-413, August 2001.

7 Marquart, E. J. and Lawrence, F.C. "Virtual FlightTesting (VFT) at the Arnold EngineeringDevelopment Center," Presented at ITEA 1999Conference, Tullahoma, TN, October 1999.

8 Gebert, Glen, Kelly, Joy, and Lopez, Juan; "VirtualFlight Test (VFT) Modeling and Assessment"; TEASReference Number: 9800723-90U; 30 September1998.

9 Air Force Research Laboratory, Wright Patterson AFB,AFRL-VA-WP-TR-1998-3009.



PROJECTINDEPENDENT

AIAA 2001-4423

Main GraphicalUser Interface

Main Sim Loop

PROJECTDEPENDENT

IOD(Device Interface)

Graphics

Real-TimeRotting

Default SimulationFunctions

DatabaseManagement

Project Code

Figure 1. D-Six software structure.

BeginSimTime = 0

StartTime = TrueTimeInitSimQ

Wait

False/ StartTime +SimTime + DeltaTime

= TrueTime

True

ExecuteSimTime += DeltaTime

StepSimQ

JFigure 2. Simulation loop state diagram.



AIAA 2001-4423

IIL.a

20 f

10

400

200

1000

500

V^M^J^V/ V^vwJ/W^J Vr^^VwMfV> t̂«A^

LOW LOAD: D-Six + No Other Applications

MEDIUM LOAD D-Six + 1 Normal Priority Thread

HIGH LOAD:D-Six + 3 Normal Priority Threads

'

CRITICAL LOAD: D-Six + 4 Above Normal Priority Threads

3 4 5 6Time [sec]

8 9 10

Figure 3. Simulation frame execution scheduling test results.

Project VariableMapping GUI

Figure 4. The D-Six reconfigurable device interface.



AIAA 2001-4423

Figure 5. Test articles during two-ship formation-flight testing in the Langley Full Scale Tunnel

-»National

Instruments Board

Analog Input(w/5Hz Filter)

Digital Output

Figure 6. Hardware configuration for the active trimming wind-tunnel experiment



AIAA 2001-4423

—Cm Re1>Q

PitchCommand

(i.e. Sine Wave,Stick Input etc.)

RollCommand

(i.e. Sine Wave,Stick Input etc.)

Figure 7. Functional diagram of the active trimming control law implemented during testing.

ClMom a nl Coaticianl

Figure 8. Sample results from active trim control experiments.



AIAA 2001-4423

Figure 9. Test article mounted with the VFT custom rig in ODU 4 x 3 tunnel.

Test Article Simulation

Model Support Rig

Simple PropulsionModel

Flight ControlModel

if

AerodynamicsModel

Simple Weight andBalance Model

Wind Tunnel Codeand Tools

Figure 10. Simulation structure for VFT test design.



AIAA 2001-4423

D-Six Simulation Environment

Test Article Simulation

———• R-opulsion ------ Aero -j————-Forces Moments

SurfaceCommands

360° Pitch

Strain Gauge Balance

V777////X/////////X

_ Balances andSensorsVoltages

Figure 11. D-Six test article simulation and data reduction integration during VFT.

J1MO. ..........^.....................^ . ^

rz*rz:z:zr:r:̂ ^ ;:

'':;'';;''';'';;'':';"

T3ictaMeaso?e<5\''

ftcabte «

O..QSGOQJSOK

030600X000aoooo

lis^OOOp-2.2S40 0,0000

&OOQOftCCOOS30CO1-0026

PS

Figure 12. Screen shot from the D-Six user interface during the VFT effort.


Documents

[American Institute of Aeronautics and Astronautics AIAA Modeling and Simulation Technologies Conference and Exhibit - Montreal,Canada (06 August 2001 - 09 August 2001)] AIAA Modeling